Ask five vendors about AI ROI and you will get five confident numbers. The useful ones come from independent, large-sample research — and they tell a consistent story: the returns are real, but they are concentrated in a minority of organizations that do a few things differently.
The numbers that hold up
Across the most-cited studies, a few figures recur:
| Metric | Figure | Source |
|---|---|---|
| Return per $1 invested | ~$3.70 average; ~$10.30 for leaders | IDC / Microsoft, 2024 |
| Time to positive ROI | ~13 months (deploys in under 8) | IDC / Microsoft, 2024 |
| Report first-year ROI | 74% of executives | Google Cloud, 2025 |
| Report cost reductions | 42% (59% report revenue gains) | McKinsey / Stanford HAI |
| Are “AI high performers” | ~6% (5%+ of EBIT from AI) | McKinsey, 2025 |
These are directional industry benchmarks, not guarantees. Actual ROI depends on use-case selection, data readiness, architecture, and adoption — which is exactly what varies between the leaders and everyone else.
Why most ROI leaks away
The gap between the ~6% of high performers and the rest is rarely about model quality. Research consistently points to three causes: chasing pilots instead of production, bolting AI onto unchanged workflows, and having no financial model or metrics tied to each use case.
How to capture it
- Prioritize ruthlessly. A handful of high-value use cases beat a spread of experiments.
- Redesign the workflow, do not decorate it — high performers are ~3x more likely to do this.
- Instrument value early with a financial model and leading indicators per use case.
This is the core of a Celadon AI Audit: finding the use cases where the ROI is real and the path to production is clear — closely tied to build-vs-buy and vendor selection.
Find the ROI that is real for you
An AI Audit identifies the specific use cases where AI creates measurable value — ranked, costed, and tied to a path to production.
Get an AI Audit →