A new study suggests more than 80% of companies are not tracking the return-on-investment (ROI) of their AI spending.
The survey — commissioned by US cloud computing giant Akamai and conducted by market research company Opinium — polled 750 senior tech professionals across the UK, France and Germany on their approach to AI.
According to the survey, published on Tuesday, 82% of businesses have yet to implement “a strategy for tracking the ROI of their AI projects”. Despite this, 65% of those polled say they plan to increase AI spending over the next 12 months.
“Traditional ROI models don’t map neatly to AI,” says James Kretchmar, CTO of Cloud Technology at Akamai. “Productivity alone isn’t enough. Companies have to prioritise the quality of outcomes and use the right tool for the job.”
As pressure mounts on much-hyped AI applications to deliver results, Sifted asked European AI startup leaders and other experts in the field how companies can keep better tabs on the technology’s returns.
Time saved
Nick Mason, founder and CEO of Turtl, a London-based automated content platform backed by Octopus Ventures, says the smartest businesses are tracking AI’s ROI in three ways: time, impact and revenue.
“Start with time. If AI can automate or speed up manual processes, like analysing behavioural data or creating content, it should show up in reduced production hours and freed-up team resources,” says Mason.
Time can be one of the easier variables to track. Whether cutting down repetitive tasks or speeding up decision-making, time saved can be measured in staff hours — and potentially headcount.
Stef van Grieken is the CEO of Amsterdam-based biotech startup Cradle, a generative AI platform enabling clients to engineer different protein designs, which in turn can be used in a range of applications, from agriculture to pharmaceuticals.
Cradle measures ROI based on how quickly teams achieve their protein-design goals. “Customers report development timelines between 1.5 and 12 times faster, leading to significant cost and time savings,” he tells Sifted.
“Productivity use-cases are usually about how much time is freed up,” agrees Sanjin Bicanic, a partner in the AI practice at Bain & Company. “Then the ROI is in either having fewer people because each is more productive [or employees are] doing the next best thing with the freed-up time.”
Wider impact
Companies need to closely monitor how things looked before and after rolling out AI.
Can you do things you couldn’t before? Are customers reporting better outcomes? These are the kind of broader questions you should be asking.
Matthijs Huiskamp, CEO of AI-powered bid-management platform Altura, also based in Amsterdam, says using the technology shouldn't just be about saving time or, as he puts it, “looking for a faster horse”.
At Altura, teams use AI to analyse hundreds of past tenders at once, then apply the insights gathered to rework their strategy for a current bid. “That level of insight and strategic depth simply wasn’t feasible before AI,” says Huiskamp, explaining how his company focuses more on the number of bids clients win, rather than the hours employees save.
“AI isn’t just about doing the same things faster; it’s about doing what was never possible before.”
This kind of strategic transformation might not be so easy to quantify, but the results can be significant. Tata Maytesyan, a self-described AI marketing expert who previously worked at Nike and Deloitte, tells Sifted: “The key is to first identify the specific process or objective AI is intended to impact.
“Is it to optimise existing processes [...] or enable entirely new initiatives like expanding into a new market?” Once that’s clear, says Tata, companies can compare factors like resources needed and customer response to a given AI use-case.
Revenue up?
Finally: the bottom line.
Freeing up time and getting improved customer feedback are worthy pursuits, but won’t count for much if revenue is stagnant.
But how easy is it to measure dollars-in versus dollars-out? It’s not always straightforward, particularly when it comes to assistants like Microsoft’s Copilot or OpenAI’s ChatGPT.
“They usually help with a lot of small tasks and that adds up. but it’s hard to track it in a way in which you can attribute the ROI,” says Bicanic.
Experts advise focusing on where AI directly impacts costs or creates measurable efficiencies — the kinds of changes that will show up on a balance sheet.
At Cambridge-based electric vehicle startup Monumo, CEO Dominic Vergine says the company has cut the bill for materials needed by more than 10% through AI, which it uses to identify “system-level improvements that were previously impossible to discover”.
“For the automotive industry, AI is not a nice-to-have, it's becoming a critical component for all manufacturers to help improve their bottom line,” Vergine says.
Knowing exactly where you hope to save money — and boost profits — is key to any AI project, so keep a close eye on these pre- and post-adoption figures to make sure you’re getting the best out of the technology.
“Most AI initiatives live or die on whether they can tie back to revenue,” says Mason. “If your dashboards can’t show a direct line between AI activity and pipeline growth, you’re not measuring ROI — you’re working on a wing and a prayer.”
