Every CFO resume you read this week will say “leveraged AI to drive efficiency in the finance function.” That sentence tells you nothing. This post is the screening layer I use before I schedule the first interview.

The signal-vs-noise table

5 red flags

1. “AI-driven transformation” with no artifact. If they cannot describe the workflow they built and who else uses it, they did not build one.
2. Wrong tool for the claim. A resume that says “compressed close by 4 days using AI” but the tools listed are all ERP modules. Close compression comes from process, and AI plays a supporting role (accrual first-draft, flux commentary). If they conflate ERP with AI, they do not know the difference.
3. Team growth flat or up while AI adoption “increased efficiency”. Real AI adoption at the CFO level shows up in headcount ratios (revenue per finance FTE goes up). If they added 3 people while claiming efficiency gains, the story does not hold.
4. No published or shared thinking at all. In 2026, a CFO who has been “using AI seriously for two years” and has never written or spoken about it publicly is either lying about the timeframe or does not have a real point of view.
5. Names an AI vendor as a differentiator. “Certified in [vendor] AI platform.” Certifications from vendors are not signals; they are marketing. Ask what they built with it.

5 green flags

1. Names a specific workflow and its second owner. “Built a variance commentary workflow in Claude that my controller now runs every close.” The second owner is the whole signal.
2. Talks about what they refuse to automate. “I do not let AI draft the audit-committee narrative. That comes from me, always.” Judgment.
3. Verification discipline is default. Talks about tying AI output to source documents without being asked. “We reconciled the AR aging pull to the GL before the LLM ever saw it.”
4. Team ratio moved in the right direction. Revenue per finance FTE went up, or headcount shrunk while coverage grew. Not both at once, but one is real.
5. Has a strong opinion on tier and data privacy. “We run enterprise Claude on the paid tier because we do not want inputs used for training. Free tier for personal research only.” That is a five-second answer that filters cleanly.

The phone-screen question that filters 80% of noise

Walk me through one workflow you built with an LLM in the last 90 days. What did it replace, what did it not replace, and where does it still break?

Why this question wins:

• “Last 90 days” filters for current practice, not resume padding from a 2024 pilot.
• “Walk me through” requires narrative, not a bullet-point summary they memorized.
• “What did it replace” forces a concrete before-state. If they cannot name what the workflow replaced, they did not deploy one.
• “What did it not replace” is the judgment check. Strong candidates always have an answer here (the human review layer, the exception handling, the customer conversation).
• “Where does it still break” is the maturity check. Strong candidates give a specific brittleness (context window on the biggest customer, template drift when the source system changed). Weak candidates say “it works pretty well now.”

Ten minutes. If they cannot answer this cleanly, everything else on the resume is a maybe.

A quick scorecard for the resume screen

What has changed since 2019

Five years ago a CFO resume screen was about scale (biggest P&L run), pedigree (Big Four, brand-name MBAs), and system experience (NetSuite, Workday, Oracle). Those signals still matter, but they are table stakes. What has changed is that a strong 2019 CFO could run a 12-day close, publish a solid board pack in PowerPoint, and manage a 6-person team, and that was a fine year. In 2026, that same profile is running the same close in the same way at higher cost while a peer running the same business has a 6-day close, a 3-person team, and publishes the board pre-read as a written memo drafted with an LLM and edited in 45 minutes.

The delta is not the tools; it is the operating rhythm around the tools. Screening for that rhythm on paper is hard. That is why the phone screen question above matters more than any resume review. Twelve minutes of walking through one real workflow tells you more than an entire pedigree line.

Two other 2026-specific things to check for. First, whether the candidate has ever managed a finance team through an AI adoption without losing people. Adoption often produces resignations from the accountants who feel replaced. A CFO who has navigated that transition, retained the good ones, and moved out the fixed-mindset holdouts is worth a premium. Second, whether they can talk about the CFO-to-CTO relationship. In 2026 the CFO owns finance systems and data pipelines more directly than they did in 2019, and the working relationship with engineering is now a real part of the job. If they cannot describe how they partner with a head of engineering on finance data quality, they are running a 2019 playbook.

Frequently Asked Questions

Is a CFO with no AI experience a hard no?

Not automatically. A CFO with a strong operator record and a real curiosity signal (asking questions in the screen, has started using Claude in the last 60 days, willing to hire an AI-native controller under them) can catch up. What is a hard no: a CFO who is dismissive of AI, or who claims they “delegate that to the team.” That is 2019 thinking in a 2026 chair.

How do I evaluate a CFO who runs a small team?

Different bar. For a 2 to 4 person finance team, you are looking for personal fluency (they do the AI work themselves) rather than team-scale deployment. Ask them to share their own prompt library. If it exists and is 12 to 40 prompts organized by task, they are the real thing.

Should I ask for a live demo in the interview?

Yes, for CFO of a $50M+ business. Give them a monthly variance file and 30 minutes with Claude or ChatGPT. Watch them prompt. Same exercise as the FP&A analyst screen, higher bar on the framing and the edit judgment. If they refuse, pass unless there is an extraordinary reason.

Do I need an AI-native CFO if we are not a tech company?

Yes. AI-native is not a tech-industry skill; it is a finance productivity skill. A CFO at a services business, a manufacturer, or a hospitality group needs the same close compression and reporting velocity. If anything, non-tech businesses have more upside from AI in finance because their peers are further behind.

Are AI-native CFOs commanding a premium?

Yes. Base salary premium of 10% to 20% at the $300K to $500K level, and equity grants trend higher because these hires are still relatively scarce. The premium pays back in the first year because you hire a smaller finance team behind them. See the retained recruiter piece for how the fee math shakes out.

More from The Pragmatic CFO

• When to Hire a CFO: Fractional, Full-Time, or Interim
• The AI-Native Finance Team: 2026 Job Descriptions
• Retained vs Contingency: The Finance Executive Recruiter Playbook

Sources

• Heidrick & Struggles, “CFO Compensation Trends Report 2026.”
• Korn Ferry, “CFO Pulse Survey 2026.”
• Robert Half, “2026 Salary Guide for Finance and Accounting Professionals.”
• Brilyant, “State of AI in Finance 2026” survey.
• NACD, “The Board’s Role in Overseeing AI in Finance” 2025 briefing.
• CFO.com, ongoing coverage of AI adoption in the CFO seat, 2025-2026.

Written by The Pragmatic CFO. 15+ years running P&Ls and building finance teams across portfolio companies.