Avalara Surveyed 1,505 CFOs and Finance Leaders. 92% Feel Pressure to Prove AI ROI. Only 7% Put Governance Ahead of Speed.

October 16, 2026 · Josh Agusti, Founder, Aiven Consulting Group

Avalara's July 2026 survey of 1,505 CFOs and senior finance executives across the United States, United Kingdom, India, and Australia found that 92 percent feel moderate or significant career pressure to prove that AI agent investments are delivering a return, with half describing that pressure as significant.

The same survey found that half of respondents say their AI agent initiatives have produced only limited measurable ROI so far. The gap between the pressure finance leaders feel and the proof they can actually show is the real story here, and it says more about how AI agents are being deployed than any single productivity number could.

The Pressure Without the Proof

Finance has always been the function that has to answer for spending, so it makes sense that CFOs feel the squeeze first. What is notable in Avalara's data is how far ahead of the evidence that pressure has run. Boards and CEOs want AI agents handling invoicing, reconciliation, forecasting, and vendor management now, and they want a number attached to it. Finance teams are producing agents faster than they are producing the measurement discipline to say whether those agents are actually working, which leaves CFOs defending a number they cannot fully stand behind.

Governance Is Still an Afterthought

The deployment speed shows up directly in the governance numbers. Only 7 percent of respondents say their organization puts governance ahead of deployment speed, and 30 percent report their company has not updated its internal controls in the past year, even as AI agents take on tasks that used to require a human sign-off. Perhaps the most telling figure is that 23 percent of finance leaders say responsibility for an AI agent mishap would fall to no one or would be unclear. That is not a technology gap. It is an ownership gap, and it tends to surface only after something has already gone wrong. Closing it starts with the same kind of documentation most finance functions already know how to build, just applied to a new category of actor; Aiven Consulting Group's governance templates are built around exactly that starting point.

The Audit Trail Problem

Audit readiness compounds the accountability problem. Avalara found that only 28 percent of surveyed organizations require documented audit logs showing how an AI agent reached a given decision, and 46 percent say their AI incident response plans are untested or still in development. Without a log, there is no way to reconstruct what an agent did or why, which makes the 23 percent who cannot name an owner even harder to defend to a board or a regulator after the fact. That pattern also shows up in a broader look at AI audit readiness, which is worth reading alongside these numbers for a fuller picture of where enterprise audit trails for AI agents currently stand.

What CFOs Should Actually Measure

None of this means finance leaders should slow down AI agent deployment and wait for perfect governance. The pressure to show returns is real and will not go away. What it means is that the return itself needs a sturdier definition than a single productivity estimate. An agent that saves twenty hours a month but leaves no audit trail and no named owner is not actually a smaller risk than one that saves five hours and has both. CFOs who want a credible answer the next time a board asks about AI ROI are better served by pairing a structured ROI calculation with the controls and ownership record to back it up, rather than a number that cannot survive the first follow-up question.

Avalara's numbers describe a finance function caught between two demands moving at different speeds: prove the return now, and build the governance to defend that return later. The CFOs who come out ahead will be the ones who treat those as the same project instead of two separate ones.

Key Takeaways

Avalara's July 2026 survey found 92 percent of CFOs and senior finance leaders feel pressure to prove AI agent ROI, but half say their AI agent initiatives have only produced limited measurable returns so far.
Governance has not kept pace with deployment: only 7 percent of finance leaders say governance comes before deployment speed, 30 percent haven't updated internal controls in a year, and 23 percent couldn't say who would be accountable for an AI agent mishap.
Only 28 percent of organizations require documented audit logs of AI agent decisions, and 46 percent have incident response plans that are untested or still in development, meaning most finance functions could not fully reconstruct what an agent did if something went wrong.