Copilot’s dashboard now counts chat users, and “passive” becomes “chat-first”

GitHub has expanded its Copilot impact dashboard to show nine Copilot features, adding Copilot Chat and the VS Code Agents window, according to a changelog entry. The “Passive users” cohort is now called “Chat-first users”.

Two new features and one renamed cohort

The Agent adoption chart previously covered seven features. It now also reports Copilot Chat and the VS Code Agents window. The same two features appear in the enterprise and organization 28-day usage metrics reports, where copilot_feature_engagement.totals_by_feature gains chat and vscode_agent entries.

The renamed cohort covers users who are active in Copilot but have not used a qualifying feature on at least two days in the 28-day period. They therefore do not qualify for an adoption phase. In the API they are still the No Cohort value, and cohort sizes do not change.

GitHub’s reasoning is that chat is a core Copilot experience that the dashboard did not show, and that people who mainly chat were being labeled passive. In its words, the change gives “a more complete picture of how developers use Copilot.”

What the numbers will and will not count

The phase model itself stays as it was. Chat and the VS Code Agents window show up as engagement figures only, and neither feeds into a user’s adoption phase, so a team’s agent-first or multi-agent share should not shift because of this update.

The engagement test is the same one the other features use: someone is engaged with a feature once they have used it on two or more days within the 28-day window, and one person can appear under several features. The chat count takes in agent mode as recorded by usage metrics, which means a developer who accepts agent edits is also tallied under chat. That overlap means the per-feature counts cannot simply be added together.

The new entries exist only in enterprise and organization 28-day reports. Reports covering earlier periods list just the original seven features. Viewing the data requires being an enterprise owner or billing manager, an organization owner, or holding a custom role that includes the View Copilot Metrics permission, and the Copilot usage metrics policy has to be switched on. GitHub does not mention a price or a separate plan requirement.

What to watch as chat joins the dashboard

The dashboard has been growing quickly. The July 2026 version introduced cohort cards for code-first, agent-first and multi-agent users plus a passive segment, based on a rolling 28-day window, as we reported on the October industry comparison. That update let companies rank their agent use against peers by size band and one of ten industries, and GitHub said pooled benchmark figures are erased after 30 days, a claim that has not been independently verified.

The API side is moving in step. A November 5 changelog entry added a copilot_code_review object to enterprise and organization repos-1-day reports, with counts of evaluated and addressed comments split by severity and review tier. Our earlier piece covered how the usage API separates own-key traffic from Copilot-hosted models.

Several things are not established. GitHub publishes no figures on how many users fall into the chat-first group or how much chat use there is compared with agent features. Because chat does not feed adoption phases, a chat-heavy team will still see a modest share of users in the advanced phases. The practical signal to watch is whether administrators who now see chat engagement start to question whether the phase model, which still rewards agent features only, reflects real developer work. GitHub does not say whether chat will ever count toward phases.