Review intelligent usage breakdown
Use the usage breakdown card to understand how intelligent usage is distributed across models and activities, and drill into recent activity details.
This guide is designed for project manager.
Steps
Step 1: Open the intelligent usage breakdown card
Navigate to the billing or usage area where the Intelligent Usage Breakdown card appears; it loads automatically, showing a summary of your current intelligent usage with progress indicators and key metrics.

Step 2: Review model usage by LLM
In the LLM usage section, review how your usage is split across different AI models (such as GPT-4o and Claude variants), using the labels, icons, and progress bars to see which models account for the most usage.

Step 3: Expand or collapse the LLM usage section
Use the chevron control on the LLM usage header to expand or collapse the section, helping you focus on model-level details when needed and hide them when you want a simpler overview.

Step 4: Review usage by activity type
In the activity breakdown section, compare how usage is distributed across activity types such as Agent Tasks, Agent Chat, Document Creation, Document Assessment, and Revision Assessment, using the coloured icons and labels.

Step 5: Expand or collapse the activity breakdown
Use the activity section’s chevron control to expand or collapse the list of activity types to quickly show or hide this detailed breakdown.

Step 6: Select an activity type to view recent activity
Click on a specific activity type row (for example, Agent Tasks or Document Assessment) to set it as the selected activity; the component automatically loads up to 50 recent activities for that type.

Step 7: Review the recent activities list
In the activity details area, review the loaded list of recent activities for the selected type, including timestamps and other usage context, while a loading indicator appears if data is still being fetched.

Step 8: Refresh the usage data if needed
Use the refresh control on the card header to reload the latest combined usage summary, ensuring the LLM and activity breakdowns reflect the most up-to-date data.

Step 9: Identify and respond to loading or error states
If the breakdown is still loading, wait for the skeleton placeholders to be replaced by real data; if an error message such as “Unable to load usage details” appears, try refreshing the card or returning later.

Related Tasks
Last updated: 2026-01-09