With CFOs under pressure to cut costs, many are turning to AI. Amid this backdrop, Aidana Zhakupbekova, COFO at fintech firm Rydoo explores why AI investment needs clear metrics to measure returns, before companies count on efficiency gains too early.
It’s been a challenging year for finance leaders. Rising inflation, borrowing costs and geopolitical uncertainty continue to eat into the bottom line and make planning ahead increasingly difficult. As a result, CFOs are looking closely at where they can cut costs, with Deloitte’s latest UK CFO survey revealing that 68% of CFOs are prioritising cost reduction, up from 53% in Q2 2026.
Despite this, AI is one area where businesses are still prepared to spend. With its promise to reduce manual workloads and boost efficiency, CFOs are counting on it as the answer to their problems.
Indeed, AI is already supporting finance teams to work more efficiently. Tasks such as expense management, reporting and data analysis that take up considerable time can be effectively automated, giving employees more time to focus on higher level, strategic work that requires their judgement. With economic uncertainty persisting, this headspace is extremely valuable as finance teams evaluate risks and set the strategy for the months and years ahead.
However, beyond automation, large scale AI transformation is expensive. What’s more, beyond the initial investment, there are ongoing costs such as integrating, running and overseeing these tools. As a result, companies need to be deliberate about how and where they invest in the technology and ensure that clear metrics are established from the outset.
As CFOs, we’re used to questioning spend and asking what return we can expect. AI adoption shouldn’t be any different just because of the hype surrounding it.
Importantly, returns on AI investment can take a while to show up. Gartner research found that simpler AI applications, including data extraction and accounts payable automation, generally take nine to ten months to deliver returns. More complex applications can take much longer.
That makes it even more important to set realistic expectations from the start. Finance leaders need to know what they’re spending and what improvements they’re expecting, with regular reviews of the actual efficiency gains, before relying on them.
Deloitte’s survey shows that more than half of CFOs expect to reduce headcount over the next year as a result of AI. In reality, this might be much too soon.
The majority of AI systems require a human in the loop who can oversee and question the outputs and remain accountable for the decisions taken. Cutting staff before these systems are properly embedded risks losing the very people needed to make them work. Businesses need to give employees time to get comfortable with the technology and understand where it’s actually making a difference.
And when AI fails to deliver the expected results, and there will be instances where this is the case, businesses may be tempted to invest in even more technology, hoping another tool will finally improve productivity. But if employees aren’t using the existing systems properly, adding more is unlikely to help. More tools not only increase spend, but they add complexity and potential duplication, neither of which are conducive to greater productivity or lower costs.
On the whole, a scattergun approach to AI adoption can quickly eat into returns, so CFOs need to understand where AI can realistically deliver before making staffing decisions. The potential is real, but the savings will only follow once the technology is properly embedded and measured.
