Add product documents to your Synaptiq knowledge base so the AI can answer customer questions accurately.
Your Synaptiq AI agent is only as good as the information it has access to. The Knowledge Base is where you upload the documents that the AI references when answering customer questions. Every product spec, pricing sheet, FAQ, and case study you upload makes your AI agent smarter and more accurate.
Navigate to Admin > Knowledge or go directly to /admin/knowledge in your browser. You will see your document library with any files you have previously uploaded, along with their processing status and last-updated timestamps.
There are two ways to upload documents:
The fastest method. Simply drag one or more files from your desktop or file manager directly onto the upload area on the Knowledge Base page. A blue highlighted zone appears when you hover files over the page, confirming the drop target.
Click the Upload Documents button to open your system file picker. Select one or more files and confirm. The upload begins immediately.
After uploading, each document enters a short processing queue where Synaptiq extracts the text, splits it into chunks, and generates vector embeddings. You will see a progress indicator next to each file. Most documents process in under 30 seconds.
| Constraint | Limit | |---|---| | Maximum file size | 10 MB per file | | Maximum batch upload | 20 files at once | | Total storage per workspace | Based on your plan tier |
If a file exceeds 10 MB, consider splitting it into smaller, topic-focused documents. This is actually better for AI retrieval because smaller, focused documents produce more precise answers.
When you are migrating an existing knowledge base or onboarding a new product line, you will likely need to upload many files at once. Select up to 20 files in a single batch using either drag-and-drop or the file picker. Each file processes independently, so a failure in one document does not block the others.
For very large migrations (50+ documents), upload in batches of 15-20 and wait for each batch to finish processing before starting the next. This ensures you can catch any processing errors early.
Synaptiq uses a flat document structure rather than nested folders. To keep things manageable, rely on clear, descriptive file names and use the tagging system.
Use consistent, descriptive file names that reflect the content and its purpose:
| Instead of this | Use this |
|---|---|
| doc1.pdf | product-pricing-enterprise-2026.pdf |
| FAQ.docx | faq-returns-and-refunds.docx |
| notes.txt | onboarding-guide-new-customers.txt |
| specs.pdf | widget-pro-technical-specifications.pdf |
Good naming conventions help you manage your library as it grows, and they help the AI understand the context of each document before it even reads the content.
After a document is uploaded and processed, click on it to open the detail view. You can add tags like pricing, technical, onboarding, or product-name to categorize it. Use the tag filter on the main Knowledge Base page to quickly find documents by category.
Product information changes. Pricing gets updated. Features ship. When your source documents change, you need to update the knowledge base to keep the AI current.
/admin/knowledge and locate the document you want to update.Synaptiq removes the old embeddings and reprocesses the new file. The document retains its existing tags and metadata. During reprocessing (usually under 30 seconds), the AI falls back to other available documents for related queries.
Do not upload a new copy alongside the old one. Duplicate documents create conflicting information, and the AI may surface outdated answers.
If a document is no longer relevant at all, such as a discontinued product sheet or a deprecated feature guide, delete it. Keeping outdated documents in the knowledge base leads to incorrect answers.
/admin/knowledge.Deletion is immediate. The document's embeddings are removed from the vector store, and the AI will no longer reference that content. If customers were getting answers sourced from that document, those answers will change after deletion, so verify your key queries still work correctly using the test interface.
Once your documents are uploaded and showing a Ready status, the AI can immediately use them to answer questions. We strongly recommend testing a few representative queries to verify the AI is pulling the right information. See Testing Queries Against Your Knowledge Base for a walkthrough.
Was this page helpful?