Reviewing Employee Tasks

Activity Log Overview

Every action your AI Employee takes is recorded in the activity log. This includes conversations from all channels, tool executions, routine runs, and any errors encountered. The activity log is your primary tool for understanding what your AI Employee is doing and verifying its work.

Viewing the Activity Log

Access the activity log from the Activity tab on any AI Employee's detail page. Each entry shows:

• Timestamp — When the activity occurred.
• Type — Conversation, tool execution, routine, or system event.
• Channel — Which communication channel the interaction came from (or "Platform" for built-in chat).
• Status — Completed, failed, or pending.
• Summary — A brief description of what happened.

Click on any entry to expand the full details.

Understanding Task Details

When you expand a task, you can see the complete interaction:

• The original user message or routine trigger.
• The AI Employee's reasoning process — what it decided to do and why.
• Tool calls — which tools were executed, with what parameters, and what results they returned.
• The final response sent to the user.
• Any errors or warnings that occurred during processing.

This transparency is critical for building trust in your AI Employee's work. You can verify that it is accessing the right data and making appropriate decisions.

Task Statuses

Completed

The task finished successfully. The AI Employee processed the request, used the necessary tools, and delivered a response. This is the expected status for most tasks.

Failed

Something went wrong during execution. Common causes include integration connection issues, permission errors, or unexpected data formats. Failed tasks include error details in the expanded view.

Pending

The task is still being processed. Most tasks complete in seconds, so a pending status that lasts more than a minute usually indicates an issue.

Resolving Failed Tasks

When you find a failed task, follow these steps:

1. Expand the task to read the error message.
2. Check if the error is related to the integration connection — if so, verify the connection status on the Integrations page.
3. Check if the error mentions a specific tool — the tool may have been disabled or may not be available for your integration.
4. If the error is a timeout, the request may have been too complex. Try breaking it into smaller parts.
5. For persistent errors, check the Monitoring guide for common issues and solutions.

Retrying Tasks

For failed tasks that were caused by temporary issues (like a brief connection interruption), you can retry by sending the same request again through the chat interface. There is no automatic retry for failed conversations — this is by design to prevent the AI Employee from making repeated attempts that might create duplicate records.

Routine failures are handled differently. If a routine run fails, the system notes the failure but continues to run the routine on its next scheduled time. If a routine fails three times consecutively, it is automatically paused to prevent repeated errors. You will receive a notification when this happens.