A new survey of 300 CFOs found that 48% would follow an AI recommendation even when it conflicts with their own judgment. That is the highest figure across the entire C-suite. It deserves a harder look.

CFOs are the executives most willing to defer to AI over their own instincts. That finding from the Board 2026 Planning Intelligence Report is worth sitting with before you decide whether it is a good thing.

Board, the enterprise planning platform, surveyed 300 CFOs, CIOs and COOs at companies with annual revenues of at least $100 million. The headline finding got most of the attention: 83% of respondents said their board had made a strategic decision based on a forecast they already knew was outdated, with 40% reporting significant business consequences from that decision. A damning number, and one that speaks to the structural gap between how fast data moves and how slowly most planning cycles still operate.

But the number I keep coming back to is a different one. When asked whether they would follow an AI recommendation that conflicted with their own instinct, 48% of CFOs said yes. Chief operating officers came in at 11%. Chief information officers at 33%. Finance leaders are, by a wide margin, the C-suite executives most willing to trust the machine over themselves.

Why finance is particularly susceptible to this

There is a logic to it. CFOs have spent their careers being humbled by data. They have seen their own gut instincts on revenue forecasts, cost trajectories and working capital prove wrong, often precisely because they were relying on intuition rather than numbers. That experience creates a certain intellectual humility that other functions do not always develop in the same way. When a model says one thing and your instinct says another, the CFO's trained response is to interrogate the instinct rather than the model.

AI is now feeding into that pattern at scale. The Board survey found that 61% of executives cite large language models, specifically naming ChatGPT, Claude, Gemini and Copilot, among the external sources that most influence their strategic decisions. LLMs ranked ahead of industry peers, professional networks, technology vendors, external consultants and Gartner. That is a significant shift in how strategic input gets assembled in the modern enterprise, and finance leaders are at the front of it.

The governance gap this creates

The problem is that deferring to AI is not the same as trusting good data. A model is only as good as what it was trained on, what it was asked, and how its output was validated. Only 39% of respondents in the Board survey report having formal governance and escalation processes for AI-driven decisions. Which means the majority of CFOs who say they would follow an AI recommendation over their own judgment are doing so inside organisations that have not formally defined what counts as a trustworthy AI output, who is accountable when it goes wrong, or how it should be escalated when the stakes are high.

That is a meaningful gap. The same rigour that CFOs apply to financial controls, asking who signed off on this, what does the audit trail look like, how was this number derived, needs to apply to AI-driven inputs. The technology is capable of producing confident-sounding outputs that are directionally wrong. And the finance function, precisely because it is the most inclined to trust those outputs, is the part of the business most exposed if they turn out to be flawed.

What the best CFOs are actually doing

The finance leaders who are navigating this well are not the ones who distrust AI or the ones who defer to it uncritically. They are the ones who have thought carefully about which decisions benefit from AI input and which require human judgment to remain primary. Variance analysis at scale, scenario planning across hundreds of variables, supplier risk flagging: these are areas where AI demonstrably improves on human pattern recognition. Capital allocation decisions, M&A integration calls, and decisions about team structure and talent: these are areas where the CFO's contextual judgment is irreplaceable and should stay in the chair.

By the end of 2028, 36% of respondents in the Board survey expect AI to lead most enterprise planning decisions. That projection will come true in some organisations and not others. The ones where it goes well will be the ones where the CFO was deliberate about which decisions they were willing to hand over, and built the governance infrastructure to hold the model accountable. The ones where it goes badly will be the ones where the 48% just kept saying yes.