The One Question Your CFO Is About to Ask About AI

Your CFO has been patient.

She approved the first AI pilot eighteen months ago. She signed off on the vendor contracts. She sat through the demos. She nodded at the technology roadmap. She's been supportive, even enthusiastic.

Now she has one question.

"What's the ROI?"

And if you can't answer it — specifically, in numbers, connected to EBITDA — you have a problem that's bigger than your AI program.

Why This Question Is Coming Now

The first wave of enterprise AI adoption was characterized by experimentation. Boards and CFOs gave their CEOs significant latitude to explore, pilot, and learn. The implicit understanding was: this is new, we need to figure it out, results will come.

That window is closing.

In 2026, AI is no longer new. The vendors have been paid. The pilots have run. The question has shifted from "are we experimenting with AI?" to "what are we getting for our AI investment?"

If you don't have a clear, defensible answer to that question, you're not just at risk of losing your AI budget. You're at risk of losing credibility with your board.

The Measurement Problem

Most enterprise AI programs were never designed to produce CFO-ready metrics. They were designed to produce technical outputs: accuracy rates, processing speeds, model performance benchmarks.

These are not business metrics. Your CFO doesn't care about your model's F1 score. She cares about:

If your AI program can't produce numbers in these categories, it wasn't designed to produce business results. It was designed to produce technology outputs.

How to Fix It Before the Board Meeting

The good news is that this is fixable — if you act before the board meeting, not during it.

The fix requires three things:

1. Retroactive outcome mapping. Go back to every active AI initiative and map each one to a specific business outcome. Not a technology output — a business outcome. If you can't make that connection, the initiative needs to be restructured or discontinued.

2. A measurement framework. Build a simple, consistent framework that tracks AI-attributable business impact across all initiatives. This doesn't need to be sophisticated. It needs to be credible and defensible.

3. A board-ready narrative. Your CFO and board don't want a technology briefing. They want a business briefing. The narrative should answer three questions: What did we invest? What did we get? What are we doing next?

The companies that get this right don't wait for the CFO to ask the question. They answer it before she asks — with a quarterly AI performance report that connects every initiative to a business outcome.

That's what board-ready AI governance looks like. And it starts with strategy, not tools.