GitHub Copilot Metrics Now Track VS Code Agents Window Activity

GitHub has updated its Copilot usage metrics API to include activity data from the dedicated VS Code Agents window, with the change going generally available on 11 September 2026. Enterprise and organization admins can now pull daily active user counts, session totals, and per-user message data for both 1-day and 28-day reporting windows, giving teams a concrete way to measure how much of their Copilot seat spend is actually reaching the agentic coding workflow.

What happened

Detail	Value
Feature status	Generally available
Release date	11 September 2026
Reporting periods	1-day and 28-day
New aggregate fields	daily_active_vscode_agent_users, totals_by_vscode_agent (session_count, total_user_messages)
New per-user fields	used_vscode_agent, totals_by_vscode_agent (session_count, total_user_messages)
Scope	Enterprise and organization reports only

GitHub added two sets of new fields to the Copilot usage metrics API. At the aggregate level, enterprise and org reports now include daily_active_vscode_agent_users (a count of unique users who opened the VS Code Agents window each day) and totals_by_vscode_agent (a rollup of session count and total messages sent). At the per-user level, those same reports gain a boolean used_vscode_agent flag and the same session-plus-messages breakdown per individual.

All new fields are optional. If VS Code Agents window data is unavailable for a given period, the fields are absent or null rather than returning a zero, which GitHub says keeps the change backward compatible with existing API consumers.

What counts and what does not

These metrics cover only the dedicated VS Code Agents window, the panel specifically built for agent-style tasks in VS Code. They do not include activity from editor-window Agent Mode (the inline agentic feature inside the regular editor) or from the broader Copilot usage rollups. If your team uses both surfaces, you will need to query them separately to get a complete picture.

Who can see this data

Access is limited to:

• Enterprise owners and billing managers
• Organization owners
• Anyone with a custom org or enterprise role that includes the “View Copilot Metrics” permission

The Copilot usage metrics policy must also be enabled in your enterprise or org settings before any of these fields appear.

Why it matters

Copilot seats are not cheap, and most finance and engineering leads who buy them at scale have limited visibility into which features their developers actually use. The VS Code Agents window is a meaningfully different workflow from autocomplete or inline chat: it handles longer, multi-step tasks. Knowing whether your team uses it, and how often, is useful when deciding whether to train, expand, or cut a Copilot rollout.

The separation of aggregate and per-user data is practical. You can use the org-level numbers to spot adoption trends over a 28-day window, then drop to the user-level report to identify which teams or individuals are driving engagement versus ignoring the feature entirely. For teams building internal tooling around Copilot, this is the kind of signal that justifies or kills a line item. If you are thinking about AI integration for your development workflow, metering actual usage is a prerequisite before scaling.

Our take

This is a straightforward, useful addition. GitHub is filling an obvious gap: organizations were already paying for VS Code Agents access but had no API-level way to prove adoption. The backward-compatible null behavior is the right call and shows some care for teams that have already built dashboards on top of the metrics API.

One caveat worth noting: the split between “VS Code Agents window” and “editor-window Agent Mode” is a real footgun for anyone building a usage dashboard. Both are “agentic” Copilot features, but they live in separate metric buckets. Before you report to leadership that agent adoption is low, double-check which surface your developers are actually on. We have seen this kind of reporting gap cause real confusion in practice, as covered in our look at Copilot’s expanding VS Code panel features.

What to do about it

1. Check that the Copilot usage metrics policy is enabled in your enterprise or org settings, otherwise none of these fields will appear.
2. Pull the updated API documentation and add daily_active_vscode_agent_users and totals_by_vscode_agent to your existing usage reports or dashboards.
3. Run a 28-day aggregate report to establish a baseline before drawing any conclusions about adoption.
4. Cross-reference per-user used_vscode_agent data against team rosters to find which groups need enablement or training.
5. Keep editor-window Agent Mode metrics separate in your reporting until GitHub consolidates the two surfaces.

Solid metrics are the foundation of any informed AI rollout. Instrument now, before your next seat renewal conversation.

Source: GitHub Changelog