Measure & Report AI Governance Seat Usage

[!NOTE]
This issue is a part of our larger AI Governance Add-On: PRD. See the GitHub Project for all related issues.

Our AI Governance Add-On (released in February) includes a number of seats, which can be a subset of the number of users within a Coder deployment. In the March release, we want to begin measuring and reporting to admins (and to tallyman) the number of seats consumed using AI Governance features (Bridge, Boundaries, Tasks).

Measurement

• AI Governance Seats: Similar to our agent-ready workspace starts measurement, our collector and Tallyman should measure AI Governance seats as users who have done 1 of the following activities:
  • Authenticated to AI Bridge and sent at least 1 request
  • Created an AI Task
  • Leveraged a workspace with Boundaries (not in scope to be measured yet)

Collection & Display

• This should be displayed in the licensing page under Agent-Ready Workspace starts as "AI Governance Users."
• This should also be sent as a tallyman metric

Enforcement & Warnings

• Warning banner: When an overage occurs (# of users > # entitled AI governance seats), a warning should show only to admins and an overage should be reported via Metronome/Tallyman.

Out of scope

• Fine-grained controls: Controls for admins to specify which users/seats can see/use AI governance features is out of scope for this specific issue.