How Documents Are Processed

After you upload a document, VoxeDesk processes it automatically so the AI can search it. Here's what happens at each stage.

Processing stages

1. Extracted — text is pulled from the file. For PDFs, this requires a text layer to be present.
2. Chunked — the text is split into smaller pieces so the AI can retrieve the most relevant section instead of the whole document.
3. Embedded — each chunk is converted into a format the AI can semantically search.
4. Indexed — the chunks are stored and ready for retrieval.

Status meanings

• PENDING — the document has been uploaded and is queued for processing
• PROCESSING — currently being processed (do not test yet)
• COMPLETED — ready to use; the AI can now search this document
• FAILED — something went wrong during processing

If a document shows FAILED

• Check that the file is not corrupted — try opening it locally first
• Scanned PDFs without a text layer will fail to extract meaningful content. Convert to a text-based format (DOCX or TXT) and re-upload.
• Very large or complex files occasionally fail. Try splitting the document into smaller files.
• If the issue persists, contact support with the file name and format.

What's next

• How to Link a Knowledge Base to a Workflow — connect your knowledge base to the AI
• Best Practices for Knowledge Bases — get more accurate AI responses