A local AI connector, part of your subscription
Install @baytekdev/agile-analytics-mcp with one npx line and Claude, or any assistant that speaks the Model Context Protocol, can ask Agile Analytics for a sprint summary, cycle time, flow efficiency, work in progress, aging and a forecast for any team. The connector runs on your own computer and signs in as you, with a personal access token or with az login for organisations connected to Microsoft Entra. It answers with counts, dates and work item ids only, never titles or names, and sends nothing to Baytek. What your assistant does with those numbers is between you and its provider.
Admins stay in control
Configuration → General has a new switch, "Allow the local AI connector", for organisation admins only. It is on by default; turn it off to refuse every such connection. The connector checks it before every answer and replies "disabled by admin" when it is off. Each change is recorded in the Admin activity log. The switch does not remove anyone’s Azure DevOps access, which already reaches the same data.
Improvements
The Dashboard’s Flow Efficiency tile now counts each item under the workflow mapping for its type, the same rule as the Flow Efficiency view, so teams with per-type mappings see one percent in both places. In Aging & Blocked, the "Includes N still-open items carried over" note counts only the items the chart draws; items still in New have no lane and no longer inflate it. User Metrics counts pull requests linked from GitHub through the Boards-GitHub connection the same way as Azure Repos ones. Flow Efficiency and the other history-based views keep the milliseconds of each state change, so two changes inside the same second (scripted imports, demo data) no longer drop an item as having no usable history.
What you need to do
Nothing for the extension itself; Azure DevOps updates it automatically. To use the connector, follow the set-up guide: one config entry in your assistant and a personal access token with three read scopes. Admins who do not want it in their organisation turn it off on the General tab.
Related: Set-up guide: the local AI connector · 6.74.1: Epic Timeline and Quality & Rework