Spend quality, not just spend volume
Where Usage tracks instance awake time (infrastructure) and Costs tracks AI token spend by provider and model, Productivity asks the harder question: was the spend worth it. The strip at the top keeps the score in four numbers: 30-day spend, the share of work not written off, agent awake hours over the week, and whether caps are on.

Receipts and attribution
Every value claim on this page traces back to the Costs ledger: receipts and attribution flow to Costs, so the number that says the work was worth it and the number that says what it cost are the same books. A growing bill and a growing return are never assumed to be the same thing.
The four spend lenses live as tabs on one page: Usage, Costs, Productivity, Limits. Usage answers who is on and when, Costs answers what the tokens cost, Productivity answers what it produced, Limits keeps all three honest.