Internal Knowledge for Teams

Retrieval-Augmented Generation (RAG) is a technique that grounds AI responses in your own data. Instead of relying solely on the model's training data, RAG retrieves relevant passages from your documents and includes them as context when generating responses. The result is answers that cite your actual files, policies, and data — not hallucinated generalizations.

In GreatChat, projects serve as knowledge containers. Attach up to 20 files and 10 reference links per project, and the assistant searches that scoped corpus first. Ask 'What's our refund policy?' and the answer comes from your actual policy document. Ask 'What did we decide about the API design?' and the response cites your team's design doc — not a generic web result.

Knowledge is scoped to projects, and projects have permissions. Team members see what their project access allows. External guests see only what you share. The agent respects these boundaries — it never cross-leaks between projects or surfaces documents from outside the scope. This makes internal knowledge safe for sensitive materials like contracts, financial models, and HR policies.

The real power is conversational access to structured knowledge. Instead of searching a file server or asking a colleague, you ask the agent in natural language. 'Summarize the Q2 forecast.' 'What are the open action items from last week's retro?' 'Find all references to the competitor analysis in our docs.' The agent retrieves, synthesizes, and responds — with citations you can verify.