.jpeg)
Table of Contents
Key takeaways
AI has become accepted and embedded across daily finance workflows, from forecasting and reporting to scenario planning. Pigment's Q2 2026 CFO Index found that only 3.2% of finance teams have no plans at all to adopt it, with another 8.4% on the path but still in the planning stage.
The largest share of respondents (28.8%) are scaling AI across multiple workflows, while a quarter (25.3%) remain in early implementation and a growing number (15.9%) describe their adoption as “mature” – up two percentage points from our Q1 survey.
Where you sit on that spectrum matters, because AI maturity now tracks closely with how finance leaders feel about their business. Sorted into the Index’s four maturity tiers, just 23% of early-stage organizations describe themselves as “very confident” in their company’s direction. That number rises to 34% at the developing stage, 43% at advanced, and 75% at leading. Our Q1 survey found the same gradient in performance, with the most AI-mature organizations reporting 18.1% revenue growth year over year, compared to 6.2% for early-stage teams.
The correlation doesn’t prove that maturity produces confidence, and a fast-growing business has its own reasons to be both optimistic and quick to adopt. But it does mean our attitude data is worth a closer look. If the teams furthest along the curve are also the ones performing best, then what they say now about AI describes what a finance function looks like once the technology is fully embedded. So let's get into it.
The Pigment CFO Index tracks financial performance, AI adoption, and market uncertainty across 2,000 finance leaders in the US, UK, France, and Germany. Explore the Q2 2026 report →
The value of AI is moving from doing the work to informing the decision
Productivity and efficiency gains remain the most-cited benefit of AI in finance, barely changing from 58.8% of respondents in Q1 to 59.7% in Q2. Faster planning and forecasting sits just behind at 54.1%, up 4.4 points from Q1. Speed is the core of the AI value case, and it’s not going anywhere.
But there's plenty of movement elsewhere. The largest single gain was in quality of decision-making, now cited by 50.8% of respondents, which is up 5.9 points from Q1. Accuracy and confidence in planning also rose 5.4 points to 46.8%. Both are judgment benefits rather than throughput benefits. They describe AI as improving what a finance team decides, not just how quickly it can assemble the numbers supporting the decision. That points to AI being trusted with more strategic tasks – like forecasting, variance analysis, and scenario work – rather than just the mechanical steps that feed them.
One benefit, however, moved sharply in the other direction. Reduced manual work was the second-most-cited benefit in Q1 and has since fallen 7.4 points to 42.7%. Improved work experience and more time for strategic projects also slipped down the scale, though far less sharply (down 2.7 and 1.8 points, respectively).
Reduced manual work is cited most heavily in France (at 46.4%, the highest of any market), where only 5.2% of organizations describe their AI adoption as mature and more than half remain at or before the pilot stage. It’s cited least (40.5%) in the US, a market that leads on nearly every AI adoption measure and one where decision quality (56%) and faster cycles (59.8%) pull ahead.
That pattern suggests that reduced manual work isn’t fading as a value because it stopped happening, but because it’s the benefit teams notice first, record immediately, and then stop tracking. Once automation is in place, it becomes the new baseline.
Data quality is now the single biggest barrier, ahead of cost
Data quality and availability has overtaken cost as the most-cited obstacle to getting value from AI, rising from 35.7% to 40.3%. It holds that top position across every crosstab in the Index.
High costs, the previous top barrier, held roughly steady at around 39%. That's unsurprising, as nothing in the past quarter has made AI meaningfully cheaper - but The CFO Index’s next edition will specifically track how the market responds to changes in AI pricing models.
Meanwhile, every barrier rooted in internal capabilities eased:
- Lack of internal expertise fell 3 points, from 33.4% to 30.4%.
- Unclear ROI and cultural resistance fell 1.3 and 1.5 points, respectively.
A manager is nearly twice as likely as a VP or CFO to flag poor-quality outputs as a barrier (16.4% versus 8.9%). That’s the same perception gap we identified in our Q1 data, and it has not narrowed. The people closest to the outputs are the ones registering the problem, which means senior leaders assessing their own data readiness are likely working with a more flattering view of it. Getting the data foundation right is now the highest-leverage AI investment most finance teams can make, and knowing where a business actually stands is a critical first step.
The Pigment CFO Index tracks financial performance, AI adoption, and market uncertainty across 2,000 finance leaders in the US, UK, France, and Germany. Explore the Q2 2026 report →
Friction that remains is structural, not cultural
Two barriers moved up while the internal ones came down. Slow procurement and implementation rose 1.5 points, and regulatory and security concerns are now cited by 28.8% of respondents.
What we think that means is that finance teams have largely stopped debating whether to use AI and have now learned how to use it. What’s left are the factors they don’t fully control, like data spread across disconnected systems, vendors and models that have to clear a compliance bar, and buying processes designed for slower categories of software. Finance is a highly regulated function purchasing into a category where governance standards, audit trails, and data residency guarantees are still being written. That invites careful, committee-led evaluations like security reviews, vendor risk assessments, and legal sign-offs, all of which take time.
The market data supports that reading: Germany has the highest share of mature adopters of any market in the Index (at 21.4%) and also reports the highest slow-procurement barrier score (29.1%) and the highest cost barrier score (47.4%). Deeper deployment doesn’t remove procurement friction but exposes it, because the next set of deployments sit closer to core financial processes and are thus harder to approve.
Planned AI spend is flattening at the top
AI budgets are still growing at every maturity stage, but the pace of that growth is changing.
The most AI-mature cohort now plan to increase their AI budgets by 22.7% over the coming year, down from 29.9% in Q1. Advanced organizations moved the other way, from 21.8% to 28.7%. That’s a reversal from Q1, as the second-most mature group is now planning a larger budget increase than the most mature one.
The straightforward explanation is coverage. Leading teams have already inserted AI into most of the workflows where it obviously pays, so their remaining spend is incremental: additional seats, expansion, and renewal. Meanwhile, advanced teams still have visible gaps to fill and a clearer place for their marginal dollars to go. The market picture reinforces this finding. France, the least mature market, posted the largest swing anywhere in the Index, with planned AI investment up 16.2 points to 27.3%, taking it from last to first on planned increases. Germany, which is the most cost-sensitive market, is planning the smallest increase at 13.3%.
The Pigment CFO Index tracks financial performance, AI adoption, and market uncertainty across 2,000 finance leaders in the US, UK, France, and Germany. Explore the Q2 2026 report →
None of this signals a retreat – the average planned increase across the full sample actually rose, from 17% in Q1 to 19.1% in Q2. But it does put a ceiling into view, and it changes the question in front of finance leaders. While budgets were expanding to cover new ground, the case for AI spend was largely self-evident. Now, as that growth rate flattens, AI spend has to defend itself against the same standard as every other line item.
Entering the ROI phase
The throughline across all three shifts is that finance teams have stopped asking whether AI works and started asking what it returns.
The experimentation period answered the first question. It’s now reasonably clear which AI applications produce value and which ones don’t, and the ones that don’t can just be phased out. The second question is harder, because most organizations can’t answer it with the data they currently have. AI spend arrives as vendor invoices, platform dashboards, and engineering-led monitoring. That’s enough to see the total, but not enough to attribute it. If you can’t allocate AI cost to a team, a workflow, or an initiative, you can’t set it against the value that initiative produces. “AI is expensive” is as far as that analysis can get.
That’s the gap Pigment’s AI Token Management Application template was built to close. It brings token usage and cost into the planning environment, where AI spend can be allocated to departments, projects, and use cases, tracked live against budget, forecast under different adoption and model-mix assumptions, and modeled against the value it produces – from hours saved to productivity gained. It makes AI spend behave like any other cost finance already governs: attributable, forecastable, and testable against a clear return.
Attitudes toward AI have moved faster than the systems used to account for it. Closing that gap is what the next stage asks of finance.
Don't let your spend on AI outrun your ability to govern it
Track token usage, allocate costs to the right teams and projects, and test spend against real returns with Pigment’s AI Investment Planner.
