Skip to content

AI Companion

Add AI knowledge

Add text, files, and OCR text so AI can answer from workspace knowledge.

For: Owners and admins

  1. Add useful workspace content

    Text notes, links, files, videos, and image cards can all help AI understand the workspace. Clear titles, folder names, and notes make answers easier to find.

  2. Use notes for long training material

    Paste scripts, training guides, FAQs, product notes, or long documents into notes when you want AI to use the exact wording later.

  3. Turn on Image OCR when needed

    Image OCR extracts readable text from image cards and saves it as OCR text or image notes. This makes image-only flyers, screenshots, and posters searchable by AI.

  4. Choose an OCR model carefully

    Image OCR uses the saved AI vision model and consumes AI tokens. Pick a cost-effective vision model, then run OCR in batches instead of rerunning every image repeatedly.

  5. Let the index finish

    After content or OCR text is ready, the knowledge index processes it for the owner workspace. Once indexed, owners, admins, and direct members can all use the same knowledge without re-uploading it.

If something looks wrong

  • OCR is useful for text inside images, but normal typed notes are cheaper and easier for AI to read.
  • Review extracted OCR text before relying on it for important answers.
  • If an answer misses a resource, check that the content has useful text and that indexing has completed.