An AI business case template for CFOs who want ROI math before approving the next AI line item.
The business case is a vibes-based slide.
Most AI business cases we read are enthusiasm wearing a finance costume.
The opposite failure mode is over-modeled enthusiasm.
The right business case is grounded, narrow, and measurable.
Before any AI math, the proposal must measure the current state. Volume, unit cost, unit time, error rate.
What does it cost per event after the AI is deployed? This is the line where most cases fail.
The base case assumes a deployment that hits 65 to 75 percent of an internal best estimate. The downside case assumes 35 to 45 percent.
Before any AI math, the proposal must measure the current state.
What does it cost per event after the AI is deployed?
The base case assumes a deployment that hits 65 to 75 percent of an internal best estimate.
If your AI proposals are getting parked because the business case is hand-wavy, the answer is not a longer deck.
Two weeks of measurement on the existing workflow. Less than that and the baseline is an estimate. More than that and the program slows.
The framework is. Specific numbers depend on your model choice and infrastructure stack. We disclose every assumption.
We co-author the first one. We have built about 30 of these cases. Your team learns the format from the worked example.
