Research suggests up to 70% of UK businesses are using AI or planning to. After last year’s panic about an “AI bubble”, the consensus has swung the other way: 2026, we’re told, is the year AI ROI gets real. Boards want numbers. CIOs and CTOs are being asked to prove the spend.
They’re looking through the wrong lens. The fixation on measuring returns at the tool level, license by license, seat by seat, isn’t always a sign of financial discipline. Sometimes it’s a sign that the original investment decision was made without a clear problem to solve. You can’t calculate the return on an answer when nobody agreed on the question.
Head of AI at Slalom UK & Ireland.
Most organizations brought AI tools in to innovate, solve specific problems, improve productivity, or keep up with competitors. Others were meeting employee demand for best-in-class tools. Either way, AI became a must-have, and technology-focused C-suites came under pressure to implement at pace so the company could credibly claim it had AI Capabilities.
The demand to also demonstrate immediate ROI, layered on top, is where the framing starts to break down. The better question isn’t “what did this tool return?” but “has this investment created genuine value, broader than a figure on a page?”
The wrong question reveals the wrong strategy
A leader pushing hard for tool-level ROI is often revealing something they didn’t intend to. Ask what the AI is doing inside the business: which decisions it changes, which workflows it reshapes, which constraints it removes; and the answer often thins out quickly. That’s not a failure of measurement. It’s a failure of definition.
If an organization can’t articulate why the technology is needed beyond “staying ahead”, the demand for immediate returns is doing the work that strategic intent should have done upfront. AI implementation isn’t like rolling out a new CRM. It reorders how people make decisions, and it tends to expose whatever was already broken in those decisions.
Skipping the groundwork, which is the honest assessment of which problems are AI-shaped and which aren’t, and then asking the spreadsheet to retrofit a justification leaves teams confused and leadership defensive.
Businesses will not see genuine returns at pace if they cannot define what “return” means beyond the monetary. Treating AI like any other piece of software, where staff are handed access to a powerful tool because it might increase profits, doesn’t just fail to deliver results. It builds the friction and skepticism that make the next AI investment harder to justify too.
Who will the winners be?
The businesses that win with AI in 2026 won’t be the ones with the most sophisticated ROI dashboards. They’ll be the ones that started with a sharply defined problem and worked forward to a solution, rather than starting with a tool and working backwards to a justification. The ones that ask the right questions.
The difference shows up in how the conversation begins. “We need to deploy generative AI across customer service” is a tool-first frame; the ROI question becomes unanswerable because the goal is the deployment itself.
“Our agents spend 40% of every call searching three systems for policy information, and we want to cut that in half” is a problem-first frame; here, the return is built into the brief, and the AI either delivers against it or it doesn’t.
That does not mean task-level ROI is irrelevant. In fact, it is often the right place to start. If an AI system reduces the time it takes to summarize a case, draft a response, check a contract, classify an incident or retrieve policy information, that gain should be measured.
Task-level metrics help teams understand whether the technology is genuinely useful, where adoption is working, and which use cases deserve more investment. But they are still only a starting point. A faster task does not automatically create a better business outcome if the workflow around it remains unchanged.
The real question is whether those task-level gains accumulate into something more meaningful: fewer handoffs, faster decisions, better customer experience, lower operational risk, or more capacity for higher-value work.
ROI frameworks layered over the first kind of decision rarely create accountability. They create the appearance of it, usually after the fact, once leadership senses the investment isn’t landing and reaches for measurement as a shield. Defining the problem first doesn’t guarantee a return, but it’s the only way to know what one would even look like.
The human work is where ROI compounds
Even a well-defined problem won’t produce returns if the people closest to it aren’t equipped to use the tools. This is the part of AI investment that gets routinely under-funded: the human side of the rollout. Capability building, workflow redesign, time to experiment, and managers who can tell the difference between a team adopting a tool and a team performing adoption.
When the entire conversation is anchored to financial return, this work looks like overhead. It isn’t. It’s where the return compounds. Employees who understand what an AI system is good at (and what it’s not) make better decisions about when to use it, when to override it, and when to escalate. That judgement is the asset. Tool licenses aren’t.
A workforce that has been brought into the design of an AI rollout, rather than handed the output of one, also tends to be more receptive to the disruption that follows. Morale holds. Adoption is real rather than performed. And the return that everyone was so anxious to measure starts to show up, usually somewhere other than where the original spreadsheet was pointing.
Measure outcomes, not subscriptions
The fixation on tool-level measurement has pulled attention away from the question that should have come first: what problem are we solving, and for whom?
Returns from AI are real, but they show up at the level of capabilities and outcomes: faster decisions, fewer handoffs, better judgement at the edge of the business; not necessarily at the level of individual tool subscriptions.
This question will become more urgent as organizations move from copilots to agents. In a copilot world, it is still tempting to measure productivity task by task: how quickly someone drafted a document, searched for an answer, completed an analysis or generated a summary. Those measures can be useful, but they are limited.
In an agentic world, where AI systems can plan, act across tools, trigger workflows and involve humans only at specific decision points, those metrics will start to look inadequate.
The question will not be whether one task became 20% faster. It will be whether an entire process became more resilient, more auditable, more responsive or fundamentally different. If AI changes the unit of work, ROI measurement has to change too.
This is not an argument against measurement. It is an argument for measuring at the right altitude. Leaders who measure at the wrong level will keep finding the numbers disappointing, regardless of how good the technology gets.
The ones who define the problem first, invest in the people using the tools, and measure what those people are now able to do that they couldn’t before will find the returns were there all along. Just not on the line item where everyone was looking.
We’ve featured the best AI website builder.
This article was produced as part of TechRadar Pro Perspectives, our channel to feature the best and brightest minds in the technology industry today.
The views expressed here are those of the author and are not necessarily those of TechRadarPro or Future plc. If you are interested in contributing find out more here: https://www.techradar.com/pro/perspectives-how-to-submit
