As a leader, you’ve probably spent years nudging your managers to create psychological safety. After all, organizations benefit greatly when it feels safe to challenge the boss, surface bad news early, and say the uncomfortable thing before it becomes the expensive thing.

The Space Shuttle Challenger disaster of 1986 has become the canonical warning. Engineers raised concerns about the O-rings used in the solid rocket booster before launch. Their warnings were overruled by managers under schedule pressure. Seventy-three seconds after liftoff, seven astronauts were dead. The lesson still holds: When people are afraid to speak or are pressured to comply, organizations miss what they already know.

AI, however, is introducing a complication most leaders haven’t accounted for. Your employees now have another place to speak, and emerging research suggests they may be more willing to challenge a machine than a human leader, while another body of research challenges our assumptions about psychological safety itself.

Put those together, and you’re no longer managing psychological safety only between people. You’re managing it across people and machines, and the two may tell you very different things about your organization’s health.

Safety needs accountability

Psychological safety became popular for good reason: Teams perform better when people can admit mistakes, question assumptions, and challenge authority without fear of it following them into a performance review. But here’s the misconception: Psychological safety and accountability sit at opposite ends of the same dial. Turn safety up, the thinking goes, and standards come down. They don’t. The strongest teams have both; people feel safe enough to challenge one another and accountable enough to deliver.

Researchers studying five workplace settings, from hospital nurses to knowledge workers to four years of sales data across 229 retail stores, found that the relationship between psychological safety and performance wasn’t simply linear. At high levels of psychological safety without matching accountability, performance on routine work could decline; collective accountability substantially buffered that decline. The problem wasn’t that people felt too safe; it was safety operating without another ingredient the team needed.

A newer conceptual paper helps explain why. Psychological safety lowers the interpersonal risk of speaking up, which is what you want, but safety alone doesn’t create the discipline that keeps a team sharp: preparing for meetings, thinking carefully before weighing in, and taking shared norms seriously. That discipline comes from accountability.

Aviation offers a useful analogy. Sterile cockpit rules eliminate nonessential conversation during critical phases of flight, not safety-critical challenge. If something is wrong, you still speak. What changes isn’t how much psychological safety is allowed. It’s how tightly decision rights, operating discipline, and accountability are enforced.

Your AI pilots and early experiments need maximum candor. A junior engineer questioning a model or vendor claim should have every opportunity to be heard. Once a decision moves into production, incident response, or customer-facing execution, accountability tightens. Psychological safety doesn’t go down; accountability goes up.

Culture won’t automatically make people trust AI

Here is another set of assumptions in this area for IT leaders to challenge: that inclusive leadership, meaning leaders who stay approachable and treat employees’ input as worth acting on, builds psychological safety, that psychological safety makes people comfortable with AI, and that comfort with AI drives adoption. The evidence is messier.

One recent study found that inclusive leadership did improve psychological safety, but psychological safety itself did not significantly predict trust in AI. What mattered more was an individual’s ability to tolerate uncertainty. That matters because AI adoption is an uncertainty event: Capabilities shift every few months, outputs can be confidently right and impressively wrong, and the impact on jobs, workflows, and careers remains unclear. Employees are being asked to exercise judgment around systems whose behavior is probabilistic, not deterministic.

Another study of employees in South Korea found that AI adoption was associated with lower psychological safety, which in turn was associated with greater employee depression. Ethical leadership, meaning leaders who are transparent and fair about what’s changing, softened that relationship but didn’t eliminate it. Neither study is universal proof — different countries, samples, and methodologies matter — but together they raise an important question for CIOs.

A healthy culture may help people absorb the disruption AI creates without necessarily making them trust the technology itself. Those are two different problems. One is organizational: Do employees believe they can question what is happening around them? The other is psychological and technical: Can they work productively when the tool is probabilistic, and do they know when to trust it, challenge it, verify it, or ignore it? That requires more than another culture initiative. It requires AI literacy, judgment, and comfort operating under uncertainty.

Your people may be honest. Just not with you.

A 2026 series of scenario-based experiments with participants in China found that employees were more willing to voice concerns to an algorithmic leader than a human leader during cognitive tasks. Perceived fairness and psychological safety helped explain the difference. That doesn’t mean employees want robots for bosses — the effect didn’t hold in the same way for emotionally charged situations — but for analytical work the finding is worth paying attention to. One possible explanation feels remarkably human: An algorithm can’t embarrass you at the next staff meeting, downgrade your reputation with the executive team, or remember that you challenged its favorite project six months ago.

Now think about what happens as generative and agentic AI become embedded in everyday work. Somewhere in your company, an engineer may tell an AI assistant that the proposed architecture is flawed, explain why, work through three alternatives with the model, then walk into a meeting an hour later and nod along while a senior leader advocates for the original design.

The concern existed. The organization just never received it. AI didn’t suppress the dissent; your management system did.

You’ve created a psychologically safe chatbot sitting inside a psychologically unsafe org chart, and your AI adoption dashboard will never show you the difference. Both employees look like successful AI users, one using AI to accelerate work, the other using it as the only place they feel comfortable saying what they really think. Usage looks identical. Organizational health doesn’t.

We spend a lot of time asking whether employees trust AI. We may need to spend just as much time asking whether they trust people enough to repeat what they already told it.

What only you can fix

First, keep psychological safety constant and let accountability flex by phase. Push for aggressive challenge during strategy sessions, architecture discussions, AI pilots, and early experiments, and make disagreement cheap before decisions get expensive. Once execution starts, tighten decision rights and expectations without closing the channel for legitimate safety, ethical, or operational concerns.

Second, stop measuring psychological safety only through sentiment. Ask whether dissent changes decisions: what major initiative was redesigned because someone challenged an assumption; what AI pilot was narrowed, delayed, or killed because someone surfaced a risk early; what executive decision changed because somebody several layers down challenged the prevailing view. If nobody can remember one, another engagement survey isn’t going to tell you much.

Third, build a path for concerns raised to AI to reach the human decision system. That doesn’t mean reading employees’ private AI conversations; doing that would create exactly the surveillance problem that destroys candor. Instead, design mechanisms that let employees elevate a concern without reconstructing it from scratch: an anonymized escalation channel, a prompt-to-retrospective workflow, a mechanism to flag questionable AI outputs, or a structured path from an AI conversation into an architecture or investment review.

The design question is simple: Can an employee move a concern discovered in an AI conversation into the organization’s decision process with one deliberate action? Increasingly, that’s where the concern may show up first.

None of this argues against psychological safety. It argues against assuming psychological safety works in isolation, or that human trust and AI trust are the same problem. AI is creating a second conversation inside your organization, one that happens in meetings, reviews, and Teams calls, and one that happens quietly between an employee and a machine.

The real question isn’t whether your employees trust the AI. It’s whether AI has quietly become the place where they say the things your organization most needs to hear.

Second Opinion offers technology leaders an evidence-based perspective on their most consequential decisions. Khalid Kark brings 25 years as a trusted advisor, researcher, and operator. In an enterprise technology market crowded with noise, this column separates signal from spectacle.

Every year, thousands of enterprise organizations walk into technology negotiations trapped by the same fundamental mistake: they negotiate a product instead of solving a problem.

Having advised over 2,000 enterprise customers annually on technology acquisition over two decades, I see this pattern play out daily. Purchasing committees meticulously list the specific features, software modules and server specs they think they need, converting their organizational requirements into a static shopping list. On paper, this strategy feels disciplined. It creates a standardized, predictable item that procurement can use to negotiate a line-item discount with providers.

When you buy a fixed product, you get the immediate, gratifying perception of a lower price. But that upfront discount can come at a hidden cost: you risk cutting yourself off from the disruptive innovation of emerging technologies like artificial intelligence (AI).

By prescribing the solution before the negotiation even begins, enterprises force vendors into the narrow role of commodity suppliers rather than strategic partners. You ask them, “How cheaply can you build what our internal committee specified six months ago?” instead of asking, “What is the most efficient way to achieve this outcome today using capabilities we might not even know exist?”

In an era where AI can reshape entire operational workflows, anchoring your procurement strategy to fixed product specifications means you may pay less for yesterday’s way of doing business while missing the value of new capabilities.

When procurement treats technology as fixed products, it tries to bundle mature products, such as software-as-a-service (SaaS) licenses, with rapidly evolving capabilities, such as AI services billed by token consumption, into one grand contract. You end up applying three-year static product licensing to AI capabilities that change on a monthly basis, locking your company into rigid pricing models and outdated technology.

A request for proposal (RFP) is the ultimate manifestation of product-first buying. It forces vendors to answer 200 static questions about prescribed features rather than offering alternative solutions. Procurement celebrates negotiating a 15% discount on a legacy customer service product, completely unaware that an outcomes-based AI approach could have eliminated 70% of the support tickets altogether. You win the unit-price battle, but lose the total cost of ownership (TCO) war.

Procurement departments use RFPs with the goal of driving down unit costs and reducing risk. But as behavioral expert Rory Sutherland has argued, corporate purchasing can prioritize blame avoidance over value creation. RFPs require the buying organization to specify the technical solution long before it enters a room with the experts who could help solve the original problem.

The RFP backfires in two distinct ways:

You force vendors to answer the wrong question

An RFP asks vendors: “How much will you charge to build what our internal committee decided we need six months ago?” It denies the seller the ability to say: “The problem you are trying to solve can now be eliminated using a new approach you haven’t considered.”

Shifting costs rather than reducing them

RFPs do not necessarily lower TCO; they can shift costs around while inflating administrative overhead. Software goes unused, becoming shelfware, because the focus was placed on winning the RFP pricing battle rather than user adoption and operational enablement. The company then has to spend more on adoption, while employees may turn to unsanctioned tools, known as shadow IT, duplicating costs and eroding the negotiated savings.

When enterprise buyers attempt to simplify procurement by bundling SaaS, infrastructure-as-a-service (IaaS) and AI into a single, multiyear contract with a hyperscaler, they can make switching providers harder. Research published by Harvard Business School on platform strategy explains how bundling can reinforce customer retention even when one component disappoints. Buyers need to evaluate the flexibility of each component, not just the combined discount.

A good example is what happened to one of my customers: an unmonitored script loop written by a developer unexpectedly burned through 18% of their annual infrastructure commitment in AI token consumption over a single weekend. By bundling emerging, unpredictable technology with stable infrastructure services, they exposed an otherwise predictable IT budget to financial volatility.

Buying technology correctly means matching the contract model to the technology’s maturity stage. Treating an emerging, fast-evolving space like AI with the same three-year static lock-in used for corporate email is a governance failure.

Another of my customers, a regional insurance company, issued a 200-question RFP to replace its customer service center software, seeking an AI-driven system to handle support tickets 20% faster.

Consider the limitations of that RFP: procurement issued a spreadsheet demanding details on call-center headcount, staff rotation rates and employee training certifications. An AI startup whose autonomous agent framework proactively fixes customer problems before a support ticket is even logged refused to participate because the questionnaire assumed a human workforce. The deal went to a legacy vendor that answered all 200 questions and offered a 15% discount. Procurement celebrated a price drop on seat licenses, even though it had just committed millions to perpetuating the existing operational model.

The buyer saved 15% on seat licenses, but spent $3 million over two years maintaining the call-center workflow. Had the company abandoned the rigid RFP, it could have eliminated 70% of its call volume entirely. Procurement successfully minimized the unit cost of the contract while missing the potential value of a different capability.

If we want to stop buying technology wrong, CIOs and purchasing committees must adopt three fundamental shifts:

1. Decouple contracts by maturity stage: Keep mature SaaS, optimizable IaaS and emerging AI in independent contractual structures. This preserves the flexibility to swap out emerging tools as better models hit the market without triggering financial penalties on core systems.
2. Replace rigid RFPs with proof-of-value (PoV) sandboxes: For disruptive technologies and services, replace prescriptive questionnaires with small, rapid experiments that allow vendors to prove outcomes. Define the business problem, success criteria, budget and security requirements before testing. Pay to solve the problem, not just to acquire the product.
3. Prioritize local enablement and FinOps over direct discounts: A 10% direct manufacturer discount means little if 30% of your cloud infrastructure is overprovisioned or if your employees never adopt the software. FinOps aligns technology spending with business value through collaboration among engineering, finance and business teams. Evaluate what your internal team or a multicloud partner can deliver in user adoption, local-currency billing, support for local tax requirements and ongoing cost management. Compare the full cost and return on investment (ROI), rather than the discount alone.

It’s time for enterprise leaders to stop managing technology procurement like a commodity auction. The goal of buying technology isn’t to get the lowest price on yesterday’s specs; it is to secure the greatest speed, flexibility and business value.

A LinkedIn post about my old company, insightsoftware, popped up the other day. It was lamenting the challenge of turning 29 acquisitions and close to 80 products into one comprehensive platform. I spent years in the middle of that work, and what I know now is how much harder that consolidation is than you might think, and how much of it you only learn by getting it wrong first. 

So, when I arrived at Veson in November and found a similar landscape, I felt a bit of déjà vu. A bit of “I’ve done this before, so I know what I’m doing” and a bit of “I’ve done this before and I know it’s no cakewalk.” Veson had spent the past several years acquiring a range of products like Shipfix and VesselsValue and now needed to consolidate them onto a platform alongside its flagship IMOS solution. A new variable is also now at play that changes the whole equation: the advancement of AI.

While AI may sound like a cheat code for this work, it’s very much a double-edged sword… with a grenade taped to it. 

Yes, we can do some things faster and differently than we ever could a few years ago. But we’re also learning as we go and upskilling our team as we go, on a technology that’s changing under our feet.

The newness of these tools multiplies the complexity in a way I didn’t fully appreciate at insightsoftware. Back then, the tools simply didn’t exist or were just emerging. Now your platform must transform so that it thrives in the age of AI. It has to be built in the age of AI, which — depending on whose study you believe — means anywhere from 2X to 10X faster.

And then your team needs to adapt to the age of AI, which again means faster, at higher quality and producing more output. That’s a lot of spinning plates, on top of each other, none of which you’re allowed to drop. 

Oh, and one other plate: this is a centuries-old industry that moves billions of dollars in freight a year and does not stop for anyone. You don’t get to tie the ship up at the dock while you rebuild the engine. Business continuity, zero migration, no breaking the thing people are already running their business on. And do it as fast as you possibly can. 

Building a connected, modern platform

At Veson we had a fleet, not a unified experience. A set of strong, separate hulls, each with its own data, its own experience, its own idea of the world. A collection of great products is not the same as a cohesive foundation, and pretending otherwise is a mistake.  I’ve been guilty of this myself: combining products superficially, so they sit side by side under one roof but still behave as separate entities, rather than as one integrated experience.

In fairness to my past self, before AI even that superficial stitching was slow, labor-intensive and risky, and it often used up the time and budget we’d meant for real integration. AI has worn much of that cost away. The plumbing that used to take months now takes weeks, which frees us to focus on what customers value most: products that work as one.

A platform, in 2026, is three things stacked in order. A singular data layer underneath. One experience on top. And then AI that can reason across all of it. You cannot bolt credible AI onto a flotilla; it has nothing coherent to reason over. The unglamorous work of unifying data is what makes the glamorous AI work possible. 

The part I’m most wary of talking about — because I would rather not tempt fate — is doing all of this safely. Moving fast on all fronts at once is a lovely thing to say in a strategy meeting and a dangerous thing to do in production. The mistake I’ve watched happen (and, if I’m honest, helped cause) is measuring efficacy too narrowly.

If you pick only a few metrics (that you know you can nail) and tell people that’s how they’ll be judged, teams will act accordingly. The side effects get ignored or missed until it’s too late. Often that side effect is quality, discovered downstream when it’s far costlier to fix, or market fit, because value was quietly descoped to hit a date. With AI coding agents accelerating everything, I wasn’t willing to repeat that. We needed 360-degree checks and balances.

The one thing I insisted on and set up from the first week with Veson was working with the engineering intelligence platform Plandek. We measure speed, quality, predictability and focus in real time, so we can see the progress we’re making and spot the side effects before they compound.

AI let us pick up the pace; comprehensive measurement is what lets us keep it without losing sight of the risk. As a result, we get to have our cake and eat it, we upskill the team while the ship is sailing.

Ten months in, here’s what we’ve learned so far.

Product lessons we’ve learned

• Don’t rebuild what already works. “Modernizing” something people rely on is the fastest way to destroy the value you paid for. 
• AI only counts when it functions inside a workflow people already trust. An assistant in a corner nobody visits is a demo, not a product. 
• Aim for a minimum lovable product, not a minimum viable one.
• Continuity is a feature. The moment a customer must “migrate” to your platform, you’ve lost them. 
• One source of truth beats five clever ones. Data you unify late is a tax you pay forever. 

Experiences from our engineering teams 

• Adoption will be uneven. Don’t force the sceptics; convert them with proof. 
• The job itself is changing from writing code to directing it — reviewing, steering and trusting AI as it writes, then verifying. 
• Spread what works peer to peer. A network of AI champions beats a mandate from someone who clearly hasn’t done the work. 
• Measure everything. 
• The fastest teams are the ones who trust the platform beneath them. 
• Faster is only a win if quality holds — protect the gates hardest exactly when the pressure to cut them is highest. 

I don’t think we’ve cracked the code on this. We’re in the middle of it, plates still spinning. But this time I’m making new mistakes rather than old ones, and I’m more convinced than ever that we’re doing this the right way.

For decades, CIOs have been challenging their organizations to shed one of the most persistent labels in our profession: order taker. Instead of waiting for the business to tell them what they want, technology leaders need to understand what the business is trying to accomplish, get invited into the first meeting, and earn the right to become a strategic partner. But in a fast-moving environment, is being a trusted advisor enough?

To reach the next level, top CIOs are increasingly transforming their IT organizations into what I call “innovative anticipators”: These IT orgs understand the business, customer, and possibilities so deeply that they can see around corners, anticipate what’s needed next, and create value before someone asks for it.

Janet Robertson, global CIO of RS Group, believes building that anticipation muscle isn’t simply about moving the IT organization to a higher level of maturity. Increasingly, it’s a business imperative.

For a recent episode of the Tech Whisperers podcast, Robertson and I explored the deeply human side of leadership: how trust goes both ways, how great leaders serve as a lighthouse when others are navigating through the fog, and how we prepare the next generation to lead in a world none of us has led in before. But hiding inside that conversation was another fascinating story to unpack.

RS Group is a global product and service solutions provider for industrial customers operating in 33 markets. Its journey has been remarkably similar to the one our IT profession has been making — from taking orders, to advising customers, to anticipating what they might need next. As Robertson evolves IT at RS into an innovative anticipator, the company as a whole is doing the same.

Two journeys, similar challenges. And as Robertson explains in the following conversation, strategic partnership may no longer be the destination for today’s CIO. It may be the starting point for what comes next.

Dan Roberts: You’ve said something that sounds simple but has enormous implications: There isn’t a technology strategy and a business strategy. There’s one business strategy. What changes in the way a technology organization operates when people genuinely believe that?

Janet Robertson: When the technology group is focused on an outcome, they’re looking at solving a business problem more than implementing a technical solution. That’s a nuance, but it’s really the most direct route to a solution, because they’re trying to solve a business problem and get a business outcome more than implementing a technology solution or a technology transformation. It’s a different mindset.

In practical terms, you end up with increased adoption because you’ve got buy-in from the business on the direct answer to the problem. You also end up with fewer bells and whistles — a solution that’s not gold-plated and that doesn’t have things you won’t need, which is more efficient, considering how fast things are moving today. Everything could be different.

In fact, we’re doing product development now as more of a continuous evolution using new, innovative tools that are available to us at each space. The days of planning a three- or four-year program are over. By the time you’re going live, the technology’s out of date, or the processes have changed, or the world’s changed around it. You’ve got to create solutions fast with the most direct available technology to support the business outcomes.

You can’t anticipate a business you only understand through dashboards. How do you get your technology leaders close enough to the work, the customers, and the people to see what they otherwise might miss?

I encourage my leaders to spend time in the everyday working life of the recipients of our solutions. Sit with some of our biggest customers and watch them place an order through our website. Sit with our middle- and back-office team and watch their process of paying an invoice, receiving an order, entering an order, processing an order. Go to the distribution center. Spend time with the people who are using your solution.

Our customer experience officer has been riding around with salespeople, experiencing what they are going through in a day-to-day environment. I’m going out to Milan, Italy, next week to open our new our automated warehouse there. I’m almost more excited about launching it and seeing what we’ve implemented in use than I was planning the program. I want to see the ship sail in clear waters.

When technology leaders spend the time doing this, they start to see the connection directly between what they’re putting out there and what actually has to take place. You get very different perspectives from different sides of the fence when you get hands-on. It’s less about KPIs and dashboards and the results out of the other end of it. It’s more about understanding who’s consuming what you’re producing and how you’re providing value to them. Because sometimes, their biggest pain point is something you put down as a priority 5, and it’s in the lowest level of the backlog, but would completely change their perspective, improve their experience, and drive value.

What have the best consumer experiences taught you about orchestrating a B2B customer experience that will continue to differentiate RS Group?

Our catalog is massive, and we’ve got so much to navigate. As a consumer, with the new developments you can go on to an Amazon experience and find anything you want using even cryptic keywords. Or you can go on to Google and find 20 different buying options. That is absolutely transformative for us, because a lot of our customers buy that way. They’re searching for something that’s maybe very specialized, so they expect the same kind of Amazon-type experience when they’re looking for a specialized manufacturing-line repair part.

For us in the technology space, it means we’ve got to make sure that our data and our foundations are prepared to enable those large language models for the AI and generative search for agentic purchasing. What’s a little bit different from the B2C space is that many of our customers are starting to implement agents to do a lot of their procurement. So we’ve got to be prepared on the receiving end of those agents to be able to place those digital orders. It’s not really an e-commerce, add-to-cart channel anymore. Now it’s a direct agentic experience in purchasing for our bigger customers that use agentic purchasing.

To lead the way on that is a huge differentiator, and we are aware of it, and we’re keeping an eye on it. We have to be able to move fast and pivot quickly. Who would have thought two years ago that we were going to be talking to agentic procurement?

For years, we’ve challenged IT organizations to do a better job demonstrating their value to the business and to become trusted advisors. How has your organization made that pivot, and what does the next evolution of that journey look like?

When I first joined RS, my department was seen as less of a technical advisor or an expert in new technologies and new ways of getting things done. We were seen more as order takers and a cost center, basically a necessary evil. The pivot we’ve gone through has happened by listening and engaging with the business to understand and help them articulate what problems they’re truly trying to solve. I say, “Please tell me the what, and let me and my experts provide you the how.”

It’s taken some time, honestly. That evolution from an order taker to a trusted advisor is a journey, because you’ve got to prove that you know what you’re talking about, that you’re up on the latest technology, you understand the latest tools, and you are thinking outside of the box — truly providing the vision and the art of the possible to a problem they don’t know how to articulate. The biggest driver for me is connecting my leaders and my team to the business in general. They’re out looking for tools to solve those business problems.

Now we’re starting to partner with our business partners to show them the art of the possible with some of the new trends that are coming, like agentic procurement and large language model search. SEO is gone; now it’s GEO, generative engine optimization. So instead of “Tell me your business problem,” it’s moving to “Let me tell you how to solve a problem you might not know you have yet. This is a way to anticipate what is probably coming down the pipeline, and it’s going to be a problem soon.”

Because we’re leading the way in talking to our business about where the future’s going in technology and customer search and customer analytics and data science, we’re being brought in to discuss these developing trends.

That level of anticipation requires intelligence, or what you’ve described as “intelligent automation.” What’s involved in building that kind of anticipation engine?

Just like you can’t expect a 300-year-old house to be a smart house overnight, you can’t turn a 100-year-old business into an efficient, highly data-connected enterprise overnight. We know that we’ve got some legacy plumbing that we need to repair, and we are well into that journey now. We know that the data we have is incredibly valuable, but we need to understand how to use it, how to interrogate it, and how to get the right answers out of it. Without a structured format, and without all the speed and the pipes and efficient use, we can’t take advantage of that.

Bad data and a bad foundation only accentuate bad decision-making from the data. If you put a lot of intelligence over the top of an unstructured or bad format and foundation, you get really bad data to make decisions from, and it just makes it that much worse. So sometimes you have to build a little bit of a foundation before you’re able to just run with it.

For years, the goal for technology organizations was to become the trusted advisor who gets invited to the first meeting, not just be handed requirements after the important decisions have been made. Robertson’s experience suggests that today’s CIOs need to challenge their organizations to go further, from “Tell me what you’re trying to accomplish” to “Let me show you what’s coming.” That requires getting close enough to your business colleagues and customers to see what they may not see yet. It’s the difference between being invited to the first meeting and being the reason the meeting gets called.

For more from Janet Robertson’s leadership playbook, tune in to the Tech Whisperers.

See also:

• IT reinvention requires mastering the art of translation
• What separates true enterprise leaders from strong tech execs
• The 7 new Cs of IT leadership for the AI era

There is a fundamental problem hiding behind today’s technology forecasts: the technology industry is projecting spending growth at rates that many of its customers cannot economically sustain.

Gartner’s July 2026 forecast calls for worldwide IT spending to increase 14.2% in 2026 to $6.37 trillion. Within that number, software is projected to rise 15.5%, Infrastructure-as-a-Service 29.3%, and data-center systems an extraordinary 62.5%. Gartner itself acknowledges that technology budgets are being strained by inflation, supply shortages, hardware and memory costs, and AI investment.

The AI projections are even more dramatic. Gartner’s September forecast puts 2026 worldwide AI spending at $2.7 trillion—up 49.5% in a single year. McKinsey has separately estimated that generative AI alone could drive $175–$250 billion of software spending by 2027, adding two to six percentage points to software-sector growth.

These are extraordinary numbers. But they expose a fundamental flaw in the way the technology market is being viewed: The forecasts measure what the technology industry expects to sell—not necessarily what its customers can afford to buy.

That distinction is critical.

The forecasts may describe what technology suppliers want to sell, what hyperscalers intend to build, and what analysts believe enterprises will demand. They do not demonstrate that the typical enterprise has the economic capacity to increase its technology budget at anything approaching those rates.

That is the collision illustrated in this cartoon. Most companies cannot indefinitely grow technology spending at 10%, 15%, 20%, 30% or even 50% when their revenues, operating margins and the economies in which they operate are growing at only a fraction of those rates.

And the affordability problem is actually worse than the headline spending forecasts suggest.

Enterprises simultaneously face the economic headwinds: IT inflation; AI premiums; rising compute, server and memory costs; cybersecurity; energy and datacenter requirements; technical debt; scarce technology talent; software and cloud price escalation; regulatory and control costs; and the continuing expense of maintaining yesterday’s technology while financing tomorrow’s.

The result is what I call the Technology Affordability Gap: Technology demand + technology inflation + AI investment + infrastructure requirements are growing faster than the economic capacity of many enterprises to pay for them.

This produces an extraordinary paradox. Every technology forecast can appear individually plausible from the supplier’s side while their collective implications become economically implausible from the customer’s side.

A hyperscaler can forecast enormous cloud growth. An AI company can forecast enormous AI adoption. A software company can forecast higher subscription revenue. A semiconductor company can forecast explosive demand for AI infrastructure. An analyst can add all of those expectations together.

But ultimately there is only one customer balance sheet paying the bills.

Something, therefore, has to give.

Enterprises will have to slow spending, eliminate legacy costs, consolidate vendors, renegotiate contracts, defer lower-value projects, substitute AI and automation for labor, radically improve technology productivity, and demand demonstrable economic returns from AI and other technology investments. Otherwise, technology spending will increasingly have to be financed by taking money away from other parts of the enterprise— or by sacrificing margins.

That is the impending Technology Economic Crisis.

It is not a crisis caused by a shortage of technology. It is potentially a crisis caused by more technology ambition, more technology supply and more technology spending expectations than the underlying economy can afford.

The question for 2027 and beyond, therefore, isn’t simply: “How much technology will companies want?”

It is: “How much technology can they actually afford?”

AI is writing more code than ever. Software specialist Sonar’s 2026 State of Code Developer Survey, based on more than 1,100 professional developers, found that 72% of those who’ve tried AI coding tools now use them every day, and that today, AI now accounts for 42% of committed code, a number that’s expected to reach 65% by next year.

But while developers embrace AI-enabled coding services, many professionals are uncertain about how their businesses will cope with this influx of AI-generated code. Sonar’s research additionally discovered that 96% of developers don’t fully trust it, and only 48% always verify it before committing.

In fact, 38% of developers told the researchers that reviewing AI code requires more effort than analyzing code written by human colleagues. That’s a pattern recognized by Tim Chilton, a managing geospatial consultant at UK mapping agency Ordnance Survey (OS). “There’s truth in the headlines that AI can get carried away, and developers spend more time fixing the code that’s been created than not,” he says.

While that’s not a pattern he sees at the OS, Chilton says his organization is cautious about how developers use AI coding tools, and conversations with other digital leaders suggest a similar approach. AI-enabled coding is indeed almost ubiquitous, and it can allow developers to reach new, higher levels of productivity, but more doesn’t necessarily mean better code.

To ensure that such coding services produce business benefits, CIOs must take control of the development process. They should exploit agents carefully, set professional responsibilities, and establish guidelines, policies, and organization-wide training.

Exploiting the agents

Snowflake co-founder and president Benoît Dageville suggests there are two ways to view AI-powered coding. One is the scary way, where AI takes over your tasks without human input. The second is the efficient way, where AI helps you do the same things as before, but at a fraction of the cost. He’s a big believer in the latter, both for himself and his teams.

“The route depends on the organization; their application of AI coding depends on their priorities,” he says. “However, I hope that overall, AI will have a positive effect on humans. I’m using AI every day to code, and if I compare myself to one or two years ago versus now, I’m more efficient and productive, but I’m also writing five times more code.”

Steve Lucas, CEO of integration technology specialist Boomi, also notes the increase in output. He quotes research that suggests 4% of GitHub public commits are already being authored by Claude Code. That proportion is expected to rise to 20% of all daily commits by the end of this year. Staff at Boomi already use AI tools to produce more code.

“We’re writing 10 times the amount today than we did two years ago,” says Lucas. The increase in AI-enabled coding is so significant that he’s spoken with Boomi chief product and technology officer Ed Macosky about goals for the proportion of self-modifying code.

“I believe AI should recursively modify 80% of a software organization’s code,” Macosky says. “Infinite improvement is demanded. If AI is extraordinarily good at breaking security protocols, the only way to keep up isn’t the kind of accordion model of human code design. AI use must be linear in scale and continuous.”

Crucially, Lucas suggests using AI to check and modify that AI doesn’t mean coders go away. “In fact, I think we’ll need more engineers, not fewer,” he says. “We need these professional experts to ensure the quality of the code that’s written.”

At OS, Chilton uses Coco, Snowflake’s coding agent. The tool does as much as 90% of a development task that he would otherwise complete from scratch. “It’s taken an hour to finish, instead of weeks or never,” he says, reflecting on the power of AI coding services.

However, the tool’s influence is carefully constrained. The agent doesn’t work in isolation, and Chilton checks its output in detail. He says all organizations need a similar process and should value experienced professionals who can use AI coding tools effectively.

“If you’re giving a tool to someone who isn’t capable of checking an agent’s work, as a digital leader, you run the risk of putting too much faith in AI just running with its output,” he says. “I think that’s a big risk, so there’s still a need for experienced developers.” 

Setting clear responsibilities

While placing humans in the loop can help ensure AI-enabled code is safe and effective, developers shouldn’t be weighed down by their new management responsibilities.

Gill Haus, CIO at Chase, the consumer banking arm of JPMorgan Chase & Co, says AI coding benefits can only be realized when agents are introduced with a strong focus on automating supporting technology and processes, which reduces the need for too much human input.

“If you have 12,000 engineers, they can write 12,000 engineers’ worth of code in a day,” he says. “If those 12,000 engineers have agentic technology, they can write much more. But if you don’t have automation left to right with confidence, you create more manual processes. You can’t move as quickly since there are only so many humans you can put in the middle.”

Haus says Chase never releases services in production that don’t pass its high bar for quality and security. Strong engineering practices, pipelines, and architecture, plus a tried-and-tested approach to automation, can help ensure AI coding produces solid outputs without a burdensome amount of human intervention.

“We’re focused on effective automation so we’re able to unleash this technology safely and effectively,” he says. “This is something many other companies should be focused on if they aren’t already, because otherwise you can’t keep up with the pace of change.”

François-Xavier Pierrel, group chief data and ad tech officer at French TV network TF1, is another digital leader who says maintaining control as AI writes more code means keeping the human in the loop in a clearly defined manner. His team documents agentic processes using Atlassian project management and collaboration tools, such as Confluence.

Pierrel says policy management is crucial as the agentic transformation gathers pace.  One key step is ensuring professionals who use agents conform to internal standards. Smart businesses set policies that state clearly when staff can explore an agent, delegate work to it, or release it in production.

“We don’t have hundreds of agents deployed yet,” he says. “But the moment you reach a certain level of volume, if you don’t have compliance and policies in place, AI becomes a critical mess, and then it can be disruptive in a negative way.”

Building strong foundations

Kirsty Roth, chief operating officer at content and technology specialist Thomson Reuters, agrees digital leaders must be proactive. Her firm’s explorations with models and agents suggest that AI can help with time-intensive coding processes, such as writing. This capability means her organization carefully considers the roles people fulfil.

“In the last 12 months, we’ve put lots of thought into where we’re seeing benefits,” she says. “It’s great AI can do some of the coding work, but what does that mean for people’s jobs? We used to have thousands of engineers. Do we now need more architects or product people?”

Roth refers to security. While cyber is always a priority for Thomson Reuters, the firm dedicates significant resources to controlling threats in the AI age. Developing high-level development and security skills requires a concerted effort. Yet Roth says ensuring people have the right skills to manage AI is in many ways a familiar IT trend.

“All this effort is no different than digital, cloud, mobile, social, or other big tech trends,” she says. “They all take several years to have an impact as the initial hype comes, and then you’ve got to work through the reality and think about how you upskill people for the new world.”

Crucially, coding is no longer the sole preserve of the software developer. AI coding tools make it much easier for non-IT professionals to vibe code applications. Tal Carmi, CIO at technology specialist WalkMe, says non-developers can create useful tools. However, effective outputs rely on a careful balance between enablement and guidance.

He adds that most of these employees don’t need in-depth capabilities, such as a database or a content delivery network. Instead, non-IT staff require a dedicated place to upload their products. WalkMe’s engineering specialists ask key questions to ensure their outputs are safe, considering credentials and potentially problematic packages.

“You need to add guardrails,” he says. “A lot of the work for in-house engineers is shifting from building to testing. The people who aren’t coders don’t necessarily have security in mind.”

The key takeaway, suggests Emmanuel Frenehard, chief digital officer at biopharmaceutical giant Sanofi, is that digital leaders must help professionals feel comfortable. Successful digital leaders ensure the business generates value from agents, and his firm is helping its staff, particularly its developers, to embrace these tools.

“I genuinely worry for professionals who don’t use AI to code, because in five years, they may not have jobs,” he says. “So we’re guiding, coaching, and training them, but people must take the journey. The train has left the station. You can decide to be on it, or just look at it.”

CIOs have always felt pressure to deploy emerging technologies rapidly. Implementation brings challenges, including calculating and measuring upfront investment costs. Deploy too quickly, and when the hype fades, many IT departments are left with a costly architecture to support, security gaps, and operational issues.

The same trend is happening with AI. Many CIOs procured enterprise-wide AI model licenses and empowered employees to experiment with language models and develop AI agents. Others asked DevOps teams to accelerate AI code generation and vibe coding. The shift from rapid experimentation to deploying more AI capabilities to production can leave CIOs with mounting AI debt and AI FinOps gaps.

CIOs’ near-term concern is managing budgets amid consumption-based AI pricing models. Even as AI token costs decrease, more employees are using AI, and agents are given more challenging, expensive assignments. Here are seven common mistakes IT leaders make that result in blown AI budgets.

Defining an AI strategy without measurable business outcomes

Most CIOs and CAIOs recognize the need to establish AI strategies focused on areas where AI can provide lasting business value. But many fail to quantify targeted business outcomes, leaving employees to use adoption metrics as the primary objective.

“Every token spent on a pilot that never proves its value is money leaders can’t get back and lost opportunity cost, which is a governance failure more than a technology one,” says Rob Scudiere, CTO at Verint. “Every agent and workflow should have to earn its budget by showing results that achieve the desired business outcome, or lose it.”

Kris Beevers, CEO and cofounder of Netbox Labs, says to use existing business metrics to see whether the areas adopting AI are improving them faster. “If adoption is climbing but the outcomes aren’t moving with it, that’s a pretty good signal to rethink where you’re spending,” Beevers says.

Recommendation: Financial impacts may be hard to target during experimental phases, but employees should set financial objectives before moving AI agents into development and deployment phases.

Experimenting without AI governance

CIOs may find they have a real mess on their hands when a few pilot AI agents quickly scale to thousands running in production. Much of AI governance focuses on identifying owners, setting data-access rules, defining target business outcomes, and establishing accuracy guidelines. CIOs must also educate employees on underlying cost factors, including model selection and platform standardization.

“Delayed AI governance gets framed as a compliance risk, but it’s just as much a budgetary concern,” says Guy Bourgault, head of agentic systems at Concentrix. “The longer organizations wait to define ownership, standards, and controls, the more they end up funding inefficiency, rather than progress.”

Varun Goswami, global head of product and AI at Newgen, says AI budgets are wasted on scattered pilots, duplicated integrations, and agents that run quietly without ever proving value. “Build on one governed platform rather than a dozen parallel stacks, reuse templates and connectors instead of custom-building each use case, and match the model to the job rather than routing everything to the priciest foundation model,” Goswami recommends.

Data quality is a third concern, according to Collibra and The Harris Poll, with 96% of IT leaders saying project failure almost always traces back to poor data foundations, and 51% reporting significant staff hours spent manually correcting autonomous AI agent outputs. CIOs should set data quality metrics and establish data products to build a trustworthy AI context layer.

Recommendation: CIOs should update and communicate their AI governance policies monthly, as AI cost, capabilities, compliance, and security considerations evolve rapidly.

Focusing tokenomics on model costs and not business value

Much of the discussion around AI token economics, or tokenomics, focuses on the costs of using different AI models. But the FinOps Foundation defines tokenomics as “the study of how the production, distribution, and consumption of tokens generate business outcomes and AI value within an organization.”

Because many organizations use multiple models, SaaS with AI agents, and internally developed AI agents, CIOs should partner with the CFO to measure total impact, including AI’s direct token costs, employee efforts, ongoing inference costs, and the value delivered.

“Companies can quickly overspend on AI when they focus too narrowly on token consumption, and the starting point should always be the business outcome,” says Prakash Kota, CIO at UKG. “I use the Three T’s: does the work require talent for judgment or expertise, can existing tools solve it efficiently, or do tokens create enough value to justify the investment?”

CIOs should then set AI budgets at the employee, department, and workflow levels. Each group’s budget should be reviewed and adjusted regularly based on outcomes. This discipline promotes fiscally smart decisions about when and how to use AI capabilities and can drive a top-down culture change.

“The same tokens can go to a skilled engineer shipping a product faster, or to a non-expert generating 10 versions of the same slide deck,” says Tom Shea, CEO and cofounder at OneStream. “The line item in the budget looks identical either way, and true token efficacy requires governance to ensure those tokens are empowering experts to accelerate output, rather than fueling a massive cost explosion.”

Recommendation: Setting AI budgets too soon may curtail employees’ enthusiasm in learning and experimenting with AI. Establish financial targets based on a maturity curve to encourage experimentation first, then fiscal discipline when AI agents are deployed to production.

Developing AI agents without standards

CIOs should communicate these two cost considerations to DevOps teams using AI code generators.

First, according to Stack Overflow’s 2026 Developer Survey, 73% of developers use AI coding assistants and agents daily, but developers care more about quality (75%) over cost (38%) when shaping prompts and picking models. CIOs should set guidelines for balancing cost and quality tradeoffs, establish DevOps non-negotiables, and require code analysis in deployment pipelines.

“Coding is no longer a constraint, and the bottleneck has moved to how we ensure the correctness of code,” says Anurag Dhingra, SVP and GM of networking and collaboration at Cisco. “There’s just no humanly possible way to review every piece of code that is written by AI.”

The problem is that 96% of developers don’t fully trust AI-generated code and 48% always check their AI-assisted code before committing, according to Sonar’s State of Code Developer Survey. As large organizations may use several AI code-generating tools, they need to benchmark quality, cost, and developer satisfaction.

“Teams need clear guardrails and engineering standards, especially as AI agents accelerate development and can introduce costly rework,” says Andrea Malagodi, CTO at Sonar. “Having live data aggregated from your one or many providers, to enable you to see the cost per PR [pull request], is a good governance mechanism.”

Recommendation: While AI code generation is a remarkable breakthrough, it also brings new costs and quality considerations CIOs should track.

Using frontier models for all AI needs

Another cost consideration comes as organizations deploy hundreds of AI agents and provide developers several AI models to choose from. Developers may choose models based on preferences rather than performance and cost considerations.

Craig Wiley, VP of AI at Databricks, says 77% of companies now juggle multiple model families because the best model for one task is rarely the best for another. “The flexibility comes at a cost of managing shifting prices, scattered vendors, and zero visibility into where your budget’s actually going. The fix is real-time budgets and smart routing that automatically sends queries to the most cost-effective model,” Wiley says.

Recommendation: Josh Fecteau, chief data and AI officer and CIO at Teradata, says that frontier models are expensive and not necessary for every task. “Diversify your model strategy with cheaper small language models, custom models, and industry-specific models,” Fecteau recommends.

Deploying AI agents without monitoring cost spikes and inefficiencies

A key challenge with AI agents is their variable costs, especially when they are given tasks of varying complexities. While some AI providers are moving to outcome-based pricing models, many still use a license model with a fixed number of tokens, plus consumption-based pricing for overages. This means CIOs must monitor costs, set AI observability standards, and define rules for handling performance variances in AI agents.

“I spent 25 years in telephony where every minute had a price, and agentic AI is the first software I’ve seen that bills like a phone call and gets budgeted like a license,” says CJ Combs, AI strategy executive at Columbus. “AI costs stop being predictable the moment an agent can loop, because it keeps pulling in context, retrying, and calling tools in ways nobody budgeted for.”

Experts suggest several implementation cost inefficiencies to review. Babak Pahlavan, founder, CEO, and CPO at NinjaTech AI, says that AI agent processing duration is an easy indicator. “A single autonomous agent running a multi-hour task can burn a month’s allocation in an afternoon because every tool call, every retry, every piece of context it re-reads gets billed,” Pahlavan says.

Other cost factors require experts to review an AI agent’s decision logs to determine optimization. Adi Elimelech, CTO and cofounder of Jedify, says, “If every query drags an entire database schema back to the LLM, token costs scale with the size of your environment rather than the complexity of the question. The cost compounds, since the model often doesn’t get it right on the first pass and loops back for another round trip before landing on the correct answer.”

Recommendation: CIOs will need to mature several AgenticOps and site reliability engineering practices to monitor AI agents and respond to cost, security, quality, performance, and other incident types.

Building agents without a change management plan

One of the biggest gaps in deploying AI agents is that deployment is just one step in a change management process. AI governance requires identifying agent owners who are accountable for decisions and responsible for outcomes. These owners should also sponsor the underlying change management process.

“The biggest AI cost overruns rarely come from technology,” says Abhijit Kakhandiki, SVP and GM of digital business automation at BMC Software. “They come from organizations deploying AI without redesigning processes, ownership models, and governance.”

Part of the change process requires identifying human-in-the-middle responsibilities to review the agent’s recommendations. The review process builds trust and lets owners switch to autonomous decision-making for low-risk, high-volume, trusted recommendations.

“It’s not going to be a binary switch where, from day one, you’re going to trust AI agents to run tasks autonomously,” says Joel Hellermark, chief AI officer at Workday. “Progressively, they’re going to earn the right to run those workflows autonomously. If it’s your fiftieth time approving an expense below $50 for one of your direct reports that looks exactly on policy, you can trust the agent to run it for you.”

Recommendation: Growing adoption isn’t the end game, but it’s a key step toward delivering business value. Metrics to track adoption include AI response quality, trust scores, repeat usage, and value delivered to adopters.

A simple rule is that you can’t manage what you don’t measure. CIOs should set budgets, track costs, and measure outcomes, especially as more agents move into production.

Much has been made of Generation Z’s supposed inability to adequately conform to modern-day corporate culture.

A recent Business Insider article on Gen Z (born 1997–2012) workplace readiness contains some damning statistics. The piece cites a 2024 poll of nearly 1,000 business leaders, 75% of which described some or all of the recent graduates they hired as “unsatisfactory.” Arguably worse, it cites another 2024 study from Hult Business School, in which 37% of HR leaders said they’d prefer to have a robot or AI do a job than hire a recent new grad, while another 30% said they would rather just leave the job vacant. Quite damning, indeed.

Warranted or not, these complaints about Gen Z are not new.

NYU Stern professor Suzy Welch conducted a September 2025 study juxtaposing the values of incoming Gen Z workers with those of hiring managers. Welch’s study included survey data from 2,100 hiring managers and a behavioral assessment of over 45,000 respondents (7,500 of which were Gen Z).

In a Wall Street Journal Op-Ed, Welch writes, “A mere 2% of Gen Z members hold the values that companies want most in new hires, namely achievement, learning and an unbridled desire to work.” According to Welch’s study, Gen Z members mostly value well-being and self-care, helping others and authenticity and expression. Oh, heaven forbid.

Not to be outdone, in March 2026, another NYU professor wrote a Wall Street Journal Op-Ed on “Why Gen Z is unprepared for the workplace.” The one, from psychology professor Tessa West, laments Gen Z’s lack of communication skills and failure to understand older generations. West writes, “Younger generations—Gen Z specifically—have learned to communicate in a digital world, sending texts and instant messages instead of interacting in person.”

West attributes Gen Z’s poor communication skills to a combination of excessive screen time, online classes during the COVID pandemic and a general overreliance on technologies. I cannot definitively say whether these accusations hold water; however, I do believe that the answer doesn’t lie in replacing humans with AI or leaving roles unfilled. Hiring managers need to adapt to younger generations, whether they want to or not. And we’re seeing this in the help desk software space in real time.

Gen Z has little patience for traditional corporate bureaucracy

Aside from grievances from older folks like me, the fact remains that Gen Z has little patience for traditional corporate bureaucracy. When these younger workers experience a technology glitch, they don’t want to log into a clunky portal, fill out a 10-field form and wait two days for a level 1 support agent to respond. These kids expect instant, consumer-grade, self-service resolution. If the enterprise doesn’t provide it, they’ll just bypass IT entirely.

To overly generalize, most members of Gen Z are comfortable with AI chatbots and tend to search for their own solutions online before reaching out to another human directly. This tendency to rely on mediated communication as opposed to person-to-person communication makes sense given their life experiences with COVID, online coursework, the post-pandemic reevaluation of work culture and their status as a mobile-first generation.

Ramifications for the help desk software space

To retain young talent, enterprises should consider deploying conversational AI, automated self-service password portals and omni-channel support to match the speed at which Gen Z operates.

Self-service should be the first order of operation. Rather than forcing employees to create a ticket and escalate an issue to an agent, workers should have the option to remedy the situation themselves. Across the IT support industry, we’re seeing more emphasis on AI oversight and escalation handling, as opposed to level 1 IT support work.

Additionally, given Gen Z’s mobile-first status, it is vital to have mobile-native applications for internal IT issues. These shouldn’t just be desktop portals reconstructed to fit a mobile interface; these should be mobile apps with consumer-grade quality UX interfaces equipped with things like push notifications for ticket updates.

We’re also seeing more agentic chatbots in the ITSM space. Gen Z workers (and all generations, really) expect agents to be able to handle simpler requests, like access provisioning, password resetting and resolving calendar conflicts.

Omni-channel continuity has become vital too. When Gen Z’ers begin a conversation in Slack or Zoho Cliq, they expect that conversation to seamlessly flow into SMS, email, mobile apps and other corporate channels.

Key takeaway

When it comes to IT service management, Gen Z workers expect agentic self-service, a mobile-first UX, omni-channel support and the ability to resolve tier-1 support issues themselves. Despite the lamentations of some old curmudgeons like me, Gen Z isn’t going anywhere. And what’s more, Gen Alpha (2010 – 2024) is waiting in the wings. If we haven’t brought our service desk infrastructure up to Gen Z’s standards, now is the time.

It’s early days for AI-assisted work, and sentiments are mixed about its benefits: Does AI actually bring value and save employees time, or does it bog them down and add unnecessary activities?

Data from ActivTrak seems to vacillate between those two potential realities: The latest research from the productivity tracking software company says employees who use AI most heavily are working up to 90 minutes longer each day, yet its previous research, published earlier this year, found that, in general, all employees are actually working less.

The discrepancy may be due in part to the timespan and sample sizes of the data collection for the two reports, but it indicates that enterprises should still be wary of vendor claims about AI-related productivity gains. Measuring them remains incredibly difficult.

AI use compared

ActivTrak’s current findings are based on behavioral work data gathered from 9,140 knowledge workers across 467 enterprises over a five month period (April 1 to September 25, 2026). This sample is much smaller than that of its previous report based on 443 million hours of work performed by 163,638 employees across 1,111 organizations over a three year period (January 1, 2023 through December 31, 2025).

According to the latest findings, heavy AI users are working between 8.42 and 9.08 hours per day, compared to 7.58 hours for those not using AI. ActivTrak’s previous findings revealed that, overall, average workdays had decreased by roughly 9 minutes (a 2% decline).

On the productivity front, heavy AI users have 22% shorter focus sessions than light users (37.8 minutes versus 48.2), suggesting their work is more fragmented due to AI use; further, they switch apps 15% more frequently than non-users.

“Over time, those interruptions add friction to high-value work that requires concentration, critical thinking, and sustained attention,” Sarah Altemus, manager of the ActivTrak Productivity Lab, said in a blog post describing the latest findings.

ActivTrak’s previous report had found that over the period of the earlier study, focus efficiency, the percentage of work time spent in focused, uninterrupted activity, fell to a three year low of 60% with AI use, and average focused session time in 2025 declined by 9% from 2023. However, that same report found that daily productive time increased by 19 minutes, or 5%, to 6 hours and 36 minutes, with AI in the workplace. Furthermore, productive sessions grew by 13%.

The company’s latest report also tracked the productivity differences between users of tools with embedded AI and users of web-based AI tools. Those using embedded AI switched apps 0.6 times an hour, while those using AI on the web switched between 1.7 and 4.2 times per hour. That’s up to 4x more switching than by employees who don’t make use of AI at all, which creates what ActivTrak describes as a “swivel-chair tax” — the additional effort spent by employees due to the movement between apps.

Similarly, in its previous report, ActivTrak had found that, overall, task switching in general rose 12%, or 10-plus minutes during the survey period.

Interestingly, according to the latest research, light users spend more active work time in core job-specific apps (81.4% compared to 78%).  “In other words, more frequent AI engagement doesn’t necessarily translate into more focused work,” Altemus pointed out.

The larger issue is the way organizations manage the additional capacity that AI may create, Altemus noted. “When employees complete tasks faster, managers may assume they have room for more responsibilities,” she said. Without clear workload boundaries, time savings may be redirected into additional work that can be taxing.

The numbers often don’t tell the full story

ActivTrak has a useful view of how work moves across applications, noted Sanchit Vir Gogia, chief analyst at Greyhound Research, but he also pointed out that it sells the capability that produces these measurements.

Recorded use depends on what the instrumentation can detect, he observed. For example, an employee can use AI at work and still appear in the non-user group, and work on an unmanaged personal device sits outside the measured estate. In addition, an API or a background agent can finish work without producing a desktop session, and passive screen time, an offline meeting captured through a calendar, and paid hours are also different measures. Furthermore, switching apps is easier to observe than switching attention within an application.

Ultimately, he noted, those judging the value of AI need to establish what happened after the first task became faster. “An improvement per paid hour can conceal effort left off the timesheet,” said Gogia. “The [enterprise] board should be able to say whose day improved and whose work made that improvement possible.”

This article has been updated to clarify that some stats are not related to AI.

With 2027 quickly looming, IT leaders should soon be finalizing their objectives for the new year, one certain to be rife with many unknowns if 2026 is any indication.

This year began with the call for a pivot from AI pilots to AI ROI, but questions around AI governance and accountability quickly took center stage as agentic AI made further inroads in the enterprise and high-profile incidents showed the risks of AI gone awry and the value of security-first approaches to AI agents.

Now, CIOs must anticipate how AI will evolve in 2027, how they’ll get their enterprises ready for all the risks and opportunities of that evolution, whether their teams will have the skills to seize on technology advances, and more.

To get a glimpse of how CIOs are approaching next year’s IT agenda, we asked IT leaders to share the key questions they are asking to shape their strategies for 2027. Here’s what they shared.

How do we ensure our AI investments provide tangible value as the cost calculus changes?

The big question as CIOs plot their IT agendas for 2027 still centers on providing verifiable business value, as 98% of CIOs report that board pressure to demonstrate measurable AI ROI has increased since 2024, according to a 2026 report based on a Dataiku/Harris Poll survey of 600 enterprise CIOs worldwide.

Doug Gilbert, CIO and chief digital officer at Sutherland, a digital transformation services company, is among those IT execs pondering this question as he executes his fiscal 2027 budget, which started in July.

“When I’m out speaking to people and they say, ‘It would be great if AI could do that,’ I say, ‘Yes, but would it provide value? Are we building value or are we just burning tokens?’” Gilbert says. He notes that he and his CIO peers are being increasingly judged on the outcomes they deliver for their organizations, meaning they have more incentive than ever to ensure the technology initiatives they undertake deliver ROI.

But the calculus on that is getting trickier than it was at the beginning of 2026 — in part because AI total cost of ownership (TCO) is challenging to calculate, but also because an AI pricing reset is on the horizon. So, to that end, CIOs such as Gilbert are now thinking about what level of ROI is high enough to justify the investment.

That’s a key question to ask now and into 2027, Gilbert says, particularly as the cost of tokens continues to rise. An AI use case that provides marginal ROI could stop producing any value in the near future if tokens increase even slightly, he notes.

As a result, Gilbert is working to factor in those dynamics when he and his team work with the business to identify which AI use cases to pursue — and he expects to do so even more moving forward. “I plan to pay a tremendous amount of time thinking about whether [a proposed project] is the right one to pursue,” he says, to ensure “when you put math on paper there is value with the project.”

How do we ensure observability and auditability of our AI systems?

According to a 2026 survey from OneStream, 47% of senior finance and IT executives say they have made a material business decision based on inaccurate, incomplete, or outdated financial data in the past 12 months.

Citing that statistic, Gilbert says he’s also asking critical questions about the observability and auditability of all of Sutherland’s AI uses cases. He wants to be sure he, other executives, and users are confident the company’s AI systems always have high-quality data — and thus accurate outputs. And he wants IT to achieve that observability and auditability so that the company can accurately track, measure, and improve its AI deployments over time.

“We will not deploy any AI that we can’t observe,” he says, a policy that he put in place about a year ago and will continue in 2027.

A wide range of tools under the moniker “AgentOps” are emerging to help IT leaders monitor AI activity, tools, and costs, starting with observability.

How do we move to more autonomous agentic processes?

“The challenge most technology executives are grappling with at the moment is how you move AI from human-assisted agentic to much more autonomous agentic processes with minimal human intervention or supervision,” says Jody Fullman, CIO of Reconomy and recently recognized as one of the UK’s top 100 CIOs in the CIO 100 UK Awards 2026.

Moreover, how do you achieve that securely and in a way that’s economically viable? he asks.

Fullman is implementing strategies to address this challenge. He is strengthening IT’s agentic controls by aligning Reconomy’s governance program to an industry-recognized framework. He is also implementing tools to monitor and manage tokens. And he’s moving some AI workloads to lower-costs models.

All that, he says, will set up the organization to develop and deliver more autonomous agentic capabilities ahead, noting that he and his team have been transitioning successful proof-of-concept agentic initiatives into MVPs to put the company on the path to more autonomous agentic processes.

How should OpenAI’s Hugging Face incident shape my security strategy?

Security has been a central factor in IT agendas for years. But Jeff Stovall, CIO for Dallas, Texas, says the 2026 Hugging Face incident, in which autonomous AI agents built by OpenAI escaped their sandbox testing environment and hacked into the infrastructure of AI open platform company Hugging Face, has him adjusting his plans for 2027.

“Hugging Face demonstrated that there was a lot more peril in the world of AI as it relates to cybersecurity than what we were ramping up for,” he says. “There’s now this notion that AI may have more downside for us than upside for us in the near future,” he adds, explaining that AI-fueled threats to the organization are more significant than the positives its use brings.

“In the upcoming six months to a year, what I see is that we have to take a hard turn in thinking about how AI poses a significant malevolent threat to our organization when wielded by actors seeking to do harm,” Stovall says. He points specifically to hackers’ ability to use AI to create swarms of bots that can autonomously collaborate and exploit all vulnerabilities nearly instantaneously and all at once. “That’s a dramatic acceleration and amplification to the threats we face.”

In response, Stovall says he’s shifting more resources to maturing his security program in 2027 than he initially budgeted — drawing funds away from internal AI deployments to “shore up the walls even further” by adding more AI automation and agents to have “much more machine speed in security response.”

How do we use AI to transform?

Marco Bill, SVP and CIO at Red Hat, believes using AI for productivity gains is now table stakes. So as he develops his plans for IT for 2027, he’s looking to push that notion even further, asking how the company can use AI to transform processes as a whole.

“This question helps me to keep reminding ourselves that we’re not just here to do IT. It’s more of a guiding principle,” he says.

And that in turn is underscoring the importance of transforming the IT organization into a more collaborative and business-centric one.

As an example of what’s brewing at Red Hat, Bill points to the work his IT team is doing with the company’s sales team to devise ways to use AI to enable account managers to have better conversations with customers to close deals faster. IT is also working with the marketing department to make go-to-market process more powerful, and his teams are adding agentic AI to their own IT operations to create fully autonomous workflows that cost less to run and manage.

All told, Bill and the IT team at Red Hat are building further on the work many CIOs began in 2026 by helping to reimagine business processes to reap AI benefits and bringing AI transformation home to IT workflows.

How do we empower users to transform their work with AI?

“Citizen AI” is a rising trend in the enterprise, and this is something Reconomy’s Fullman will continue to address in 2027, explaining that his company has found that much of its “success with AI was not by being a tech project but by empowering our users and business colleagues to transform their work and processes with AI.”

Fuller believes that his company will see more AI-enabled transformation, and see that transformation faster, by continuing to put more AI in the hands of users throughout the organization instead of expecting IT workers to become experts in all the processes that could be transformed by AI.

As such, Fuller says the upcoming year’s IT plans will continue to put AI tools in the hands of the users, and do so with governance, security, and transparency in place. He’ll also leverage the skills of workers in process transformation roles, which were recently created and installed in Reconomy’s business units but still work closely with IT architecture and data teams.

“They’re plugged into tech and the business to make sure we’re prioritizing the right things and doing things the right way as we drive transformation,” he explains.

How do we prevent our AI governance council from becoming a bottleneck?

Company leaders at intelligent automation platform provider Celigo want every team to build with AI, something CIO Amy Farrow knows makes security and controls critical. But Farrow also wants to make sure she enables speed and doesn’t unnecessarily slow down innovation, noting that barriers could lead to shadow AI.

“If [teams] all have to wait too long, they’ll find a way around the process,” she says.

To ensure that governance isn’t a bottleneck, Farrow explains that her “governance council doesn’t approve projects. It sets the rules: risk tiers, design criteria, a checklist every agent has to pass.”

She adds: “The teams approve their own work against those rules, and we audit. My team owns the integrations, access controls, and observability. My job is making it faster to build the right way.”

What controls do we need to put in place for AI agents?

As AI agents increasingly enter enterprise workflows, IT leaders will need to the level of permissions agents should be granted relative to the people on whose behalf those agents will be operating.

Farrow, who previously served as global CIO at Infoblox and Lyft, believes agents should resolutely have fewer permissions than the people whose tasks they will be performing.

“Someone in finance can issue a credit or update our ERP because we trust their judgment. An agent doesn’t get that by default,” she explains. “We start every agent with the least autonomy required to achieve the outcome. It can interpret a request or recommend an action. Changes to records go through checks the model can’t override, and bigger changes need a person to sign off.”

This approach also has a fiscal benefit, she says, noting, “Less autonomy means fewer model calls, which keeps costs in check.”

As AI increasingly goes agentic, CIOs will increasingly need to plan for ways to address this issue in the year ahead.

What happens when vendor agents start acting inside our systems?

“Many of our SaaS vendors are shipping agents, and some can make changes to our data by default,” Farrow says of another key security- and governance-related question many CIOs are facing. “For 2027, I want vendor agents held to the same standard as any other integration.”

Farrow has other questions related to the rapid rise in vendor agents, asking, “What can they touch? Can we see what they did?”

IT leaders need to ensure they can answer those types of questions, noting that California’s and Colorado’s new rules on automated decisions start in January. She wants to be able to give strong answers to those questions, stressing that “‘the vendor’s agent did it’ is not a defensible answer.”

How do we meet immediate business needs while building a foundation for growth?

Mieko Shibata, CIO at R&T Deposit Solutions, sees this as “the overarching question” for him as CIO.

“That question guides our investment decisions and how we allocate our team’s time,” she explains. “Our strategic direction is consistent, but each year’s agenda must balance immediate business priorities with progress toward our longer-term technology foundation. We consider both the value each investment delivers and how it contributes to the environment we are building.”

Shibata is taking a multiprong approach in her response to that overarching question.

First, she’s ensuring technology modernization turns into measurable business value and that modernization initiatives are prioritized based on that value.

She is also considering how IT responds to growth opportunities while keeping its longer-term transformation on track. “Our response is to consider the immediate commercial value alongside the longer-term technology implications. Can we reuse the capability? Will it scale? What will it take to support? Does it help simplify our environment?” she says.

And she’s thinking about how the company prepares for AI’s evolution while maintaining clear accountability for its use. “For 2027, my focus is on building a repeatable approach to evaluating and adopting AI as it evolves, with clear business benefits, appropriate controls, and accountability for the results,” she says.

Is our AI vision ambitious enough?

For Lauren Wilkinson, CIO for financial advisor services at Vanguard, AI ambition is top of mind as she develops her IT strategy for 2027.

AI strategy should accelerate the business strategy and “push leaders to reconsider what outcomes with AI are possible, rather than simply finding ways to make today’s processes faster or more efficient,” she says.

“It helps to start with an aspirational business vision, including what a few big outcomes could look like. Then you can work backward to determine where AI and technology can fundamentally change what is possible,” she explains.

In her work with financial advisors, she’s found that they’re primarily using AI to support administrative tasks such as drafting emails, conducting research, and taking meeting notes. But she sees an opportunity to build on those use cases in the upcoming year to deliver on that objective for a more ambitious AI program.

Like many CIOs, Wilkinson wants to take AI to the next level at her organization and ensuring that ambition is high enough is paramount to achieving that goal.

Do we have the talent to win in an AI-native world?

Wilkinson believes how CIOs address the AI talent question will be critical for organizational success.

“As AI continues to evolve, the skills that differentiate great technologists will also evolve, placing an even greater premium on AI expertise and deep business acumen,” she says. “CIOs should be building teams that understand the business, can identify the highest-value problems, and know how to orchestrate AI-native workflows to solve them at dramatically greater speed. The differentiator won’t simply be who has access to the best AI, but who has the people who know where and how to apply it.”

Most CIOs know this, so it is not surprising that this year’s State of the CIO survey found that AI tops the lists for both most sought-after skills and the hardest to find. As a result, CIOs will have to get creative to build up their teams’ AI know-how in the year ahead.

How should IT be reshaped to maximize value creation?

CIOs, including Shelia Anderson, EVP and chief information and digital officer at Unum Group, are also asking how they should shape IT as an organization to ensure they can deliver value to the business.

“That has been a big focus for us this year and will continue to move down that path in 2027,” she says.

The question has shaped Anderson’s work at Unum Group since she started in her role in May 2025. She has organized her IT department around end-to-end value streams and says that approach has helped her IT staff work more closely with the business and more readily identify where improvements can be made. She also has her IT organization working to better understand business processes, and has been working on building more capabilities to measure and manage AI costs as Unum seeks to transform using AI. And she’s looking at what investments and improvements are needed in IT’s foundation.

All these moves will enable the IT department to get better at value creation, Anderson says.

She plans to augment this work in the coming year by providing AI training and education for IT workers and employees throughout the company “so they know how they can drive value and find opportunities in their everyday use of AI.”

Ensuring she has the right mix of talent — meaning she has the right skills in the right seats to deliver value — will also shape the actions she takes as CIO in 2027, Anderson says.

“In 2025 and 2026 there was a lot of testing and learning, but 2027 will be a year of maturing. You’ll see more formal processes and settling in around how you’re going to leverage the skills in your organization to drive value,” she adds.

In Foundry’s 2026 State of the CIO survey, 83% of CIOs say their role is becoming more of a hybrid to lead both business and tech initiatives, rather than simply manage IT. As the two become more inseparable, the CIOs who see collaboration as a competitive advantage are putting daylight between them and those who see it as a cursory soft skill. And C-suites are responding as they rely more on IT to drive growth and accelerate efficiencies. So the CIOs who deliver the most value will bring collaboration skills to drive change and technical expertise.

Silos between IT and the business are falling

As the World Economic Forum notes, not every tech initiative requires deep collaboration. But as businesses move beyond routine IT investments into cross-functional, innovation-focused transformations, collaboration becomes essential.

For example, how can IT decisions help an organization open new markets faster, or emerging technologies reduce risk in areas like data privacy. It’s no surprise, then, that Deloitte found that 65% of CIOs reported directly to CEOs in 2025, up from 41% in 2015.

Because tech is now part of almost every business decision, it takes everyone’s expertise and insights — from the CIO to the CFO — to see the full picture of how an initiative will affect the company. Shared knowledge leads to better-informed decisions that give tech its biggest impact.

When silos persist, CIOs struggle to get buy-in for their initiatives. Protiviti’s Global Transformation survey found that while 61% of CIOs report being confident in their tech transformation outcomes, only 34% of CEOs and boards feel the same — and the gap is even wider for AI implementations.

That gap is, at its core, a collaboration problem. Without effective collaboration and communication, others can’t see the value of transformation efforts. CIOs who excel in these areas can more easily demonstrate value, win buy-in and keep their organizations at the forefront of technology.

Delivering speed and innovation

Collaboration goes well beyond getting buy-in. While it may be thought to slow IT down, it actually brings in the insights that make change more effective and faster.

As I’ve highlighted in the past, when organizations move away from rigid data ownership, cross-functional teams can share what drives further innovation.

The same holds across the C-suite. Perspectives from outside IT show CIOs how a tech implementation works for the company as a whole, and where new opportunities lie. That lets CIOs test, refine, and reshape solutions faster.

Most importantly, when teams share a common purpose and goals, everyone stays aligned on the best overall outcome for the business. Iteration happens much quicker because everyone is focused on the big picture rather than individual agendas. With increasingly rapid tech cycles, that speed is a competitive advantage in its own right.

Building your collaborative abilities

For CIOs used to the old way of doing things, shifting to a collaborative mindset can be a challenge. But like any capability, it can be developed and improved by applying some fundamental principles.

Earn trust through follow-through. Trust isn’t built at a kickoff meeting or with a mission statement. It’s built when people see you do what you said you would, again and again. For CIOs, that means delivering on commitments or flagging early when you can’t. Clear and consistent communication is one of the most reliable ways to build trust across an organization.

Make transparency the default, especially when news is bad. Transparency means more than status reports. It means being candid about costs, risks, trade-offs, and setbacks before they surprise anyone. When leaders sidestep difficult discussions, trust erodes and goals drift. And in technology programs, small unspoken issues have a way of becoming expensive ones.

Seek cultural fit in how you work, not who you are. Cultural fit isn’t about sameness or shared backgrounds. It’s about compatibility in management philosophy, values, decision-making, and communication. A CIO who values speed and experimentation, and a CFO who prioritizes predictability and control will clash unless both recognize the difference and agree on how to bridge it. The same is true when it comes to finding the best technology providers.

Focus on the what, not the how. Few things undermine collaboration faster than micromanagement. When leaders dictate how a colleague or partner should work, they signal distrust, stifle creativity and innovation, and tie up their own time policing activities instead of driving results. The better approach is to agree on clearly defined, measurable outcomes, and then step back. Letting go of the how isn’t giving up control but building enough trust that everyone gets there together.

This requires adjusting your workflows and skills, but the payoff is valuable insight from your collaborators and wins that lift the entire organization.

The new life of the CIO

The days when the CIO could simply be the tech expert behind the scenes are long gone. With AI and other technologies ingrained in every aspect of the business, the CIO must lead well beyond IT.

By making collaboration a priority, you can gain critical insights across teams to guide your tech strategy, and deliver the greatest value to your organization.

With hundreds of AI-related companies increasingly crowding the market for enterprise dollars, IT leaders should prepare for the day when at least one of their AI vendors suddenly shutters or gets acquired.

While the AI market overall is showing few signs of slowing down, a shakeout will eventually happen — with potentially significant ramifications for CIOs. IT leaders only need to be reminded of Broadcom’s acquisition of virtualization giant VMware to visualize what could be at stake.

CIOs, of course, are not unfamiliar with the many issues that arise when an IT vendor shuts down or is acquired. However, suddenly losing an AI vendor creates some special challenges, according to industry experts, including data access and privacy concerns and the loss of AI models trained for specific customer needs.

“Assume your AI vendor will get bought or go under, and plan for it now,” says Dave Shanker, cofounder and managing partner at IT consulting firm STS Consulting Group. “Hundreds of these companies launched in three years, and most won’t last. The ones that do will get acquired, and then pricing, terms, and roadmaps can change overnight.”

Set your exit strategy up front

For now, the demand for AI products seems to be moving ahead at full steam. In May, Gartner projected that the worldwide AI market would grow by 47% in 2026, totaling nearly $2.6 trillion.

But many insiders, including Amazon founder and CEO Jeff Bezos and OpenAI CEO Sam Altman, have projected that the AI bubble will eventually burst, with many AI startups buried in the rubble.

One warning sign came in mid-October, when consumer-focused Character.AI filed for bankruptcy, with the startup agreeing to sell some of its assets to Disney for $20 million.

With many vendors likely to go extinct in a few years, it’s incumbent on CIOs to plan ahead for potential market disruptions, Shanker says.

He recommends that IT leaders ensure that their organizations own the data they’re putting into AI models. The organization’s prompts, outputs, fine-tuning data, and anything it has loaded into an AI tool should be exportable in a usable format.

In addition, IT leaders should write the potential exit from an AI vendor into the contract from day one, with notice periods, data return and deletion terms, and the right to leave without a penalty if the vendor is acquired built into the deal.

“Procurement should build this into every contract, AI vendor or not,” he adds. “If they aren’t, that’s a big risk they’re missing.”

IT leaders should also weigh portability against cost, he says. “Sometimes it makes sense to build tightly on one platform, because adding a layer that keeps you portable can double or triple what it costs to run,” he explains. “Redo the analysis at least once a year. Something that’s easy to move today can, after three years of heavy development, leave you stuck.”

AI-specific concerns complicate traditional vendor risk

Rodion Sorokin, CTO of deep-tech venture studio The Capital Index, agrees that CIOs need to prepare for an eventual market shakeup. Sorokin points to information from company shutdown assistance provider SimpleClosure, which noted that AI companies are currently a small and shrinking percentage of the types of closures they handle.

However, the median age at closure is lower for AI companies: 2.30 years versus 3.78 for other companies, according to SimpleClosure. 

“We hypothesize that the 2023-2025 AI cohort is still in its early days, and we’re only beginning to see how it plays out,” SimpleClosure CEO Dori Yona wrote in a blog post,

Even if the early warning signs are few and far between, CIOs should be prepared to quickly switch vendors, Sorokin says. Not only should IT leaders be able to export prompts, outputs, logs, fine-tuning data, and other information in usable formats as STS Consulting’s Shanker advises, but they should also test those processes before they sign a contract with any AI vendor, he recommends.

“Assume every AI vendor will be acquired or shut down within your contract term, and negotiate for that day,” he says.

CIOs should also be aware of specific concerns that pertain to the sudden loss of an AI vendor, Sorokin says. For example, their organization’s data and customizations may live within the vendor’s product, and it’s possible all that gets deleted if the vendor goes under or is acquired, he notes.

In addition, switching AI tools can render AI-aided workflows suddenly inoperable, he says.

“Switching isn’t a migration; it’s a revalidation,” Sorokin says. “A new model behaves differently on the same prompts, so every workflow has to be retested. Without your own evaluation suite, you can’t switch safely.”

CIOs should also know that notice can be measured in days, he adds. In June 2025, Anthropic cut Windsurf’s direct access to its Claude models with less than five days of notice, amid reports of an OpenAI acquisition of the smaller company.

While some observers say there may be safety in going with the big players in the AI space, others suggest startups may offer services that aren’t available from the big boys. In addition, large AI vendors have also cancelled services on short notice, Sorokin notes, with OpenAI shutting down its Sora app in April and the Sora API in September.

“Size helps, but it isn’t continuity,” he says. “Choose portability over logos: an open-weight model you can run yourself as a fallback, and contracts that guarantee export.”

IT leaders exploring options

While signs of a busting bubble are hard to find, many IT leaders are preparing for the future, says Colin Woods, CEO and co-founder of high-performance infrastructure provider FreedomTechSolutions.

A growing number of IT leaders are beginning to reassess the level of dependency they are willing to place on AI platforms and cloud-hosted services, he notes.

“From our perspective, concerns around AI vendor consolidation are already influencing customer decision-making,” Woods explains. “While the rapid growth of the AI market has driven innovation, it has also created a crowded ecosystem of providers, many of whom face uncertain long-term viability.”

In some cases, IT leaders are moving away from hyperscalers for certain AI workloads, with a renewed focus on infrastructure ownership, data sovereignty, and operational control, he says.

“Organizations are increasingly concerned about where their data resides, how it is processed, and what happens if a vendor is acquired, changes its commercial model, or discontinues a service altogether,” he adds. “For many businesses, AI is becoming too strategic a capability to be entirely dependent on an external provider whose roadmap they do not control.”

Some organizations are deploying smaller, dedicated AI environments in regional or edge data centers, Woods says. “These AI nodes allow organizations to run inference and specialized workloads closer to their data, while retaining full governance over models, datasets, and infrastructure,” he adds.

There’s no point building your own agentic AI stack for a non-core function when a vendor already has a product ready to go that meets your needs. For non-tech companies in particular, building from scratch can take time, money, and expertise they don’t have.

Take for example, Fertitta Entertainment, with more than 600 Landry’s restaurants and the Golden Nugget hotel and casino. After the pandemic, the company had problems adequately staffing its call centers, says its director of contact center operations Brian Jeppesen. “The agents didn’t want to come back,” he says.

The virtual assistants didn’t work because the user experience was bad, and with the incoming call volume, 40 to 50% of them were abandoned.

“Everybody wants AI, but nobody wants BS,” Jeppesen says. “I talked to my CFO. This is what we’re abandoning. This is the potential revenue we’re losing. I can’t throw bodies at it anymore.”

Agentic AI has upped their game for the virtual assistants. Instead of only answering simple questions like check-in hours, it allows the company to connect APIs and carry out more complicated tasks, like making reservations.

“My abandonments fell to under 10%,” says Jeppesen. “From waiting 45 minutes to an hour, they can get their questions immediately answered, or immediately transferred and taken care of.”

And turnover went down. Call centers typically see 100% turnover, says Jeppesen. Not because everyone quits, but because the bottom 25% quit and need to be replaced over and over again.

Now that bottom tier doesn’t have to be rehired, while the people who stay are the most knowledgeable. “They give a better experience for my customers,” he says. “And it reduced my attrition by 50%.”

Forrester did a total economic impact study for the specific vendor that Fertitta used, PolyAI, and the ROI came back at 391% over the course of three years, with a payback period of less than six months.

Jeppesen says he reviewed his own numbers and found that the automated reservations system alone brings in 3,000 reservations a month. “That’s $600,000 in revenue a month for us without the need to speak to an agent,” he says.

Next, Fertitta plans to expand it to handle to-go orders and web chats. There are also opportunities for agentic AI on the technical help desk, Jeppesen says.

Agentic AI on the menu

Fogo de Chão, a Brazilian restaurant chain with around 120 locations, is also deploying agentic AI in small, focused use cases, and sees ROI with some of the projects.

For instance, it deployed virtual call center agents from PolyAI, and saw positive ROI from day one since there were no significant up-front costs or usage-based pricing. After several months of testing and small-scale pilots in 2025, the system has been rolled out across the US, with 100% of calls now being answered instead of up to 25% being dropped, and a 10% increase in total bookings that wouldn’t have happened otherwise, says Michael Breed, the chain’s chief marketing officer.

“If we compared it to a media channel, it’d be twice the ROI from other media channels,” he says. “It’s almost like another customer acquisition vehicle.”

The company is also building agentic workflows for its team members.

“We’ve identified jobs that are very repeatable and time-consuming, and we’re training agents to take those tasks,” he says. These are low-risk jobs, like proofing menus or other text, or replacing manual processes, such as pulling data from one system into another that had no automatic integrations.

“The agent does a great job bringing in the data, then the human in the loop can inspect it,” says Breed. “Is it getting classified right? Is it getting to the right places? And when guests are dissatisfied, can we make it better?”

Built for SMBs

For smaller companies in particular, off-the-shelf agents make the most sense. AltaClaro, for example, builds training simulations for lawyers, and counts more than 100 of the top 200 US law firms as customers, all with just 10 employees.

One product is a deposition simulator, says CEO Abdi Shayesteh, a completely agentic, voice-based platform.

“It’s a fully simulated environment that creates the experience of doing the deposition,” he says. “Different agents play the role of the witness, opposing counsel, and the court reporter.”

He says it’s kind of like a Zoom call. The product is built on top of technology from Verbit, a legal tech company with a custom agentic platform.

“They’re taking care of the technical infrastructure and we have the material to ground the AI.”

A different vendor, Credal, provides a secure AI platform that allows AltaClaro to switch between different models. It can obfuscate names and keep other sensitive data from leaking out, Shayesteh says. “It ensures everything is anonymized and does secure calls,” he says. “We standardized on Credal because it’s highly flexible and secure. We serve law firms and we’re as invested in security as they are.”

But security isn’t the only issue that enterprises face when deploying agentic AI.

The hidden costs of buying AI

The temptation to jump into agentic AI is high, says Dana Simberkoff, chief risk, privacy, and information security officer at AvePoint Inc., a data security company. And off-the-shelf agentic AI can be so easy to adopt that companies deploy it before they’re prepared for it.

Enterprise vendors will often add the technology at no extra cost, just to get people using it, and companies turn it on for their employees before truly understanding what it does, and before systems and workflows are ready. “Then it immediately exposes problems that were already there,” she says.

Say, for example, an agent helps with research, normally considered a low-risk task.

“AI will follow sharing links across documents,” she says. “So if you haven’t taught your users how to properly use sharing links, the AI agent is just going to follow that path, and it could expose information that might have been obscured in the past. It’ll gobble up everything it can.”

The result is it could expose HR data, personal information, or proprietary information to people who shouldn’t have access to it. One company she knows had to shut down their AI project immediately and go back, classify, and label all their information, put governance controls in place, and teach people how to apply metadata.

According to Gartner, by next year, 40% of enterprises will demote or decommission autonomous AI agents due to governance problems that only show up after they’ve gone into production.

Token management

Another hidden cost of off-the-shelf AI? Token use. When agentic AI is embedded into a system that a company is already using, then the cost of switching it on can be as low as zero. But someone has to pay for all that token use.

When agentic AI comes out of pilots and goes into production, usage can change dramatically, and token costs can quickly spiral out of control. It doesn’t help that agentic AI use typically requires much more tokens than AI chatbots.

“A lot of the embedded AI should be utilized,” says Doug Vargo, head of the national AI and alliances team at consulting firm CGI. “If big parts of your business processes are living and breathing in ServiceNow, Salesforce, or another application, you should leverage that, but you need to understand the cost.”

This is playing out in the marketplace right now, he adds, as underlying model providers change their token costs.

“That changes the economics of the decisions,” he says. “And six months from now, it’s a different discussion, because they’re going to continue to move the needle.”

The employer’s prerogative

Another hidden cost that companies might miss is they need to put training and support systems in place for employees. Many off-the-shelf agentic AI tools might seem easy to use, but actual adoption still requires change management. CGI did offer training to employees when it rolled out GitHub Copilot, a popular coding platform with agentic functionality. “And we expected them to adopt and leverage it on day one,” says Vargo. But people quickly went back to their old ways of working, he says, and CGI had to pivot its strategy.

“It’s going from training programs to habit formation or transformation programs,” he adds. “And we’ve seen adoption skyrocket.”

The change isn’t in just how individuals adopt AI to improve their own work, though. “You’re now a manager of 10 to 20 specialized agents that are driving that work for you,” he says.

The error is when companies put the technology first. “They fall in love with the tech or a vendor, but they’re not really thinking about how they’re truly integrating these tools and technologies into the business.”

Embracing change

Change management is a particular challenge with agentic AI because the tools are changing at such a fast pace.

“It’s technology that’s constantly changed almost every month, or maybe every week,” says Scott Likens, chief AI engineer at PricewaterhouseCoopers. “The challenge I see is the timescale of change is much faster than what we’re used to.”

We’re still in the early stages of agentic AI, he adds, and waiting isn’t a good strategy. “You have to start, but keep flexibility over longevity in mind,” he says.

For Luis Huerta, group head of gen AI and agentic automation at International Airlines Group (IAG), change management is one of the most complex parts of agentic AI adoption, as it uses agentic AI to improve customer experience, operations, and efficiency.

“The technology is powerful, but the bigger shift is organizational,” he says. So the company isn’t just focusing on building agents, but also on business ownership, adoption, and governance, and helping colleagues understand how these capabilities change the way work gets done and how it can support them, rather than replace them.

Huerta says IAG also uses agentic capabilities embedded in enterprise platforms, but also builds its own agents, too.

They’re not alone. According to Anthropic’s 2026 State of AI Agents report, 47% of organizations take a hybrid approach to AI agents, while 21% rely entirely on pre-built agents, and 20% build their own.

“We take a technology-agnostic approach,” says Huerta. “Our priority is to select the right solution for each use case.”

Longtime IT leader Matt Ausman envisions a future where developers may not have any backlog because AI will let them create new capabilities as quickly as they imagine them.

“The backlog will vanish because you will have infinite capacity,” says Ausman, SVP and CIO of Zebra Technologies.

He wonders whether business users will be able to keep up with the training required in a world where the software they use will change that rapidly. He also wonders whether IT workers are ready for that scenario: Will they be able to push back on bad ideas, or just build new capabilities simply because they can do so as rapidly as they’re proposed?

“Today we have to prioritize because we don’t have unlimited capacity. But in the future will they stop to ask, ‘Are we building the right things?’” he says, adding that “because someone thought it up, doesn’t mean it’s worth doing.”

Ausman acknowledges that such a reality is at least several years out, but he already sees AI changing how his IT team works, the work they need to do, and the skills required to do it.

For example, the average size of Scrum teams is shrinking, dropping from eight people to five or six who are producing more thanks to AI. Ausman believes Scrum teams may eventually be made up of two members: a functional product owner and a technical product owner who together use AI to gather requirements, generate mockups, create code, and so on, spending the bulk of their time supervising agents tasked with ensuring the architecture is right, bugs are fixed, security is in place, and the like.

Ausman also sees IT teams of the future producing work faster and drawing more from their business domain knowledge and less from their deep technical skills as they use AI to remake processes and workflows throughout the organization as well as in IT.

A shift in skills and mentality

These shifts will change not only how IT teams are structured but also the skills IT workers need, Ausman and other IT leaders and experts say.

True expertise in business domains, once merely a differentiator, is fast becoming a must-have for IT pros, Ausman notes. So is discernment (the ability to recognize when ideas aren’t worth pursuing), conviction (the ability to tell others that those ideas won’t produce benefits), and communication (the ability to be clear, concise, and diplomatic when telling others as much).

“They’ll need to be good partners to [the business] and be able to ask, ‘Are you sure we should do this thing?’” Ausman explains.

Of course, IT workers will continue to need technical skills, Ausman says, noting that most of his hiring today is for entry-level positions that draw in AI natives who help spread their AI knowledge throughout the enterprise. Still, he questions whether earning a computer science degree and knowing multiple programming languages will be as valuable 10 years into the future as they were a decade ago.

Problem-solving skills will be more highly prized in IT workers instead, he says. Emotional intelligence, authentic leadership, decision-making capabilities, and prioritization skills will also be in high demand.

“I think about my team and the fact that we can never outsource team care, the ability to have difficult conversations, emotional intelligence, [provide] psychological safety. Those things can never be automated,” he adds.

‘Hyperacceleration of learning needs’

CIOs across all industries are beginning to adjust their IT organizations in response to these evolving and anticipated changes as they contemplate the future of their IT workforces.

In a poll of CIOs at this year’s CIO 100 Awards and Conference in August, “talent, skills, and organizational change” topped the list of issues worrying CIOs, with 33% citing it. That put it ahead of “demonstrating AI business value and ROI” (27%) and “budget pressures and competing priorities” (21%).

When asked which capability will define the next generation of CIOs, 29% said “building organizational talent” — just behind “leading AI transformation” (30%) and ahead of “developing business strategy” (22%).

AI is obviously a factor here, as its use expands to automate more core IT tasks, but AI is only one of the dynamics driving change in IT. Demographics, market conditions, industry trends, and economics have all put pressure on organizations to be more responsive through the use of technology, says James Stanger, chief technology evangelist at CompTIA, an IT training and certification association.

Even as IT leaders feel the pressure of those dynamics, CIOs must also be ready to capitalize or respond to evolving technology as rapidly as it hits the market, Stanger adds. That puts further pressure on CIOs to rework their IT departments so that they’re able to meet such expectations.

Leading CIOs “have realized that technology is changing so utterly fast and because of that they need to change their organizations,” Stanger adds. “We’re finding that the traditional ways of doing things are being revised.”

Although most CIOs aren’t making radical changes yet, they are adjusting their approaches and the structure of their IT departments, Stanger notes.

CIOs are implementing more training initiatives and requirements to ensure their teams are better keeping pace with the changing tech landscape, Stanger says, noting that many CIOs have shifted from seeing training as a retention tool or a nice-to-have to something essential for enterprise survival.

And that commitment to learning is increasingly being incorporated into the day to day. “Microlearning is becoming more of a thing; there’s more of a steady stream,” he observes, adding that microlearning is beneficial because it can happen as rapidly as technology changes and accelerates hands-on retention of what is learned.

CIOs are also enlisting more talent from the business side to help develop technology solutions. These aren’t the citizen developers of yesterday, Stanger says, as they now often report via dotted lines to IT managers, allowing CIOs “to get a really good sense of what’s going on in those departments and what’s working and what is not.”

Conal Gallagher, CIO and CISO of Flexera, is leaning heavily into training to ensure his staffers have the skills necessary for the software maker to thrive.

AI fluency and AI use is at the top of Gallagher’s training list, as is AI administration and governance so that his staff can create AI tools that workers in other departments can safely use. And he’s working to develop more agentic AI skills among his workforce.

Like Stanger, Gallagher calls out the “hyperacceleration of learning needs” today, saying that most IT professionals are learning as they go, sometimes turning to online videos for instantaneous education.

The fast pace of technology evolution, however, makes it difficult to predict what will be needed from his IT workers a year from now. As a result, longer-term coursework is becoming less valuable.

“Traditional training as we know it probably doesn’t apply anymore,” he says, adding that the goal now may be to “just keep pace with the current.”

Reinventing roles to maximize IT capacity

Info-Tech Research Group’s 2026 IT Talent Trends report shows that IT leaders and IT pros alike expect nothing short of wholesale change for IT ahead. Nearly nine of 10 (89%) IT leaders anticipate restructuring their IT organization, according to the survey, and 95% of IT professionals acknowledge major skill changes would be required to keep pace with technology by 2030.

To help bridge IT to the future, CIOs are adding more service owners (the modern version of product owners) and business relationship managers, says Brittany Lutes, research director in Info-Tech’s CIO practice. IT leaders view these roles as essential for delivering AI and other IT capabilities that create value for the business.

“If they didn’t have them, they’re adding them,” she says. “If they had these roles before, they have more now.”

CIOs are also creating broader roles that enlist staff to operate in multiple domains, Lutes says. A networking engineer may now be expected to work in database administration and database engineering, for example.

CIOs want their teams to be more flexible and more modular, so they can move personnel more easily to where work needs to be done. To achieve that, some CIOs are grouping workers in pools or centers of excellence and then assigning them out as needed.

Broadening IT roles also helps keep those positions working at full capacity as AI picks up more of the manual tasks previously assigned to IT pros working as specialists.

“CIOs are not losing headcount, but the way they’re your utilizing that headcount is changing,” Lutes says, pointing out that these dynamics are changing job titles, too. For example, one IT organization Lutes knows of changed the all its IT workers’ titles to simply “engineer.”

Leading CIOs are also creating new roles centered around AI implementation and management, Lutes notes. These new roles include decision intelligence lead; forward-deployed engineer; human-AI interaction lead; knowledge engineer; master data manager; memory engineer; MLOps engineer; prompt engineer; synthetic data engineer; and trust engineer.

The seemingly contradictory trends reflect a profession in flux.

“We’re still in a messy phase as an industry,” Lutes says. “We’re not going to get to that major shift in the next two years, but five years out we may see a complete restructuring of the IT organization.”

Preparing for the future state

Nick Nadgauda, global CIO of MetLife, is preparing for that future state.

“We continually evaluate whether we have the right blend of engineering depth, business understanding, adaptability, and leadership,” he says. “The goal is not simply to fill today’s roles; it is to build a workforce that can learn, unlearn, and relearn as the work changes. That includes building a strong early-career pipeline and giving emerging technologists the technical foundations, mentoring, and hands-on experience to grow with the organization.”

This ongoing evaluation has reinforced three priorities at MetLife, Nadgauda says.

“First, AI fluency needs to become a broad workforce capability, not a specialty confined to a small group. Second, strong fundamentals in software engineering, cloud, cybersecurity, data, and architecture remain essential. We’re putting an emphasis on making our organization more technical. Third, we need teams organized around clearer accountability and closer alignment to business outcomes,” he says.

Nadgauda is investing in continuous role-based learning, creating more opportunities for hands-on learning, and bringing critical expertise in-house.

But he also acknowledges a degree of uncertainty about what’s ahead for IT, saying, “We can predict the direction more confidently than the exact destination.”

He adds, “We know AI will become a natural part of how work gets done across engineering, operations, and customer-facing teams. We also know that judgment, creativity, accountability, and deep domain expertise will remain critical. What is less certain is how quickly individual tasks and roles will change. That is why adaptability matters so much. The best hedge against uncertainty is a workforce that can learn continuously, experiment responsibly, and apply new capabilities to real business problems.”

According to a quartet of Accenture intellectual heavyweights, what CIOs need now to account for and control AI costs is tokenomics, sweet tokenomics.

The idea of linking AI spend to business outcomes may be a noble one, but “Tokenomics” only pretends to do so, unless you think tallying costs and measuring business outcomes are one and the same thing.

But before we get to Tokenomics we need to set the stage.

Start with Function Point Analysis (FPA), a venerable and mostly forgotten discipline, which should say something about its efficacy.

The folly of cost metrics you can’t understand

Back in the day, IT leaders worried that boards of directors worried that business executives worried about the cost of information technology and how to manage it.

IBM introduced Function Point Analysis in 1979 to solve this issue by breaking a software application into chunks — function points. Analysts could then sum these function points to compare the application’s overall complexity — and therefore its cost and risks — with that of any other piece of software.

When presented with the number of function points a proposed software project would require, your average, hopelessly naïve business leader asked IT to point to a few of the project’s function points, expecting mocked-up screenshots, sample reports, and so on.

Sadly, IT had to decline because, the design analysts would explain, a function point is (as Google’s AI Overview explains) “an abstract, synthetic unit of measurement rather than a physical object or a single piece of code.”

The Function Point Analysis value proposition, then, was that business leaders could manage IT costs and risks better by relying on cost metrics they couldn’t understand.

Function points for the AI era

Fast-forward to today, and the equivalent function-point cost metrics for AI are tokens — the chunks of text that a large language model reads, processes, and generates. Which, in case you missed the briefing, supposedly makes tokenomics a way to manage the costs and risks of AI.

As the aforementioned Accenture gurus “explain” (scare quotes intended): “Enterprise AI is creating a new cost of doing business: tokens. As consumption scales, token economics is becoming a material risk and opportunity, drawing increasing attention across the C-suite and the board.”

Which raises the question: Where are our tokenomics pundits finding C-suite and board members who give even an infinitesimal damn about tokens? They’re probably the ones who still pay attention to function points, too.

A short lesson in Brickonomics

But before we delve deeper into Tokenomics, capital T, we need to understand building maintenance and the critical role Brickonomics plays in it.

It’s like this: Brick walls are made of bricks. So if a brick wall needs attention, a building management project estimator should need to understand the costs and risks associated with managing a standard brick in the crumbling wall they’re supposed to fix. Multiply the standard management cost of a standard brick (a brickon?) by the number of bricks in the wall, and voilà! Building management knows …

Well, it knows less than it would have if had it instead invested in a conventional tuckpointing analysis. Among the missing pieces: In tuckpointing analysis, what usually needs and gets attention is the mortar between the bricks, not the bricks themselves.

Putting it all together — because you will have to

Now we’re ready for Tokenomics, and its conjecture that dividing an LLM into tokens and assigning a cost to each of those tokens will prove more useful than dividing software into function points or a wall into individually managed bricks.

The next questionable aspect of Tokenomics is that it depends on the proposition that all AI implementations are built on large language models or similar architectures.

To be fair, Tokenomics doesn’t have to be limited to analyzing LLMs. It can be adapted to any collection of thingies that’s composed of chunks that can be parameterized.

But that really isn’t good enough, is it? Because given the current pace of AI innovation, assuming that anything about it is stable, especially that its future core architecture will be tokenizable, is a shaky assumption at best.

It’s a mortar versus bricks challenge.

Then there are those pesky function points. Just because AI is involved in an implementation, that doesn’t mean software isn’t also involved. Which would seem to suggest that Tokenomics will have to integrate Function Point Analysis into its methods and practices as well.

Imagine you’re the one who has to explain this mess to those boards and business leaders who are, supposedly, clamoring for it. Because, as CIO, you will be.

Chargebacks and culture

We haven’t even gotten to the other big risk of Tokenomics: that it focuses decision-makers on AI’s costs, with only token attention — no pun intended — paid to its benefits.

Deep down inside, Tokenomics’ value depends on its use for chargebacks and showbacks. The logic for imposing these is superficially appealing — it makes business owners take responsibility for the AI spend driven by their consumption.

Those CIOs who dove into the chargeback rabbit hole discovered just how futile chargebacks can be. That’s because of a time-honored practice updated to the digital age: arguing about the bill.

Because that’s all a chargeback is: a bill. It’s a statement of how much money will change hands so that someone who has provided a good or service is reimbursed for the value they’ve created.

By the numbers: (1) a chargeback is a bill; (2) the point of Tokenomics is to support chargebacks; and (3) no business leader will be able to understand the bill they’re supposed to pay given Tokenomics impenetrable vocabulary and construction.

But what about showbacks, then? Letting business leaders understand the costs they’re driving would seem harmless and potentially useful.

And in fact showbacks can be harmless and useful, so long as they’re applied in the right business culture. That’s a culture in which managers consider themselves to be leaders of the whole company, basing their AI utilization decisions just as they base their other decisions — on the extent to which they benefit the whole company.

With a culture like this, IT can limit its application of tokenomics, little t, to its technical architecture practices. This way, the CIO can inform business managers of AI’s costs just as the CIO informs business managers of the other costs of running IT.

Where, the business manager might ask, did these cost numbers come from?

“You have two choices,” the CIO might reply. “You can trust me, or I can explain it to you.”

Now there’s a threat worth paying attention to.

Every generation of enterprise computing introduces new abstractions for scaling systems.

We began in the scale-up era, deploying ever-larger servers. Then came the scale-out era, where containers and Kubernetes transformed infrastructure into orchestrated fleets of commodity systems. Today, we are transitioning to agentic-scale.

This is not an incremental tweak to the infrastructure playbook; it is a fundamental architectural evolution that rivals that same major shift we saw from scale-up to scale-out. In the age of agents, many of the assumptions that defined the scale-out era no longer apply.

Applications are no longer just for human operators

Traditional enterprise applications were built around human behavior. People logged into systems during business hours, establishing predictable infrastructure usage patterns. AI agents operate around the clock and traverse resources in nondeterministic ways.

Ask an agent to onboard a supplier, investigate a security alert, or resolve an insurance claim, and that request cascades into autonomous decisions. The agent invokes APIs, consults internal systems, delegates tasks to specialized sub-agents, and iterates before returning a recommendation or taking action.
Where scale-up capacity planning revolved around forecasting user traffic, agent-driven volatility will become the new “peak usage” scenario for infrastructure.

The urgency of fluid compute and AI sandboxing

Autonomous agents can spawn clusters of computing activity at any moment, but keeping idle pools of resources on standby is financially reckless.

Just as containers became the unit of application deployment in the cloud era, isolated execution environments are emerging as the operational unit for agents. Serving as secure sandboxes, these environments can be created and retired on demand, allowing enterprises to absorb unpredictable bursts of agent activity and execute untrusted code safely without maintaining costly idle capacity.

Moving intelligence, not data

Agents only create value if they can operate across your enterprise estate, applications, and data. Historically, enterprises centralized data so applications could find it. Agentic systems reverse that assumption by bringing intelligence closer to where data already resides.

Unifying the data layer for agentic AI success requires a semantic mapping of the enterprise’s data assets, also known as a “knowledge catalog.” Agents must be able to instantly discover and query unstructured data across hybrid environments, which is why data infrastructure is evolving to support “smart storage” and systems that automatically tag objects with metadata for this new era of agentic discovery and execution.

The scale-up era taught enterprises how to optimize machines. The scale-out era taught them how to coordinate fleets of machines. Agentic-scale optimizes orchestration itself. Organizations must decide where intelligence executes, when environments are created and retired, how agents gain access to data and tools, and how all of it comes together across clouds, on-premises environments, and the edge without forcing every system into a single repository.

The new trust model for agentic systems

Today’s rapid build-out and agent sprawl has major security implications for CIOs.

Because developers can easily spin up agents using open-source frameworks and unvetted repositories, organizations risk introducing new attack surfaces into production. Infrastructure must treat agents as digital personas, enforcing centralized identity, policy-based access controls, and clear governance over what each agent can access and what it can do.

That means trust now extends beyond security to sovereignty. For critical infrastructure, defense, and other regulated industries, that means deploying AI in disconnected and fully air-gapped environments while protecting data at rest, in transit, and in memory. One straightforward architectural goal for CIOs should be to bring AI capabilities to wherever sensitive data resides, rather than forcing sensitive data to move to wherever the AI happens to be.

Preparing for the era of more agents doing more things

Organizations should assume they will have more agents doing more work across more systems.

As autonomous agents operate around the clock and execute highly unpredictable, non-linear workflows, traditional capacity planning must give way to continuous adaptation. Navigating this shift requires an integrated infrastructure stack. CIOs must pair elastic compute tailored for agentic workloads with a unified data layer that unlocks enterprise knowledge. This foundation must then be secured by modern, policy-driven networking and strict data sovereignty controls wherever sensitive workloads run.

The cloud era was won by coordinating fleets of machines. The next era will be won by orchestrating intelligence.

To read more about this topic, explore Google Cloud infrastructure solutions on our blog.

CIOs experiencing sticker shock from recent software bills should be questioning how they got there.

That’s because IT leaders at large enterprises are paying a huge so-called software tax for vendor-driven upgrades, migrations, and licensing changes they would not have chosen for their organizations.

About 28% of software budgets at large organizations — an average of nearly $12.9 million per year — are being eaten up by vendor-pushed initiatives, according to a survey commissioned by independent software maintenance partner Origina.

Ninety-five percent of IT leaders surveyed said their organizations have paid for a vendor-driven upgrade in the past three years that they believed would actually reduce performance, according to the survey, which covered US and UK organizations with annual revenues higher than $1 billion or with more than 10,000 employees.

The sheer size of this “software tax” surprised Tomás O’Leary, Origina’s CEO, even though the company expected to see substantial costs for vendor-driven upgrades.

“Software was supposed to make businesses better, and it does,” he says. “But it’s also imposing a hidden cost on companies, quite a significant one. I’m quite taken aback because we know this exists, but we never quantified it before.”

When vendors control your IT spend

Big vendors have several major advantages over software customers, with the cost of switching or refusing to upgrade topping the list, O’Leary notes. In many cases, CIOs are faced with a loss of support if they turn down vendor-driven upgrades, leaving them with little room to refuse, he says. CIOs dependent on VMware are familiar with this squeeze play, as are those facing SAP deadlines.

“CIOs are left in a situation where the alternative may be worse, so upgrading is the lesser of two evils,” O’Leary adds. “When people buy technology, they don’t realize how long this stuff is going to hang around. Nobody realized this thing was going to be embedded, and the dependency was going to be greater and greater with time.”

In addition, software vendors have an experience advantage when negotiating contracts, O’Leary says. CIOs and procurement officers may try to negotiate favorable terms for upgrades, but big software vendors are expert at writing contracts that give them an edge, he notes, adding that the survey results suggest IT leaders’ current approach to negotiating better contracts is a failed tactic.

“If you keep trying the same thing over and over again, and it doesn’t work, you’ve got to try something different,” O’Leary says. “There’s a relief on the buyer side when the negotiations are over, and on the seller side for sure. But the seller side has to get back up again and do it again the next morning with somebody else, and the buyers don’t.”

Omar Amassine, founder and technical director at TripleW Digital, says between 20% and 30% of the IT services firm’s software and maintenance budget goes to vendor-driven migrations he didn’t request.

The biggest culprits? CRM, marketing automation suites, and cloud infrastructure, he says.

“Vendors frequently bundle mandatory AI features or restructure seat-based pricing into usage tiers, driving effective 20% to 35% price jumps with zero incremental utility for the core workflow,” he says.

IT leaders can fight back, however, by using modular, open standards and headless architecture, he adds. “The less proprietary logic locked inside a vendor’s walled garden, the higher your leverage when they mandate expensive upgrades,” Amassine says.

TripleW Digital also runs migration feasibility studies six months before software contract renewals. Having a migration roadmap can bring the vendor back to the negotiating table offering multi-year price freezes, he notes.

Emphasize partnership — and a credible threat of defection

Zachary Lewis, CIO of University of Health Sciences and Pharmacy in St. Louis, hasn’t seen the same software spending mandates as the organizations reflected in the survey.

The university has negotiated many of its software contracts to ensure upgrades and security patches are handled by the vendor, either with Lewis’ approval or during an agreed-on maintenance window, he says.

Even without costly upgrades, there’s still a time tax, he adds.

“With the sheer volume of new vulnerabilities being discovered, we’re spending more time following up with vendors to make sure they’re applying patches in a timely manner,” Lewis says. “My team simply isn’t large enough to manually manage patching across every application and software component we use.”

Here, establishing a vendor management strategy that emphasizes partnership over provisioning can help. And if vendors still drive costly upgrades, IT leaders can always switch providers.

But Lewis cautions that, as tempting as that is, it’s often a complicated process.

“Switching vendors is always an option, but it’s rarely simple,” he says. “Depending on the application, you may have years of integrations, processes, and institutional knowledge built around it. Sometimes the cost of moving away from a vendor is significantly higher than the cost you’re trying to avoid.”

Customers of large ERP solutions in particular are facing a significant reckoning on decades of customizations.

Collective bargaining power comes with communication

To regain their purchasing power, Origina’s O’Leary advises CIOs and procurement officers to negotiate shorter contracts, which would enable them to renegotiate terms more often and exit more quickly if their vendor jacks up prices.

“You have to get to a point where you can have a more frequent engagement with the vendor,” he says. “Negotiate your exit, and negotiate more frequently.”

While short-term contracts may impose more up-front costs on customers, they can pay off with the additional flexibility they provide, he says.

CIOs can also turn to third-party software management firms or consider internal support of their software packages if they have the resources, O’Leary advises.

CIOs should also share their software purchasing success stories and nightmares with each other more than they do now, he says.

“We all benefit from having a little bit more collective bargaining power,” he adds. “CIOs just need to take that leap of faith that others have the same problems, and maybe we can have some sort of collective pushback on this.”

Organizations are rapidly implementing AI, but employees often feel left behind about how AI is changing their roles. While AI can help get work done faster, it won’t necessarily help produce better quality work if employees aren’t confident in their AI skills.

In fact, a recent survey from workforce analytics firm Visier reveals that 48% of employees report exaggerating their AI expertise at least some of the time, and 45% said they often feel pressure at work to use AI despite not feeling confident with the tools.

“The risk of underinvesting in training and upskilling is that the alternative can be much more expensive,” says Andrea Derler,  principal researcher at Visier. “Layoffs, rehiring, and competing for top AI talent all come with significant direct and indirect costs. Investing in the workforce organizations already have can be an important part of making AI adoption sustainable.”

When the pace of AI implementation is rushed, however, employees often default to AI compliance, meaning they’ll use AI tools as they’re rolled out, even if they haven’t received the proper guidance or training from their organization to make the most of the investment. So for companies investing millions into an AI future, adds Derler, it’s critical to ensure employee participation is part of the bigger picture to avoid performative AI usage.

Make the most of AI investments with training

Derler acknowledges that AI adoption goals vary widely from company to company. Some will need stronger investment in upskilling and training, while others might need to invest in new resources, talent, or tools to help employees get comfortable with the new pace of work, and to source the right skills for an AI-driven workplace.

Either way, the early stages of AI adoption should involve a basic framework to assist in employee adoption, one that involves open transparency and foundational guidelines for how AI will be implemented moving forward. At bare minimum, employees should be briefed on concepts around general data security, how to spot false outputs from AI tools, and expectations around AI’s impact on their job roles.

“A few foundational skills and concepts every employee should understand include protecting sensitive data and uploading information with the greatest care, understanding the basic capabilities of approved tools available at work, and checking the work before sending AI-generated output to others,” says Derler. “Most importantly, employees need to take full responsibility for the output. The AI isn’t responsible, the person using it is.”

Avoid setting blanket goals as well, where everyone is measured simply by a certain percentage of AI usage per individual or team. This can overlook reality, she says, creating a false narrative around AI usage and adoption, and a focus on quantity rather than quality. Instead, adds Derler, leaders need to strike a balance between mandating adoption and encouraging employees to explore, discuss, and learn together where AI makes sense.  

Embracing transparency around AI

When asked if they felt their organization had a clear plan for how AI affects employees and jobs, 59% in the Visier survey said they didn’t believe or were unsure that their company had a clear plan for AI implementation. And only 28% said they felt leaders in their organizations fully understand how employees use AI today, while 33% thought leaders understand how much training and support employees need.

Overall, employees are eager to embrace AI, but when asked what would boost confidence around using it at work, respondents first ranked AI training and upskilling opportunities as most important, then greater transparency from leadership, more opportunities for career growth and advancement, followed by better communication about how AI will impact employees.

Clarity from the top

Employees have made it clear they want training and transparency from leaders about how AI will affect workflows and processes. This is especially true when it comes to career paths, as AI stands to remove certain skills and expertise deemed irrelevant.

Workers know the future will look different in the wake of AI, but they want to understand exactly what that looks like so they can prepare and shift their own expectations and career goals. Part of keeping employees engaged is ensuring they’re invested in their own future at a company, but that’s difficult if the future is particularly unclear.

“Most organizations are still learning which parts of a workflow AI can take over or augment, so creating opportunities for experimentation and learning is an important step before setting formal adoption goals or metrics,” says Derler. “Many organizations, however, are still piecing together workforce data and don’t have a complete picture of their workforce, making it harder to understand where AI can create meaningful efficiencies.”

Few CIOs are likely to suddenly decide to give an AI agent broad authority over corporate systems. The more likely path to autonomy is far less dramatic. An agent starts with a narrow task. It proves useful, so another system is connected. It then needs more data, so another permission is added. Then human approval becomes a bottleneck, so the organization removes it, and exceptions that once went to people start being handled automatically.

Each decision may be sensible in isolation. Taken together, however, they can leave an agent with considerably more authority than anyone originally envisioned.

That’s the danger of autonomy creep. And as enterprises move from AI systems that recommend actions to agents that execute them, CIOs need to govern not only what those agents were originally designed to do, but what they have gradually become capable of doing.

It starts with small permissions

Jim Taylor, president and chief product and strategy officer at identity security company RSA, sees a familiar pattern. Organizations typically create an agent for a specific purpose and give it the permissions necessary to accomplish the task. Later, the job expands.

“They just give it tasks, permissions; it just sits there and accumulates things,” he says. “Over time, it develops potentially toxic combinations. It gets too much access and can see too much data. It can do too many things.”

Humans accumulate permissions, too, of course, but organizations have developed processes around employees joining, changing roles, and eventually leaving. Agents may not have the same natural lifecycle. “They never give permissions back,” Taylor says.

Nancy Wang, CTO of 1Password, argues that CIOs should treat each expansion in an agent’s responsibilities as something more consequential than another feature. “Every time you expand the scope of the agent’s effective authority, you’re essentially adding a new authorization boundary,” she says.

That distinction between capability and authority is important. An agent may be technically capable of accessing a system, changing data, or executing a transaction, but that doesn’t mean the enterprise should automatically authorize it to do so.

Success can weaken oversight

Ironically, autonomy creep often occurs when AI works well. Paul Dongha, head of responsible AI and AI strategy at NatWest Group, sees that potential at the bank.

NatWest deployed agents in complaint handling, he says. Parts of the workflow that involve retrieving information from systems can be automated relatively easily, while humans continue to make the final decision on whether to uphold or reject a complaint.

If a system consistently produces the correct result, human reviewers may begin treating approval as a formality. The organization may eventually decide that the check is no longer necessary. “The temptation is to give over to the AI because it’s so reliable,” Dongha says.

NatWest sometimes reduces checking when an AI system demonstrates sufficiently high performance, he adds, but the decision is deliberate rather than incremental. It involves stakeholders and risk managers, and the system continues to be monitored afterward.

That’s a critical distinction for CIOs. Removing an approval step doesn’t necessarily represent poor governance. Removing it because everybody has gradually stopped paying attention to it does.

The warning sign is when a human moves from making a meaningful decision to merely rubber-stamping one — and then entirely disappears from the workflow without anyone reassessing the authority the agent has gained.

Put the controls outside the model

As agents acquire more authority, CIOs can’t rely solely on the agent to respect instructions about what it should and shouldn’t do. At 1Password, Wang advocates what she calls verified loops. Before an agent begins work, humans define its task, which tools it may use, what evidence will demonstrate that it completed the task successfully, and which actions it should never perform.

A separate verifier then compares evidence from the agent’s actions against that human-defined policy. “What you want to avoid here is something recursive where you have a fox guarding a henhouse, or the agent looking over its own work,” Wang says.

The principle applies beyond software development. A model might be instructed never to transfer more than a certain amount of money, delete a production environment, or access particular data. But if violating that rule would have serious consequences, the limit should also exist outside the model.

RSA’s Taylor makes the same distinction from the identity and access side. “Model-level guardrails are necessary, but not sufficient,” he says. “The real limits need to sit outside of that in enterprise control.” He argues for layered controls around agents, including identity policies, provisioning rules, lifecycle policies, governance, monitoring, and attestations.

In other words, CIOs shouldn’t ask an AI agent to exercise restraint where the infrastructure could simply prevent the unauthorized action.

Govern access in real time

Traditional access governance often operates on a schedule. A manager might review an employee’s permissions quarterly and determine whether those permissions are still appropriate. Agents may evolve too quickly for that approach. “Agentic identities are dynamic,” Taylor says. “So we need a mechanism that certifies them continuously, rather than performing periodic reviews.”

If an agent suddenly gains access to another database, for example, that event itself could trigger reassessment. Organizations should determine whether the new permission is appropriate or creates a dangerous combination with access the agent already possesses. “Privilege and authority should never be permanent,” Taylor adds. “They should always be reviewed.”

Wang favors an even more granular approach of task-based authorization. Rather than giving agents standing access or long-lived credentials, organizations can make authorization decisions at runtime and provide short-lived credentials only when required for a particular task. “We espouse no standing privileges,” she says. “No long-lived credentials.”

That effectively applies least privilege at agent speed. A change in task becomes an opportunity to reassess whether the agent still needs the access it’s requesting.

Make someone own the agent

Technical controls, however, don’t eliminate the organizational question of accountability. Taylor believes enterprises should be able to associate agents with a responsible identity. If an employee creates an agent, the organization needs to know who’s responsible for it, what its purpose is, and what happens to it if that person changes roles or leaves. “Every agent action needs to be attributed to a responsible identity,” he says.

Ownership also needs a lifecycle. An agent that remains important after its creator leaves should be transferred to another designated owner rather than simply continuing indefinitely.

Dongha makes another distinction. For AI embedded in a business process, he believes the business owner should remain accountable because that person owns the customer relationship, the process, and often the P&L. “It’s the business owner who has the responsibility because their people run it, and they’re responsible for the customer relationship,” he says.

That creates a useful division of roles for CIOs: IT may provide the infrastructure and controls, but the business can’t outsource accountability for the consequences of autonomous decisions.

Multi-agent systems need system-level controls

The challenge becomes more complicated when agents begin working together. An individual agent might remain entirely within its permitted scope while contributing to an overall outcome nobody intended. One agent generates an output. Another accepts it as input and takes an action. A third agent then acts on the result. Again, the individual actions may all appear legitimate.

Dongha says enterprises therefore need to test not only individual agents but the entire network as a system. “You need something overarching on top of a multi-agent system,” he says.

NatWest is working on an AI control tower that would give authorized users visibility into a network of agents, including messages passed among them and the guardrails surrounding those interactions. Dongha considers governance of multi-agent systems an emerging problem that nobody has fully solved yet.

Wang also points to a related challenge that identity becomes more complicated when agents are ephemeral. An organization may spin up a temporary team of agents, tear it down, and recreate it later. If something goes wrong, the enterprise needs to reconstruct which agent made which decision during which session, and how that action propagated through the system.

That makes traceability increasingly important as autonomous systems become less like individual applications and more like networks of interdependent actors.

Set the bar before the agent reaches it

Before an agent goes into production, Taylor says CIOs should be able to answer several basic questions: Who owns it, what can it do, what’s its maximum authority, how will the organization know if that scope changes, and which actions require a human decision.

Other industries don’t need to replicate the governance machinery of a large bank, but they can borrow one important principle: controls should be built into the system from the beginning. “Governance shouldn’t be a check at the end,” Dongha says. “It needs to be embedded when you build an AI system.”

For CIOs, that may be the central lesson of autonomy creep. The question isn’t simply whether an agent can be trusted today but whether the organization has mechanisms to notice when tomorrow’s agent has quietly acquired authority that nobody ever intended to give it.