---
id: "question-defining-ai-roi"
type: "open-question"
source_timestamps: ["¶2", "§ Performance drive"]
tags: ["metrics", "roi"]
related: ["concept-performance-drive", "claim-95-percent-failure"]
resolutionPath: "A follow-up study detailing the specific KPIs (e.g., hours saved, revenue generated, error reduction) used by the successful 5% of companies."
source_url: "https://hbr.org/2025/09/what-companies-with-successful-ai-pilots-do-differently"
source_title: "What Companies with Successful AI Pilots Do Differently"
sources: ["execution"]
sourceVaultSlug: "hbr-seg-execution"
originDay: 8
articleStem: "hbr-foci-60-successful-ai-pilots"
sourceUrl: "https://hbr.org/2025/09/what-companies-with-successful-ai-pilots-do-differently"
sourceTitle: "What Companies with Successful AI Pilots Do Differently"
---
# What specific metrics define a successful AI pilot?

## Open question: What specific metrics define a successful AI pilot?

The article contrasts **'activity metrics'** with **'business outcomes'** and cites a [[claim-95-percent-failure|95% failure rate]] in delivering 'bottom-line returns', but **does not specify which exact financial or operational metrics** are most reliable for evaluating Gen AI ROI. This is the measurement gap inside [[concept-performance-drive]].

**Resolution path:** A follow-up study detailing the specific KPIs (e.g., hours saved, revenue generated, error reduction) used by the successful 5% of companies.

### Enrichment
Practitioner guides based on the MIT findings recommend defining **KPI ladders** — lead indicators plus lag P&L metrics — *before* build, and prioritizing back-office/process-automation use cases for early, measurable ROI.


## Related across articles
- [[concept-ai-economic-value-measurement]]
- [[claim-genai-hardest-to-value]]
- [[action-controlled-experiments]]