CFOs have always been the custodians of commercial disciplines, but the rise of artificial intelligence has added a new dimension to that responsibility. The CFO who treats AI as someone else’s problem is making a strategic error that will compound over time, while those who lean in, ask hard questions and build coherent frameworks for AI investment will find that their businesses are still competitive by the end of this decade.
The modern CFO must now think in terms of two parallel tracks: deploying AI to strengthen what the business sells and also using it to transform how the business itself operates. Both demand proper investment, proper governance, and, crucially, proper leadership from finance.
Investing in what customers will pay for
On the external front, AI-enabled features are shifting from novelty to expectation, meaning organisations that fail to embed them into their products risk ceding ground to competitors who have already made that leap. There is a commercial upside, too: customers who are brought in early, as participants in pilot programmes and beta phases, tend to become loyal advocates, and premium pricing becomes easier to justify when the value delivered is demonstrable.
Go deeper with GlobalData
Access deeper industry intelligence
Experience unmatched clarity with a single platform that combines unique data, AI, and human expertise.
That said, the financial calculus here is genuinely complex. Variable costs tied to AI workloads are notoriously difficult to model at scale, particularly in the early stages of deployment. Regulatory frameworks around AI remain a moving target across both the UK and EU, with significant penalties attached to non-compliance. Meanwhile, the build-versus-buy question is being reshaped constantly by the extraordinary pace at which foundation models are improving and becoming cheaper at the same time. The CFO’s role is to cut through that complexity and ensure pricing structures remain flexible, cost trajectories are stress-tested, and the organisation never finds itself paying more to deliver AI than customers are actually willing to spend.
Getting the house in order
Internally, the opportunity is more immediately tangible. Phased rollouts that target high-volume, low-complexity processes such as approvals, data entry, reconciliation and routine reporting, can generate measurable returns relatively quickly. More importantly, the data governance infrastructure built to support these internal deployments lays the groundwork for external productisation and kicks off the process of building genuine AI fluency across the workforce, which is no small thing.
The obstacles here tend to be human rather than technical: low adoption rates, the proliferation of unsanctioned tools used without IT oversight, and a creeping over-reliance on automated recommendations are the enemies of a well-run AI programme. Finance should treat internal AI adoption as a managed experiment setting clear metrics from the outset, measuring outcomes rigorously and using the results to build the evidence base for the broader AI strategy.
This kind of disciplined internal experimentation provides something invaluable: real data. Not projections or vendor case studies but lived experience of what AI does and does not do well inside your specific organisation. That intelligence is enormously useful when it comes to making the case for external investment and for deciding where to place the next bet.
What AI will (and won’t) do for headcount
One of the more persistent misconceptions about AI in the enterprise is that it is fundamentally a headcount reduction story. In reality, people are more essential than ever to supervise AI agents, validate outputs, handle exceptions, and provide the commercial judgement that automated systems cannot replicate.
In product and engineering, for example, AI-assisted coding tools are already shortening development cycles, reducing time spent on boilerplate code, and helping teams identify testing gaps earlier. The latter matters increasingly to investors, who are paying closer attention to software quality and defect rates following a number of high-profile system failures in recent years.
In customer support, AI can handle routine queries at scale, freeing human teams to focus on the complex, relationship-defining interactions that actually matter. In procurement, a function that is almost entirely rules-based, the procure-to-pay cycle is a near-perfect candidate for intelligent automation. And in finance itself, AI tools capable of running continuous variance analysis, flagging deviations in real time, remodelling forecasts, and narrating the results in plain English, have the potential to meaningfully bridge the gap between raw data and genuine commercial insight.
Looking ahead
2026 is shaping up to be a pivotal moment for enterprise software and the organisations that have spent the past year building coherent AI strategies, developing internal capability, and gathering real evidence of what works will be distinctly better placed than those that have been running in circles.
By the end of the year, the signal-to-noise ratio should improve considerably and customer behaviour data, adoption metrics, and risk indicators will give finance leaders far more to work with. Macro-economic pressures, particularly around cost and supply chain, will push functions like procurement further under the microscope, making the case for intelligent automation easier to make.
The CFO’s enduring advantage has always been grounded in evidence rather than enthusiasm and as the market accelerates, the organisations that will genuinely benefit from AI are those whose finance leaders have had the rigour to invest deliberately, the patience to build properly and the clarity to distinguish real innovation from expensive noise.
Brandon Nussey, CFO at JAGGAER
