GitHub’s October 22 changelog says Copilot usage metrics now report the names of custom skills, agents, MCP servers, slash commands and plugins created by an organization. That is useful for teams experimenting with multiple agents because it makes it easier to see which extensions are actually being used instead of treating all agent activity as one anonymous category.
For a small consultancy, the practical question is whether an AI add-on earns its keep. A custom MCP server that gets invoked constantly might be worth maintaining; one nobody uses is overhead. This feature is a reporting improvement, not a promise that Copilot measures each extension’s direct dollar cost or works on every individual plan.
Check your GitHub plan and organization access before investing time in a dashboard. If the metrics are available, identify your top-used custom agent or MCP integration and compare its use with the work it saves. If they are not, keep this as a model for instrumentation inside your own projects.
What to do: Check whether Copilot usage metrics are available for your organization and inspect the named custom-tool report.