Post

Build a Tenant-Wide Copilot Credit View with the Power Platform API

Diesen Beitrag auf Deutsch lesen

Retrieve daily Copilot Credit consumption, store it in Dataverse, and create governed tenant-wide dashboards for agents and environments.

TL;DR

The Power Platform API exposes Copilot Credit entitlement and per-resource consumption data, including agent, environment, date, billed status and available dimensions. A scheduled Power Automate flow can retrieve one day at a time, follow continuation tokens, and store up to 180 days initially plus seven refreshed days in Dataverse for dashboards and exports.

Original by Petros Feleskouras, on The Custom Engine. Read the original

This is our own summary, not a republication or full translation.

Governance takeaway

  • Admins/CoE: build the scheduled flow to retrieve one day at a time via continuation tokens, but reach for PPAC first and persist this into Dataverse only when custom history or analysis genuinely requires it.
  • Security/Compliance: apply read-only access to the stored Dataverse records and set retention under policy, since this is ongoing consumption data accumulating for up to 180 days at a stretch.
  • Leadership/Business: a tenant-wide credit dashboard is achievable here, but budget for upkeep — environment IDs need separate resolution and optional response fields may change, so this isn’t a one-time build.

Frequently asked questions

Which Power Platform API route returns the tenant's Copilot Credit capacity?

The MCSMessages route under licensing and entitlements returns entitled, allocated, consumed, available and status values.

Why can the admin center reports fall short for Copilot Credit consumption?

The standard reports under Licensing and Copilot Studio do not let you reshape the data, own the records or keep trends beyond the standard window.

What limitations apply to the data from the Copilot Credit consumption API?

The API can return several rows for one agent on the same day when channel, feature, model or environment differ.

This post is licensed under CC BY 4.0 by the author.