Collection
AI ROI for operations: the baseline, the stop rule and the bill
Your CFO asks two things about any AI line item: what it returned against a number finance already booked, and when it stops billing if the answer is nothing.
The articles below cover where an AI business case breaks in a mid-market operation: the board report, the pilot gate, and the bill that arrives after go-live.
Common questions
Frequently asked questions
What baseline will a CFO accept for AI savings?
A figure finance booked before launch, with the ledger or invoice set it came from, the period, and the date it was recorded. A baseline reconstructed from invoices after launch, or estimated, weakens the claim. If you have not launched yet, record the baseline now. If you launched without one, say so in the report and grade the claim down yourself before a director does.
Who should decide when an AI pilot stops?
The pilot sponsor and the finance controller, together. The sponsor holds the upside of continuing and the controller holds the invoice, so neither decides well alone. Route the decision to both at the review date, name one approver for each outcome (continue to production, extend with a new KPI, or stop), and let a slipped review default to stop.
What does an AI agent cost beyond the credit or token bill?
Rework. The invoice shows tokens, compute and credits. It does not show the hours your team spends finding and undoing a wrong action after it entered a real workflow, such as a reassignment made on stale data. That cost scales with how long bad output ran unverified, so a business case that nets savings against the credit bill alone overstates the return.
What we mean by AI ROI
AI ROI here means a dollar movement on a cost line finance can find in its own books, net of what the AI costs to run and to correct. A count of prompts, seats or decisions shows the system is used. It does not show the use paid.
Each article in this collection puts a named owner on one part of that number: the baseline, the pilot gate, or the meter. That ownership, with an approval step and an audit record behind it, is what we call decision infrastructure.
Where does your own operation's decision cycle stall?