AI Lead Time Comparison shows the weekly average Lead Time for AI-assisted PRs alongside non-AI PRs, broken down by the same sub-stages used in regular Lead Time (Coding, Waiting for Review, In Review, Ready to Deploy). It's the most direct measure of whether AI tool usage is making PRs faster or slower.
What it measures
For each week in the period, the average Lead Time of:
PRs marked AI-assisted (at least one commit on a day the developer used an AI tool).
PRs marked non-AI (no AI activity recorded for the author on the commit days).
Both groups are decomposed into the four standard Lead Time stages.
How Leanmote calculates it
ai_lead_time(week) = avg(lead_time_for_change(pr)) for AI-assisted PRs merged that week non_ai_lead_time(week) = avg(lead_time_for_change(pr)) for non-AI PRs merged that week
Same Lead Time for Changes formula as the DORA metric — see Lead Time for Changes.
AI-assisted classification is per PR, based on the author's AI tool activity on the commit dates.
Sub-stage decomposition uses the same definitions as Lead Time decomposition (Coding, Waiting for Review, In Review, Ready to Deploy).
How to interpret it
AI lead time consistently shorter — AI tools are correlating with faster delivery for that team or repo.
AI shorter only in Coding stage — the speedup is in writing code, but review/release stages dominate the cycle. Real-world impact may be smaller than the coding number suggests.
AI longer in In Review or Ready to Deploy — AI-assisted PRs may be larger or harder to review. Pair with Avg Lines per PR from AI Lines of Code.
Wide week-to-week variance — sample size is too small. Aggregate over more weeks.
What to do about it
If AI-assisted PRs trend faster overall, document the workflows that produced the win and spread them.
If AI-assisted PRs are slower in review, coach toward smaller AI-assisted PRs and clearer commit messages.
Use this metric quarterly, not weekly. Single weeks are noisy.
Related metrics
Lead Time for Changes
AI Intensity
AI-assisted PRs
AI Governance metrics overview