Every AI vendor on the market will hand you a number. It will be large, it will be flattering, and it will not survive a single pass of due diligence from a competent finance team. We know because we've watched it happen — dozens of times.
The core problem is that most ROI numbers are built from a single survey of self-reported productivity lift, projected across an organization, and multiplied by a loaded fully-burdened cost figure. It's a persuasive narrative. It isn't a measurement.
Umbral Intelligence's 0–150 ROI Score starts somewhere else. It starts with the contract — what you actually pay, seat-by-seat. It then layers adoption telemetry (who logged in, who used the feature, who abandoned it), then task-level automation baselines from published research, and finally a conservative impact envelope that accounts for uncertainty.
The output is a number you can defend. Not a number that sells. That distinction — between a marketing metric and a measurable one — is what separates organizations that are operationally AI-native from organizations that are paying for the appearance of it.
If your AI portfolio can't answer the simple question 'what did we get?' in terms that trace back to data your auditor could verify, you don't have an AI strategy. You have a line item.