Stop Prompting, Start Building: Agentic AI for FP&A
Nicolas Boucher's 2026 playbook: build small agents for recurring FP&A work, chain them into a workflow, and fix the data before you trust them.
A great prompt is a favour. An agent is a process. FP&A needs processes.
📰 THE NEWS
Nicolas Boucher, founder of AI Finance Club and one of the most-followed voices in AI for finance, has been teaching teams to build agents for recurring work (commentary, variance analysis, contract review, data cleaning) with ChatGPT GPTs, Gemini Gems and Copilot agents, then chain them into one month-end workflow.
His 22 October masterclass with Numeric added the catch: an agent can only reason over what it can see. A journal entry tells you what moved, but rarely why, so the AI guesses and you spend your hours checking its work.
🔄 THE PIVOT
Value in FP&A doesn't come from the cleverest prompt. It comes from the boring, recurring jobs done the same way every month, with fixed instructions, fixed inputs and a named owner.
✅ ACTIONABLE TIP: build your first agent in one afternoon
1. Pick one task you repeat monthly with the same inputs (variance commentary is the classic).
2. Write the instructions once: role, inputs, output format, and what it must never do (e.g. invent drivers).
3. Attach reference files: chart of accounts, last month's approved commentary, materiality thresholds.
4. Test it on three past months where you already know the right answer.
5. Only then chain it to the next agent (commentary → CFO summary → board slide).
💬 What's the first FP&A agent you'd build: variance commentary, forecast refresh or board-pack summary?
👉 Join Free at www.FinanceClub.Ai
#AgenticAI #AIAgents #FPandA #FPA #AIinFinance #FinanceTransformation #NicolasBoucher #FinanceClubAi
Sources: Ramp webinar "5 AI Agents Every Finance Leader Should Build"; Numeric masterclass with Nicolas Boucher (22 Oct 2026).