Look past the volume
A high token count can be all maintenance. The mix tells you what that spend actually produced.
What the agents actually worked on
Work type
Is AI moving the roadmap, or holding the line? The work-type mix answers it, session by session.
Why it matters
Not all engineering work is equal
Watch features against maintenance
New features, bug fixes and maintenance trended side by side, so a slow drift away from building shows up early.
Show leadership what AI bought
The answer to what AI actually bought you is a breakdown of the work, per agent and per team.
How it works
How work type is classified
Sessions are read, not guessed at
Each session's prompts and edits are classified into a work type automatically. Nobody tags anything by hand.
It rolls up by team
Classifications roll up per agent, per team and per repository, and are trended across whatever window you pick.
Compare periods
Put this month's mix against last month's to see whether AI capacity moved toward new features, or quietly away from them.
AI ready at scale
Observability that was built for AI,
Observability that was built for AI,
not repackaged for it
Run your agents and teams on the right foundation, so they don't approximate, they know
14 days, all features, no credit card required.
<1msAlert detection time
99.9%Platform uptime
10xData ingestion capacity