TL;DR: Finance teams that treat AI mainly as a compliance risk end up reviewing decisions the rest of the business already made with it. Abacum CEO Julio Martínez argues finance should be the company's proving ground for AI: use it on its own work first, decide once which numbers must be exact, and put the company's model behind everyone else's AI.

Finance has become AI's compliance desk

In August, OpenAI's own finance team ran a webinar on how they use AI to automate forecasting, reconcile models and build dashboards. The most revealing part of CFO.com's write-up wasn't the demo. It was the audience. Questions about controls, auditability, security and data integrity dominated the chat. Where is the data stored? Who has access? Does the workflow align with SOX?

Those are fair questions, and I'd want answers to all of them. What struck me was the order. Finance leaders watched OpenAI's finance team show what's now possible, and their first instinct was to ask how to control it.

The data says the same thing. In a survey of 687 finance and operations leaders published in September, 51% said their finance team doesn't use AI at all. Finance teams have quietly cast themselves as AI's compliance desk. I think that's the wrong seat, and it gets more expensive every quarter.

Why caution feels free, and isn't

Finance people spend their whole careers being rewarded for not being wrong. A misstated number gets noticed, discussed and remembered. A decision that came three weeks too late usually doesn't get noticed at all, because nobody ever sees the version of the quarter where finance spoke up in time. The cost of a slow finance team just shows up somewhere else, in decisions the business made without them.

How Strategic Drift reaches finance

In most companies I talk to, sales teams are already using AI to prioritize accounts and draft outreach. Marketers are shipping campaigns in days that used to take weeks. Engineers are writing code faster than they can review it. Each of those changes reaches finance eventually: a pipeline that suddenly looks different, a channel absorbing budget faster than planned, a hiring plan that no longer matches the roadmap. By then it's a variance, and the people who made the call knew weeks ago.

That's how Strategic Drift happens: each team moves quickly and rationally, and by the time the monthly review catches up, the company has moved off the plan. If finance teams are the slowest AI users in the building, they spend their time reviewing decisions the rest of the business already made with AI.

Deloitte's Finance Trends 2027 survey found that 54% of finance leaders are leading their company's AI and technology capital allocation decisions. So in many companies, the function deciding how much the business spends on AI is the one using it least.

Governance is a reason to lead, not to wait

Finance is the hardest place to get AI right, because the numbers have to add up and someone has to stand behind them. That's exactly why finance teams should be the proving ground. The ones getting this right have stopped asking whether AI is safe enough to use. They ask where it has to be exact and where it just has to be fast. We made a related argument about accountability for AI-produced numbers last week.

What an AI-native finance team does differently

Use it on your own work first, and visibly. First-pass variance commentary. Anomaly detection that flags a cost line before anyone asks. Scenario runs that used to eat a long weekend and now take an afternoon. Finance teams earn the right to set AI policy by being good at AI.

Draw the line once, then get out of the way. Decide which numbers have to be exact and traceable (the close, the board pack, the plan of record) and hold those to a hard standard. Let everything else move as fast as the rest of the company does.

Put your model behind everyone else's AI. Sales and marketing are going to ask AI about performance anyway. Make sure their tools are working from the company's numbers and assumptions, and not from a spreadsheet someone exported in June.

Why finance's context matters more as analysis gets cheap

As analysis gets cheap, every leader can generate their own forecast, and every one of them will look convincing. The risk stops being too little analysis and becomes too many plausible answers, each built on slightly different assumptions. Finance teams' real advantage is that they understand how the business actually makes money, and that context is exactly what everyone else's AI is missing. The CFOs who move first get to shape how the company thinks with AI. The ones who wait get to audit it.

Which function in your company is using AI most aggressively right now, and is your finance team learning from it or just reviewing it?

How Abacum helps

Abacum is an AI-native FP&A platform built for finance teams that want to lead with AI without giving up control of the numbers. Abacum Intelligence classifies, maps, and cleans incoming data, flags anomalies in real time, and turns plans and performance into clear, shareable explanations. Every AI-generated result is traceable: right-click any cell to see exactly which data sources, formulas, and assumptions drove it.

For the rest of the business, Abacum connects via MCP, so teams can ask questions in Claude, Gemini, or ChatGPT with the same governance guardrails intact. That is the practical version of putting your model behind everyone else's AI. You can read more about how Abacum builds financial intelligence, or get the free guide AI: faster, smarter decisions.