Issue No. 2 · Tuesday, June 24, 2026 · thecfodesk.com
THE SETUP
AI adoption across finance departments hit 97% this year, up from 76% in 2025. That sounds like the function has arrived. It hasn’t. A separate survey found only 17% of finance teams are actually using AI in their core workflows. The other 80% have a tool open in a browser tab. That’s not transformation. That’s a pilot that never graduated.
THE SIGNAL
Two numbers tell the whole story this week. Ninety-seven percent of finance departments report using AI, more than 75% of those investments are already generating positive returns within 12 months, and 42% of finance leaders say AI is now broadly or fully embedded in their organization — double the rate from a year ago.
That’s the adoption headline. Here’s the depth reality: a General Atlantic poll found 45% of finance teams are still in “limited pilot” mode, and only 17% are actively using AI in their core workflows. Finance lags every other business function — full workflow automation is already mainstream in engineering, marketing, and sales. Finance isn’t there yet.
Three things every CFO should internalize from this gap:
The bottleneck isn’t the technology. It’s confidence. Sixty-eight percent of CFOs say they don’t know where to start. That’s not a tooling problem — every major platform now has enterprise-grade finance capability. It’s a roadmap problem. The CFOs pulling ahead aren’t the ones with better software. They’re the ones who picked one workflow and finished it.
Security concerns are real, but they’re solvable, not permanent. Finance handles sensitive compensation, forecast, and board data — and enterprise tools resolve most of those concerns. Citing security as a reason to stay in pilot mode in 2026 is increasingly an excuse, not a constraint.
The function is being judged by board-level expectations it hasn’t met yet. Eighty-seven percent of CFOs say AI will be extremely or very important to their finance department’s operations this year, and only 2% say it won’t matter. Boards have absorbed the headline number. They haven’t absorbed the 17% reality. That gap between perception and execution is where CFOs get exposed in board meetings.
Our take: the finance leaders who move from pilot to production this year will look meaningfully more capable than peers stuck running parallel experiments. The technology is not the differentiator anymore. Follow-through is.
The CFO Desk reaches over 90,000 CFOs and senior finance executives at US public and private companies. To discuss sponsorship opportunities, email [email protected]
QUICK HITS
Four things every CFO needs to know this week.
· Risk management leads AI adoption at 81%, followed by financial reporting at 74%, treasury management at 68%, and tax functions at 66%. If your team hasn’t deployed AI in at least one of these four areas, you’re behind the median finance function, not ahead of it.
· Digital transformation is now the top CFO priority for 2026, cited by 50% of North American finance chiefs, according to Deloitte’s latest CFO Signals survey. This has displaced cost management as the leading concern for the first time in the survey’s history.
· More than half of CFOs — 54% — say embedding AI agents into the finance function is a transformation priority this year, not next year. Agent-based automation has moved from experimental to expected at the board level faster than most finance leaders are prepared for.
· One CFO’s real number: OneStream’s Bill Koefoed reports 39% efficiency gains in R&D from AI tools — but cautions that ROI in functions like marketing and sales is harder to prove. The lesson for finance: pick workflows with clean, measurable outputs first. Don’t start with the hardest problem.
THE PLAYBOOK
The 30/90/365 framework for moving from pilot to production
Days 1–30 — Pick one workflow, not five. Choose a single high-volume, low-judgment task — invoice coding, variance commentary drafting, or month-end reconciliation support. Resist the urge to pilot AI across the whole function simultaneously. Width kills depth.
Days 31–90 — Measure it like you’d measure any capital investment. Track hours saved, error rates, and cycle time before and after. If you can’t quantify the return in 90 days, you picked the wrong workflow or you’re not actually using the tool — you’re experimenting with it.
Day 90 — Make the call: scale or kill. The CFOs stuck in pilot purgatory are the ones who let a pilot run indefinitely without a decision point. Set the 90-day mark as a hard deadline for a yes or no.
Days 91–365 — Build the second workflow on the first one’s infrastructure. Once one workflow is in production, the security review, data access, and team training are already done. The second deployment should take a fraction of the time the first one did.
The CFOs who treat AI adoption as a single binary decision stay stuck at 17%. The ones who treat it as a repeatable, measurable process are the ones who move into the minority that’s actually using it.
Interested in reaching over 90,000 CFOs and senior finance executives at US public and private companies? Email [email protected] to discuss sponsorship opportunities.
THE CLOSE
One question: of the workflows in your finance function, which one is the cleanest, highest-volume, lowest-judgment candidate to move into AI production in the next 90 days? Hit reply and tell us. We read every response.
— The CFO Desk
[email protected] · thecfodesk.com
CAPITAL · FINANCE · LEADERSHIP