I'm an FP&A Manager. Here's What AI Actually Does

There’s a version of the AI story that gets told a lot in Finance circles right now. It goes like this: AI will automate everything. Your analysts will be replaced. The CFO will get real-time insights from a dashboard that runs itself. Finance as a function will be unrecognizable in three years.

I've read it in trade publications. I've sat in webinars where someone who has never built a rolling forecast in their life explains what AI means for FP&A.

Here's what I've actually found — after building forecasts, owning board decks, managing monthly closes, and quietly experimenting with AI/automation tools for the past four years in a real finance team:

Some of it is genuinely transformative. Some of it is overhyped. And most of the honest reporting on which is which is missing, because the people writing about it aren't the ones doing the work.

That's what IntelliCues is here to fix.

Let's start with the good news, because there is real good news.

1. First-draft narrative work

If you've ever spent a late afternoon writing variance commentary — translating a table of numbers into prose that a CFO can read in 90 seconds — you already understand why this is the first thing AI changed for me.

The variance happened. The numbers are in the model. You know the story. The problem is that converting it into clear, professional language takes time and cognitive effort that, frankly, you've usually already spent by the time Sunday evening rolls around.

A well-structured prompt can produce a solid first draft of that commentary in under two minutes. Not a perfect draft — you'll always need to add context that AI can't know, the deal that slipped at the last minute, the headcount decision that isn't in the system yet. But the structure, the language, the flow — AI handles that. You handle the judgment layer on top.

I'll be covering the exact prompts that work for this in detail. It's the single highest-ROI use of AI I've found in an FP&A context.

2. Scenario narratives and planning language

Building three scenarios for a board presentation is analytically straightforward. Writing three distinct, coherent narratives that explain each one — in language that a non-finance board member can follow — is surprisingly hard and time-consuming.

AI is good at this. Give it the key variables and ranges, specify the audience and tone, and it will produce narrative scaffolding that would have taken you an hour in twenty seconds. You reshape it, sharpen it, add the organizational context. The heavy lifting of getting words on a page is gone.

3. Challenging your own thinking

This one is underrated. One of the most useful things you can do before presenting a forecast or a business case is ask AI to play the role of a skeptical CFO and interrogate your assumptions.

It won't know everything your CFO knows. But it will identify logical gaps, missing assumptions, and questions you haven't prepared for — reliably and quickly. I now do this before every significant presentation. It has saved me from embarrassment more than once.

This is where the vendor pitches fall apart. And it's important, because misunderstanding AI's limitations in a finance context doesn't just lead to disappointment — it can lead to bad outputs finding their way into board packs.

1. Replace judgment on organizational context

AI doesn't know that the regional VP who's responsible for the APAC miss has already had that conversation with the CEO. It doesn't know that the headcount variance is technically favorable but politically sensitive. It doesn't know which numbers the CFO will fixate on and which ones she'll wave past.

Organizational context — the invisible layer of history, relationships, and unspoken rules that shapes every piece of finance communication — lives entirely with you. AI produces technically correct outputs. You decide whether they're appropriate.

2. Handle messy, real-world data

The finance data in vendor demos is always clean. Yours probably isn't.

Inconsistent categorization across business units. Legacy systems that don't talk to each other. A chart of accounts that's been reorganized three times. Manual adjustments that live in a spreadsheet tab called "Other."

AI tools work well when you give them clean, structured inputs. Preparing those inputs is still your job — and in most real finance environments, it's non-trivial. Don't let anyone tell you otherwise.

3. Do the actual analysis

AI can summarize. It can structure. It can draft. What it can't do is tell you why the APAC business is underperforming, what that means for next year's plan, or whether the explanation your regional team gave you holds up under scrutiny.

Analysis — the kind that requires pattern recognition built over years, knowledge of the business, and the ability to distinguish signal from noise in messy data — is still a deeply human skill. AI accelerates the production work around it. It doesn't replace it.

Here's the framing I keep coming back to: AI is very good at the parts of FP&A that feel like production. The variance commentary. The narrative formatting. The scenario language. The meeting prep. The first draft of anything.

It is not good at the parts of FP&A that require judgment, context, and genuine analytical thinking.

The problem is that most finance roles have been evaluated — and compensated — primarily on the production work. How fast you close. How clean the deck looks. How quickly you turn around the CFO's ad hoc requests.

If AI handles that layer, what's left is the judgment layer. Business partnering. Interpreting ambiguous data. Advising on decisions, not just reporting on outcomes. Telling the story behind the numbers, not just writing the numbers down.

That's a more valuable version of the FP&A role. It's also a more demanding one. And getting there requires actually learning to use these tools — not waiting until your company mandates it.

The finance professionals who figure this out in the next two years will be significantly better positioned than those who don't. That's not a vendor pitch. It's just where the work is going.

I started this site because I couldn't find what I was looking for: an honest, practitioner-written resource on AI for FP&A that wasn't funded by a software company or written by someone who had never sat in a budget review.

Every week, I'll publish one piece covering something specific and practical — a prompt that works, a workflow I've tested, a tool review from a finance user's perspective, or a clear-eyed take on where this is all heading.

No hype. No vendor relationships. No content written by someone who's never built a rolling forecast. If that's useful to you, subscribe below.

IntelliCues is written by an FP&A manager writing honestly about AI in finance. New posts weekly. Subscribe at intellicues.com.