The candidates you want in 2026 are running monthly variance analysis in a prompt library, verifying model output against source documents, and killing 8 hours of weekly work by turning a memo into a repeatable workflow. Those people scroll past your JD because it reads like it was written in 2018. Fix the JD and the pipeline changes overnight.

Why your 2019 JDs are broken

A 2019 finance JD is a laundry list of certifications, tool names, and vague soft skills. “CPA required or preferred.” “Advanced Excel including pivot tables and VLOOKUPs.” “Highly detail-oriented.” “Strong communicator.” “Able to work independently.” None of that predicts performance in a finance function where the analyst is expected to compress a 6-hour close task into 40 minutes using an LLM plus a source-of-truth check.

What predicts performance now: has the candidate shipped an AI workflow into production, can they describe the exact prompt they use for the first pass of a monthly variance, do they know which tasks they will not automate, and do they have a point of view on data-privacy trade-offs between paid and free model tiers. None of that is on the JD you posted last quarter.

JD 1: Chief Financial Officer (AI-Native)

Headline: Own the finance function of a $30M to $150M business, use AI as a lever across the team, and hold the line on the 20% of work that must stay human.

Responsibilities

• Own the annual budget, monthly forecast, and 3-year plan; publish variance analysis within 5 business days of close.
• Run monthly and quarterly close; target a 7 business day close on GAAP books.
• Report to the board and audit committee; own the board deck and pre-read.
• Manage banking, lender, and audit relationships; lead capital raises and refinancings.
• Build and manage the finance team (controller, FP&A, accounting); set the AI standard the team operates by.
• Own the AI operating model for finance: pick tools, set the data-privacy policy, decide what stays in Excel and what moves into an LLM workflow.
• Own systems selection (ERP, FP&A platform, expense, close software) with a 2-year TCO view.
• Partner with the CEO on pricing, unit economics, and hiring plans.
• Sit in front of the CRO’s pipeline weekly and stress-test the forecast against actual bookings velocity.

Qualifications

• 10+ years in finance leadership, at least one full cycle as CFO or interim CFO of a $20M+ business.
• Prior experience closing a Series A/B/C, a lender package, or a strategic sale.
• Fluent in modern FP&A tooling and at least one LLM-based workflow deployed in a prior role.

Signals of an AI-native candidate

• Has deployed at least one LLM workflow into monthly close or FP&A that a team member other than them still uses.
• Can name their prompt library and where it lives.
• Verifies model output against a source document before publishing; can describe the exact check.
• Has an opinion on paid vs free tier data-privacy trade-offs and can defend it.
• Can name at least two finance tasks they will not automate and why.
• Uses AI to draft board memos and audit-committee narratives, then rewrites them; does not ship first-draft output.

JD 2: FP&A Manager (AI-Native)

Headline: Own the forecast, the variance story, and the monthly business review. Use AI to do the work of a 3-person team as a team of one, and know when to slow down.

Responsibilities

• Own the monthly forecast rollup and variance-to-plan commentary.
• Build and maintain the KPI dashboard by department; publish 3 business days after close.
• Partner with department heads on annual budget builds and quarterly reforecasts.
• Own the customer-cohort and unit-economics model; refresh monthly.
• Build first drafts of board and investor materials for CFO review.
• Maintain the FP&A prompt library; version-control it in a shared repo.
• Run scenario analysis on demand (best/base/worst, hiring pause, price change) within 24 hours.
• Own the finance side of headcount planning; reconcile HRIS to plan monthly.

Qualifications

• 4 to 8 years in FP&A, banking, or consulting; at least 2 years in an operating finance seat.
• Fluent in one modern FP&A tool (Pigment, Cube, Mosaic, Anaplan, or equivalent) and one LLM interface.
• Can build a 3-statement model without touching a template.

Signals of an AI-native candidate

• Has shipped at least one LLM workflow that reduced a recurring monthly task by 50% or more.
• Can name their prompt library and show it on demand.
• Verifies AI output against a source (GL, HRIS, CRM export) before it leaves their inbox.
• Has opinions on which tasks NOT to automate (audit-facing memos, comp analysis for a specific person, board narrative on a bad quarter).
• Can articulate the difference in data handling between a paid enterprise tier and a free consumer tier.

JD 3: Controller (AI-Native)

Headline: Own the close, the books, and audit readiness. Use AI to compress close from 12 days to 7, and to make the audit trail cleaner than it was last year.

Responsibilities

• Own monthly and quarterly close on GAAP books; target 7 business days.
• Own AR, AP, payroll, sales tax, and treasury operations.
• Own the chart of accounts and the accounting policy manual.
• Own audit prep and the auditor relationship; deliver PBC lists on time.
• Own systems administration for ERP, expense, and close-management software.
• Build LLM-assisted workflows for reconciliations, flux analysis, and journal-entry review.
• Manage 1 to 4 accountants; set the accounting standard the team operates by.
• Partner with FP&A on the actuals-to-plan handoff each month.

Qualifications

• 7+ years in accounting, at least 3 as controller or assistant controller.
• CPA preferred, not required if the operational experience is deep.
• Big Four or high-growth private company experience closing US GAAP books.

Signals of an AI-native candidate

• Has deployed at least one LLM workflow into close that another accountant on the team still uses.
• Can name the recons where they use AI first-pass and the ones where they refuse to.
• Verifies AI output against the GL and the source system every time; can describe the check.
• Has opinions on data-privacy trade-offs and has read their firm’s AI policy.
• Can name at least two accounting tasks they will not automate (complex accruals, related-party review, revenue-recognition judgment calls).

What to remove from your old JD

• “Proficient in Excel” listed as the top technical skill. Everyone is proficient in Excel. Ask for a specific workflow they built.
• “Attention to detail” as a headline trait. This filters for people who like being detail-oriented, not people who ship clean work at speed.
• “PowerPoint-first” or “polished deck creator” as a top-of-list ask. Boards read pre-reads, not decks, in 2026.
• “Big Four required.” Big Four is a fine training ground; requiring it in 2026 cuts 60% of your best operator candidates.
• Certification checklists (CPA, CFA, MBA all required). Pick one, and only if the role actually needs it.
• “Fast-paced environment” and “wearing many hats.” Every candidate has read this 200 times. Say what the job actually is.
• “Own end-to-end” without saying end-to-end of what.

Frequently Asked Questions

What if my finance team is anti-AI?

You have a leadership problem, not a hiring problem. Fix the leader first, then hire under them. An anti-AI finance leader in 2026 is a legacy hire from 2019 who is running a 12-day close and blaming the ERP. Give them a real 90-day mandate to prove out one AI workflow. If they cannot, move them.

Do I pay more for AI-native talent?

A slight premium at the manager and controller level (5% to 12%), meaningful at the CFO level (10% to 20%) if they have a shipped track record. The premium pays for itself in the first year because you hire smaller (one senior instead of two mid-levels).

Can I retrain my existing controller?

Yes, if they want to. Give them 6 hours a week for 90 days, a real project (compress the reconciliation of one balance sheet account), a paid enterprise LLM tier, and a peer to work with. Retraining does not work if they are treating it as extra homework.

Which certifications matter?

CPA still matters for controller and audit-facing roles. CFA still matters for treasury, M&A, and investor-facing CFO seats. MBA is a plus, never a filter. Anyone requiring all three is filtering for pedigree, not performance.

Remote vs onsite for AI-native roles?

Hybrid wins for controller and above. Fully remote works for FP&A managers and analysts if you have a rhythm of monthly onsites. Fully onsite is a losing hand for AI-native hiring in 2026; the pool you want is not local to you.

More from The Pragmatic CFO

• When to Hire a CFO: Fractional, Full-Time, or Interim
• Interviewing an AI-Native FP&A Analyst: 15 Questions
• How to Screen an AI-Native CFO Resume

Sources

• U.S. Bureau of Labor Statistics, Occupational Employment and Wage Statistics for “Financial Managers” and “Accountants and Auditors”, May 2025 release.
• Robert Half, “2026 Salary Guide for Finance and Accounting Professionals.”
• Brilyant, “State of AI in Finance 2026” survey.
• Michael Page, “Finance and Accounting Talent Trends 2026.”
• CFO.com, ongoing coverage of AI adoption in finance functions, 2025-2026.
• NACD, “The Board’s Role in Overseeing AI in Finance” 2025 briefing.

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