Knowledge

Knowledge is the internal context AI workers use when answering or deciding what to do. On the Knowledge page, teams can organize folders and create text pages, uploaded files, JSON documents, and database nodes. This supports both manuals and structured data such as customer lists, product catalogs, and support histories.

Overview

Let AI workers reference pages, files, JSON, and structured databases.

Basics

Knowledge types

Use folders for organization, text pages for manuals and FAQs, files for PDFs or source documents, JSON for structured documents, and databases for tabular data with fields and rows. Create these from the Knowledge page's creation controls.

When to use databases

Use databases for customer lists, product masters, inquiry histories, FAQ lists, or sales notes where each row should be added, searched, and updated. Define fields with types such as text, number, or date, then add rows, manage columns, search rows, and switch views.

Assigning to workers

Creating knowledge is not enough. In AI worker settings, assign the knowledge each worker should reference. Support workers may use FAQs and response rules; sales workers may use product material and customer databases; accounting workers may use invoice and evidence documents.

Search and maintenance

Use Knowledge search to find pages, files, and database content, and drag items into folders to keep the workspace organized. Update pages and database rows when pricing, specs, policies, or customer information changes. Outdated answers often come from outdated source material.

Checklist

  1. Open the Knowledge page and organize folders, pages, files, and databases by use case.
  2. For databases, define field names and types, add rows, and confirm the data can be searched.
  3. Assign only the needed knowledge to each AI worker and test answers that require registered content.