Execution History

Every action an AI agent takes in BFF is logged and visible in the management UI. This provides full transparency into what your AI team members are doing.

Viewing Execution Logs

  1. Navigate to AI Agents
  2. Open the agent’s detail view
  3. Go to the Execution History tab
  4. See a chronological list of all executions

What’s in the Logs

Each execution log entry contains:

FieldDescription
Task NameWhich task was being executed
Start TimeWhen execution began
End TimeWhen execution completed
StatusSuccess, failed, or in-progress
Steps CompletedProgress through the Work Instruction steps
Step DetailsWhat happened at each step
ErrorsAny failures or issues encountered

Step-by-Step Progress

For tasks with structured Work Instructions, the execution log shows per-step detail:

  1. Step 1: Pull data — ✅ Completed (2.3s)
  2. Step 2: Transform data — ✅ Completed (1.1s)
  3. Step 3: Generate report — ✅ Completed (4.7s)
  4. Step 4: Send to client — ❌ Failed (API timeout)

Filtering and Searching

  • Filter by date range — see executions from a specific period
  • Filter by status — show only failures or successes
  • Filter by task — see executions for a specific task

Using Logs for Troubleshooting

When an agent task fails:

  1. Open the execution log for the failed run
  2. Find the step that failed
  3. Review the error details
  4. Check if the Work Instruction needs updating
  5. Fix the issue and re-run

Tip: Regular log review helps you optimize AI agent performance. Look for steps that consistently take longer than expected — they may need better instructions or different tooling.