OpenAI CFO Sarah Friar published a vision this week for an "AI-native finance function" built around a "zero day close," a model where books are continuously reconciled rather than periodically closed. For most mid-market CFOs, that ambition is out of reach - and that's not the point. The useful takeaway is simpler: finance automation is moving toward continuously fresh information, and the practical question is which decisions are currently made with stale data.
Friar's essay describes a finance stack where systems reconcile continuously, AI handles repetitive knowledge work, and humans review exceptions rather than reconstruct data. Mid-market teams don't have OpenAI's engineering talent or appetite for rebuilding workflows from first principles. They have an ERP, a planning system, Excel, and a lean team doing month-end close. But the direction of travel applies to them too: away from periodic reporting and toward decision-ready information.
Decision latency, not speed, is the real KPI
The limiting factor in AI performance isn't the technology. New research from PYMNTS Intelligence found 71% of executives at companies with $1 billion or more in annual revenue say organizational readiness is the primary constraint, and only 11% cite the technology itself.
Michael Younkie, VP of Product Management at Billtrust, pointed to the underlying data issues: "We see inconsistent and incomplete data structures, bad data, dirty data. We see challenges around legacy ERP systems with limited AR API capabilities."
Friar recommends CFOs define which data AI can access, which actions it can take, when approval is required, and when an exception must escalate. Outputs must tie back to a reliable source, and changes to approved forecast baselines should stay under finance control. Success metrics should track useful work completed, cost after human review and rework, output usability, and whether the workflow produced a faster or better decision.
Where to start, if you're not OpenAI
A practical beginning point is one recurring finance process with an identifiable owner, measurable cycle time, and clear output. Automate part of it. Track exceptions. Measure the review burden. Decide whether to expand.
Finance chiefs looking to build these skills can start with a structured foundation, such as the AI for CFOs Learning Path, which focuses on how AI changes the finance function's role and workflows.
The candidates for AI support are familiar: variance commentary, audit support, covenant reporting, board materials, contract review, cash forecasting, and recurring management analysis. They consume skilled labor without requiring skilled judgment at every step.
Senior finance leaders exploring how to apply these principles to their own workflows might also consider the AI Learning Path for VP of Finance, which addresses practical implementation for finance executives.
The dividing line that matters is not strategic vs. administrative work. It's whether a workflow repeatedly forces finance employees to seek to reconstruct information that already exists somewhere in the company. If yes, AI has an economic case.
Why this matters for finance professionals
Close the book faster is the wrong near-term savings for most CFOs. The real target is decision latency, defined as the time between a business event and the financial information that reflects it. Start by mapping one recurring finance process, identify where information gets manually reassembled, and measure how much skilled hours that consumes. That is the foundation - not a zero-day close.
Your membership also unlocks:
