AI Readiness in Finance: 5 Questions Every CFO Should Answer Before Buying Anything

AI tools are widely available, relatively affordable, and very well marketed. Copilot, ChatGPT Enterprise, specialised finance AI platforms — the options are everywhere. The pressure on CFOs and finance directors to "do something with AI" is real.

But before any purchase, any licence, any pilot — it's worth sitting down and honestly answering five basic questions. They'll save you time, money, and frustration.

Question 1: Do we know where our time actually goes?

AI is an optimisation tool. But you can only optimise what you can clearly see. If you don't know where your team spends most of its time on routine, repetitive tasks, you don't have a foundation for an AI strategy.

A good first step: ask your team to track what they do and how long it takes for one week. The results will probably surprise you — and they'll show you where AI can help most.

Typical high-potential areas in finance: monthly and quarterly reporting, invoice checking and matching, contract processing and analysis, preparing management commentary, tax research and tracking regulatory changes.

Question 2: Is our data in good shape?

This is the question companies most often answer too optimistically. "We have data." But in practice, that data is scattered across ERP systems, spreadsheets, shared drives, and email threads. It's inconsistent, incomplete, or in formats that AI can't process effectively.

AI is only as good as the data it receives. Bad data won't be improved by AI — it will be amplified and turned into faster mistakes.

Key questions for assessing data readiness: Where is the data we want AI to work with actually stored? Is it accessible from one place, or fragmented? How is it structured? Who owns it?

Question 3: Do we understand the security risks?

Financial data is among the most sensitive data a company holds. Before introducing any AI tool, you need clarity on several things: What data will we be putting into the tool? Where is that data processed and stored? How does the tool handle data under GDPR? Can the tool use our inputs to train its underlying model?

These policies don't need to be invented from scratch — there are established approaches to AI governance in finance teams. But they need to be in place before launch, not after the first incident.

Question 4: Is the team willing to change how they work?

Technology is the easier part. Changing habits is harder. Finance teams have established processes — and for good reason. Accuracy, control, auditability. AI doesn't threaten these values, but it does change how they're achieved.

The key question: Are people in the team willing to experiment? Or is the resistance to change strong enough that any pilot will fail regardless of the tool?

Adoption matters as much as technology. A company that doesn't communicate to its team why AI is being introduced and what it means for them will almost certainly end up with an unused tool.

Question 5: Who will own AI in the team?

Every project needs an owner. An AI project in finance is no exception. Who will coordinate the rollout? Who will be the first point of contact for questions from the team? Who will evaluate whether AI is delivering results?

In smaller companies, this can be the CFO or a senior controller. In larger organisations, it makes sense to appoint an "AI champion" — someone the team trusts, who understands the processes, and who is willing to engage seriously with the technology.

How to Use These Questions

This isn't a test where a wrong answer means stopping. It's a map. If you don't know where your time goes — that's step one. If your data isn't ready — that's step two. Working through these steps is what prepares you for AI that actually has a chance of working.

Companies that skip the preparation and go straight to buying tools end up paying twice: once for the licence, and once for the time spent on a pilot that leads nowhere.