CortexDocumentation
Portal · Spend

Spend

Break estimated captured AI spend down by provider, team, or workstream and interpret changes with coverage in view.

  • Estimated captured costPlanning signal, not provider billing
  • Flexible breakdownsProvider, model, team, and workstream
  • Evidence behind changeOpen the work and sessions behind an outlier
Cost drivers in context

Spend estimates the model cost of activity Cortex captured. Use it to find cost drivers, compare like-for-like windows, and open the work behind an outlier.

What Spend measures

  • Captured activity only: Work outside supported or active capture paths is absent.
  • Estimated cost: Model usage is priced for analysis and planning, not billing.
  • Selected scope: Visibility, filters, and date range shape the total.
  • Multiple breakdowns: The same total can be grouped by provider, team, or workstream.

Use the controls

ControlWhat it changes
Date rangeWhich captured activity contributes to the total
ProviderWhich model providers or models are included
TeamWhich organizational groups are included
WorkstreamWhich outcomes and work units are included

State the date range and breakdown beside any figure you share. Changing a breakdown regroups the total; changing a filter changes what is counted.

Interpret a change

  • Open the provider or model behind a spike.
  • Check whether retries, long contexts, or tool calls explain the cost.
  • Move into the related workstream or session for evidence.
  • Confirm Coverage stayed stable across the comparison window.
Estimated spend is not a provider invoice and is not your organization's complete AI bill.

Continue from Spend

Was this helpful?