CFOs have spent the past year investing in AI but many are still working out how to make it pay. McKinsey found that 98 per cent of CFOs had invested in automation technologies such as AI in the previous 12 months. It also found that 44 per cent were using generative AI in more than five use cases, up from 7 per cent a year earlier. However, for all the experimentation, progress inside the finance function remains uneven.
Many finance teams are experimenting with AI, but the evidence on day-to-day change is harder to quantify. McKinsey found that only 21 per cent of respondents using generative AI had fundamentally redesigned at least some workflows, while fewer than one in five were tracking well-defined KPIs for gen AI solutions. That leaves CFOs with a practical task: deciding where AI belongs in the finance function, what it changes in the way work is done and whether it is producing measurable gains.
“AI tends to expose the quality of finance leadership quite quickly,” says Loveday Loveridge, Consultant at Warren Partners. “In the right hands it sharpens judgement and frees up time for better decisions. In the wrong hands it creates more activity but very little value.”
From experimentation to execution
The attraction to the technology is obvious. Finance teams sit on large volumes of structured data, work through repeatable processes and are under constant pressure to improve speed, accuracy and visibility. Used well, AI can help with forecasting, reporting, working capital management and analysis. But finance leaders are right to be sceptical of grand claims. BCG found that median reported ROI from AI in finance is 10 per cent, while nearly a third of leaders said they had seen only limited gains. Those numbers suggest many companies are still working out where AI is genuinely useful in finance.
This is where the CFO’s role gets more demanding. Which use cases justify investment? Which need tighter controls? Which are distracting the business from more useful work? A strong finance leader sees AI as a series of choices about capital, process and accountability which can strengthen board level decision making. “Boards need a CFO who can give a clear account of what is changing in the finance function, where the risks sit and how progress is being measured,” says Loveday
Trust, controls and the boardroom
Governance is part of the same challenge. In finance, AI can influence forecasts, controls, scenario planning and reporting. If the data is weak or the output is taken at face value when it should be challenged, the risk is obvious. This gives CFOs a bigger brief as they’re being asked to weigh investment, build trust in outputs, set standards for use and explain to the board what can and cannot be trusted. An EY survey found that finance leaders are increasingly worried about the reliability of non-financial reporting data, which helps explain why trust and assurance are becoming bigger parts of the AI conversation. For all its efficiency benefits, AI makes scrutiny even more important.
Then there is the people question. The familiar headline version of this story says AI removes routine work and leaves everyone else to become more strategic. In reality, some tasks will disappear, others will change and teams will need a higher level of data fluency. The firms that handle that well will build stronger capabilities. A 2025 report from The CFO found that top-performing organisations were 104 per cent more likely than others to have launched AI or data science centres of excellence, while only 31 per cent of CFOs overall said they had done so.
Better questions from the leadership team
For boards, this should change the questions they ask. It’s easy to ask whether the finance team is “using AI” but a more useful insight would be discovering what has changed in the workflow, how value is being measured properly and who’s accountable for governance.
“The strongest finance teams will be built around sound judgement, clear accountability and people who know when to challenge the output. AI can support that work, but it cannot replace it,” says Loveday
The firms that make progress will be the ones that have done the work of redesigning processes, measuring outcomes and building teams that can use the technology without becoming over-reliant on it. Finance has a chance to lead that shift because it sits at the point where cost, control, performance and trust meet.
AI is already in the finance function. The question is which CFOs can turn it into better strategic decisions, stronger performance and a finance team built for the company’s successful future.
