On 22 October 2026 GitHub added an Industry comparison section to the Copilot impact dashboard. It lets an enterprise, or an organisation not owned by one, compare its Copilot agentic adoption against aggregated benchmarks from participating peers. Participation is opt-in and disabled by default.
What gets measured
- Adoption: the percentage of users who are “agent-first” or “multi-agent”, meaning they hand work to AI agents rather than just accepting autocomplete.
- Leverage: how the pull request merge rate of agentic users compares with non-agentic users.
- Results are ranked against the peer median for a 28-day period and, as more firms join, against the top decile.
- Peer groups are chosen by company size (1–999, 1,000–4,999, 5,000+ employees) and one of ten industries. Comparisons refresh daily.
GitHub notes rankings only appear once enough businesses opt in, so early numbers may show only a median or a wider peer group. The comparison is dashboard-only: it is not in the usage metrics API or exports.
How to read it
A merge rate is a measure of work accepted, not of work that was good. A team can lift it by opening smaller, easier pull requests. “Agent-first” tells you how people work, not what they deliver. These are useful signals about adoption and poor signals about quality or speed to market.
So what
Expect suppliers to start quoting “we rank in the top decile for AI adoption”. Treat it as a conversation opener, not proof. Better questions: what do you ship in a month with a team of your size, how is AI-written code reviewed, and what does it cost to change things after launch? Outcomes such as delivered features, defects found in testing and time to release are the measures that matter to you.
At BuildApps we use agents extensively, but we are measured on working software and named human review. That is the approach described under AI-assisted development. If you want to see how it applies to your product, start a project.