AI Adoption Tracker
Rollouts usually measure licenses. This measures use. It is built on the shape of the Claude Enterprise Analytics API: active users, seats, and project, skill and connector usage by day. It shows which teams adopted and which ones need support.
Overview
fictional org · 400 seats · sample data
Is it sticking?
daily and weekly active as a share of seats · sample
Who adopted
weekly active rate, last 7 days · sample
| Team | Seats | Weekly active | Status |
|---|
Adopted = 60% or more weekly active (the target) · Growing = 35 to 59% · Needs support = under 35%
What people use
users active with each, in the selected range, any team · sample
Predicted vs actual
each team took the Readiness Scorer before rollout · sample
| Team | Readiness before | Adoption now | Read |
|---|
Where they disagree is the interesting part. A high readiness score with low use is the first team to go talk to. In my experience it points to a trust or workflow problem the survey missed.
What this data can't tell you
- It is not real time. Per Anthropic's docs, a day's data usually lands by about 13:30 UTC the next day, so the newest day you can read is yesterday, or two days back before that hour. Days are UTC days.
- What counts as active, per Anthropic: at least one chat message, a Claude Code session with tool use or git activity, or a Cowork session with tool use or messages.
- Claude Code used through Amazon Bedrock does not appear.
- No history before 1 Jan 2026, so no year-over-year yet.
- This view shows no cost. The API's cost endpoints cover usage-based plans; on seat-based plans they show usage credits only.
- The API can report activity per person. This dashboard aggregates to teams and is not for ranking or managing individuals.
- Only the organization's primary owner can create the API key, and it must live on a server, never in a page like this. The rate limit is 60 requests a minute for the whole organization.
- Readiness scores are self-reported survey answers. Activity is not value. Someone can be active and still not trust the output. Pair this with talking to people.
All figures are generated sample data. API details checked against Anthropic's Analytics API docs, 3 Oct 2026.
Why use rather than licenses
Seats tell you what was bought. Weekly actives by team tell you whether the work changed. The gap between the two is where adoption programs earn their keep.
Where the shape comes from
The metrics mirror the API's activity summaries: daily, weekly and monthly active users as a share of assigned seats, pending invites, and project, skill and connector usage by day. With a small server holding the Analytics API key, the sample generator can be replaced by the real endpoints.
Read it with the Scorer
Each team's pre-rollout readiness score sits beside its adoption. When they disagree, go talk to that team. Try the Readiness Scorer.