What does a fractional CFO do for an AI startup?

The simplest definition

A bookkeeper records transactions. An accountant or controller helps ensure those records are accurate and reported correctly. A CFO uses the information to decide what the company can afford, what it should prioritize, how much capital it needs, and how to explain the plan to investors and the board.

A fractional CFO performs that CFO-level work without joining as a full-time executive. The company buys senior judgment and ownership at the level it actually needs.

What the work looks like in practice

The recurring work usually revolves around forecast, runway, hiring plan, board/investor reporting, and decision support. Event-driven work appears around fundraising, pricing changes, financing, valuation, diligence, large contracts, or changes in the business model.

A strong fractional CFO should not merely send reports. The CFO should be able to say: here is what changed, here is why it matters, here are the available choices, and here is what each choice does to cash, margin, dilution, or the next milestone.

What is different for AI startups

AI-native businesses introduce financial variables that classic SaaS companies often did not have to manage as directly. Each user action can have a variable model cost. Free users can create real COGS. Gross margin can change when engineering changes model routing. A more capable model may improve retention but hurt contribution margin. API prices may fall while product usage rises faster.

The CFO therefore needs a model that understands usage, inference, model/provider mix, infrastructure, pricing unit, and technical roadmap — not just revenue growth and headcount.

Fundraising role

During a raise, the fractional CFO helps determine how much capital the company needs, what milestones the round should fund, how dilution changes under different structures, and how the investor model connects to the pitch narrative. The CFO also prepares the financial data room, reconciles historical results to the model, and helps answer diligence questions.

When you probably need one

The trigger is usually not a revenue number. It is the point where financial decisions have become expensive to get wrong. If the company is preparing to raise, hiring quickly, missing forecasts, managing a board, dealing with volatile compute spend, or making pricing decisions without clear unit economics, CFO-level support can become useful before a full-time CFO makes sense.

What a good engagement leaves behind

The best outcome is a system: clean actuals, a driver-based forecast, clear metric definitions, cash/runway visibility, a recurring reporting rhythm, and a model the founder can use to test decisions. The fractional CFO should reduce dependence on heroic spreadsheet work, not create more of it.