Activity Log

Every time the AI agent runs — whether it's a Decompose, Refine, or Scope action — the results are saved and accessible for later review. The Activity Log gives you a history of all agent runs so you can revisit suggestions, accept cards you initially skipped, or reference past analysis.

Viewing past agent runs

Agent runs are stored as records associated with the outcome or work card they were run against. You can access them from:

• The AI drawer on the entity's detail page, which shows the most recent run
• The entity's activity log, which tracks all agent actions alongside other changes

Inspecting agent output

Each saved run includes:

• Action type — Decompose, Refine, or Scope
• Timestamp — When the agent was run
• Results — The full output (suggested cards, refinement suggestions, or scope assessment)
• Reasoning — The agent's thinking process that led to the output

Click into any run to see the complete results. For Decompose runs, the suggested cards grid is preserved exactly as it appeared, including which cards were already accepted.

Accepting suggestions after the fact

For Decompose and Refine runs, you can accept or reject suggestions at any time — not just immediately after the agent finishes. Open a past run and:

• Accept individual cards or suggestions that are now relevant
• Preview cards to see their full details before deciding
• See which suggestions were already accepted (marked with a green "Added" indicator)

This makes the Activity Log a living backlog of AI suggestions that you can draw from whenever priorities shift.

Next Steps

• AI Agent Overview — Return to the full overview of AI features
• Decompose — Learn about the most common AI action