Analytics overview
The Analytics section in platform Settings gives you three views of platform activity: Cost, Usage, and Productivity. Each view is available at platform scope, aggregated across every project, and at project scope, covering one project. All three tabs share a period picker in their fixed header.

Cost
Cost shows total AI spend for the selected period, broken down by feature and by model. Open Settings → Cost to see organization-wide spend. A project's own Settings → Cost tab scopes the same view to that project and shows budget utilization against any configured monthly ceiling.
Read more on Cost.
Usage
Usage shows token consumption and platform activity. Metric cards cover total spend, fresh tokens sent to providers, cached reads, and cache hit rate. A token-mix chart and a platform activity section track artifacts created, API events, knowledge queries, and AI generation runs over the period.
Read more on Usage.
Productivity
Productivity shows delivery throughput and flow health. Metric cards cover features completed, average cycle time, weekly throughput, and current blocked count. A feature flow chart and a flow health grid break down where time is spent, and a completed feature table lists each feature with its cycle time, review time, blocked time, and AI cost.
Read more on Productivity.
The period picker
Every analytics tab has a period picker in its fixed header. Preset ranges are This month, Last month, Last 7 days, Last 30 days, and Last 90 days. Each tab retains its own period selection independently; changing the period on Cost does not affect Usage or Productivity.
FAQ
Does the platform scope cover all projects?
Yes. The platform Cost, Usage, and Productivity tabs aggregate data across every project in the organization. Project-scope analytics tabs, opened from a project's Settings, cover that project only.
Why are spend totals missing?
Spend totals come from the AI gateway. If the gateway is unreachable, an amber banner appears on Cost and Usage noting the data is temporarily unavailable. Platform activity counts (artifacts, API events, knowledge queries, AI generation runs) come from the application database and remain visible.
Next steps
- Cost: spend by feature and model.
- Usage: token mix, cache efficiency, and platform activity counts.
- Productivity: cycle time, flow health, and quality metrics.