How to troubleshoot incorrect AI Employee responses

    CategoryAI Employee
    Article typeDocumentation
    Estimated reading time 5 min
    Last reviewedRecently reviewed
    Applies toGEVADE CRM accounts where the feature is enabled
    Access requiredAccess may depend on your user permissions
    Plan/configuration noteFeature availability may depend on your selected plan and account configuration

    Summary

    Check the AI's execution log to see which knowledge source it referenced, then correct the conflicting FAQ, URL, or business rule that caused the error.

    Who this article is for

    This article is for administrators tasked with maintaining the accuracy of the AI Employee.

    Overview

    When the AI gives a wrong answer, it is almost always because it was fed incorrect, outdated, or conflicting information. Systematic troubleshooting will help you identify and fix the root cause.

    Before you begin

    • Locate the specific incorrect conversation
    • Identify the exact question asked
    • Review your current knowledge base

    Step 1: Step 1: Isolate the Error

    Find the exact message where the AI made a mistake in the Conversations tab. Note exactly how the customer phrased the question.

    Step 2: Step 2: Check the Source Log

    If available in your AI settings or simulator, check the log to see which specific URL, FAQ, or text snippet the AI used to generate that answer.

    Step 3: Step 3: Look for Conflicting Information

    The most common issue is contradictory data. For example, your website URL might say a service costs $100, but an old FAQ you uploaded says $80. The AI gets confused. Ensure all sources match.

    Step 4: Step 4: Update and Clarify

    Correct the false information in your knowledge base. If the AI misunderstood a vague concept, add a highly specific FAQ to address that exact phrasing.

    Step 5: Step 5: Re-Train and Simulate

    Click 'Sync' or 'Train' to apply the changes. Then, go to the Simulator and ask the exact same question the customer asked to verify the AI now provides the correct answer.

    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. The AI is literal. If you have a rule that says 'Never discuss pricing' but also an FAQ listing all your prices, the AI may behave unpredictably. Keep your rules logical and consistent.

    Important clarification

    Assuming the AI is 'broken' when it gives a wrong answer. 99% of the time, it is functioning perfectly but referencing bad data provided during setup.

    Resolution guidance

    What if the AI is hallucinating information not in my knowledge base?

    Ensure your 'Business Rules' strictly instruct the AI to *only* use provided context. Add a rule stating: 'If the answer is not in the provided knowledge base, you must reply with the fallback message.'

    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.