How to handle AI Employee errors or unexpected behaviour
Summary
When the AI behaves unexpectedly, immediately pause it globally or for the specific contact, review the conversation logs, identify the conflicting instruction or bad data, and retrain the AI.
Who this article is for
This article is for administrators and support managers who need to respond quickly to AI issues.
Overview
Before you begin
- Pause the AI if the issue is widespread
- Take over the conversation manually
- Investigate the root cause in the logs
Step 1: Step 1: Intervene Immediately
Step 2: Step 2: Pause Globally (If Necessary)
Step 3: Step 3: Investigate the Logs
Step 4: Step 4: Identify the Root Cause
Step 5: Step 5: Apply the Fix and Retrain
Step 6: Step 6: Re-Enable the AI
Configuration note
We do not guarantee AI accuracy, lead conversion, bookings, customer satisfaction, response quality, revenue, support resolution, or business outcomes. The AI Employee does not fully replace human staff. AI behaviour depends on knowledge sources, instructions, handoff rules, enabled features, connected channels, permissions, configuration, testing, and ongoing monitoring. AI-generated responses should be reviewed before relying on them for sensitive, urgent, legal, financial, medical, billing, or policy matters. The AI Employee may need human handoff for these situations. Knowledge sources must be reviewed, updated, and tested regularly. Usage charges, feature access, fair use rules, or add-on requirements may apply depending on your account setup. Always test before going live. Contact support if an option is not visible. Always have a human monitoring the Conversations tab, especially during the first few weeks after launch or after making significant changes to the AI's configuration.
Important clarification
Assuming the AI is 'broken' and deleting the entire profile. The issue is almost always a specific configuration error. Troubleshoot systematically instead of starting over.
Resolution guidance
The AI got stuck repeating the same message in a loop.
This happens when the AI doesn't know how to handle a specific input and defaults to a generic response. Add a rule: 'If you cannot answer a question after two attempts, trigger human handoff.'
Was this article helpful?
Need account-specific help?
If your question requires account access, billing review, workflow inspection, message delivery diagnosis, AI response review, or private configuration review, submit a support request with examples and screenshots.
