Most CFOs aren’t ‘very confident’ about their AI governance

WorkAI.TV Editorial Desk
3 Min Read

Share with your CFO

Only 40% of CFOs say they’re “very confident” in their AI governance frameworks, according to Deloitte’s Q2 2026 CFO Signals survey of 200 North American finance chiefs. The gap between adoption and oversight is stark: 93% now report AI running across multiple functions, up from two-thirds who were still experimenting less than three years ago. More than half cite lack of governance authority as a top challenge, while 46% flag cost unpredictability, specifically consumption-based pricing from cloud AI providers, as their leading internal concern.

What this means for your business

The confidence gap here is a structural problem, not a sentiment one. CFOs arrived late to AI ownership because the technology entered organizations through IT, operations, and product teams, not through finance. Now 93% of those same organizations have AI embedded across key functions, and the finance chief is expected to govern something that was never provisioned through their authority in the first place. The 51% who cite “lack of governance authority” as a top challenge aren’t confused about risk; they’re describing a real jurisdictional vacuum.

The cost unpredictability finding deserves more attention than it typically gets. Consumption-based pricing, where the bill scales with how often employees query a model or run an agent, behaves nothing like a traditional software license. A CFO who built their AI budget around seat counts or flat SaaS fees is now holding an open-ended credit line with unclear draw-down triggers. The 46% flagging cost uncertainty aren’t being conservative; they’re identifying the single AI risk that most directly lands on their function and that most vendor contracts are currently structured to obscure.

Shadow IT worry (28%) and rogue AI concern (27%) ranking lower than cost uncertainty tells you where CFO attention is actually anchored right now, which is the invoice, not the incident. That ordering will invert the first time a material AI-related loss event hits a peer company and gets disclosed in an earnings call. The CFOs who will be best positioned aren’t the ones with the tightest governance language today; they’re the ones who’ve already mapped which AI deployments carry consumption cost exposure versus which carry liability exposure, because those two risk profiles require entirely different contractual and oversight responses.

Based on reporting from Most CFOs aren’t ‘very confident’ about their AI governance, originally published 2026-08-03 08:53:00.

TAGGED:
Share This Article