Spend grouped by owner and workload.
Owner
Team + app
Route
Model + provider
Review
Invoice
New capabilities
Tie spend to the team, project, app, and key behind it, so AI cost can be charged back and defended instead of landing as one unexplained provider bill.
Team
Support
Owner of the support bot spend.
Model
Claude
Model route behind the cost.
Provider
Anthropic
Provider path behind the model.
Break a charge down to the models and providers that drove it, so finance can answer 'what is this?' without filing a ticket to engineering.
Compare owner-level spend across date ranges to find the team or workload behind a month-over-month jump, while there's still time to act.
Who Concentrate is designed for
Cost attribution maps AI spend to the team, project, app, key, model, and provider responsible for it. Without attribution, AI shows up as a single growing provider invoice that nobody can fully explain. With it, every dollar traces to an owner and a workload, which is what makes chargeback, budgeting, and renewal conversations possible.
Spend is tagged with the key and team that made each call, so totals roll up to real owners instead of a shared account.
A charge can be traced to the model route and provider behind it, so finance can explain the bill without engineering reverse-engineering it.
Owner-level totals over a date range give finance the numbers to charge AI cost back to the teams creating it.
Attribution is the basis for token allocation and , so owned spend becomes owned limits.
Feature basics