How to organise information for better AI Employee answers

    CategoryAI Employee
    Article typeDocumentation
    Estimated reading time 4 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

    Structure your knowledge sources using clear headings, bullet points, and direct Question & Answer formats to help the AI extract accurate information quickly.

    Who this article is for

    This article is for content managers and administrators responsible for the quality of the AI Employee's training data.

    Overview

    While AI is smart, it processes structured data much better than dense, rambling paragraphs. Organising your knowledge base effectively drastically reduces the chances of the AI giving incorrect or confused answers.

    Before you begin

    • Review current knowledge documents
    • Break down long paragraphs into bullet points
    • Rewrite complex policies into simple Q&A format

    Step 1: Step 1: Use Clear Headings

    When uploading documents or pasting custom text, use clear, descriptive headings (e.g., 'Return Policy', 'Pricing Tiers', 'Operating Hours'). This helps the AI categorize the information.

    Step 2: Step 2: Format as Q&A Whenever Possible

    The most effective way to train the AI is to provide the exact questions customers ask, followed by the exact answers you want the AI to give. Create a dedicated FAQ document formatted simply as 'Q: [Question] A: [Answer]'.

    Step 3: Step 3: Keep Sentences Short and Direct

    Avoid long, complex sentences with multiple clauses. State facts simply. Instead of 'If a customer wants a refund, they must return the item within 30 days, provided it has not been opened, and they have a receipt,' write: 'Refunds require a receipt. Items must be unopened. Returns must be made within 30 days.'

    Step 4: Step 4: Remove Fluff and Marketing Speak

    The AI needs facts, not sales pitches. Remove adjectives and filler words from your training data. Keep it strictly informational.

    Step 5: Step 5: Consolidate Information

    Ensure that information about a specific topic (e.g., pricing) is located in one place, rather than scattered across five different documents. This prevents the AI from pulling conflicting data.

    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 interprets information literally. If a rule has exceptions, you must explicitly list those exceptions in the knowledge source.

    Important clarification

    Uploading a 100-page employee handbook full of internal jargon and expecting the AI to perfectly answer simple customer questions. Always curate the data specifically for customer-facing interactions.

    Resolution guidance

    The AI is combining answers from two different services. Why?

    Your source documents likely lack clear separation between the services. Use stronger headings or separate the services into different custom text snippets.

    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.