Knowledge Agent · Use case
How can staff ask questions across SharePoint and Confluence with AI on our own infrastructure?
A knowledge agent answers staff questions from the documents your organisation keeps in SharePoint, Confluence, Google Drive and Notion. Every answer cites the pages it used, so the reader can check it. The agent runs on your own infrastructure with a model you serve, so neither the documents nor the questions leave your organisation.
Knowledge in many places.
Answers nobody can trace.
- ✗ 01Company knowledge sits in several places: SharePoint, Confluence, Google Drive, Notion. Staff search each one, then ask a colleague.
- ✗ 02Public chat tools are ruled out because the documents are internal, and an answer nobody can trace to a source is not trusted.
Where it stops for a person
When the documents do not cover a question.Where the connected documents do not cover a question, the agent says so instead of guessing, and the reader takes it to a person. Guardrails set the subjects it declines.
It answers.
And cites every source.
Step · 01
The sources you choose are connected, and their content is indexed in the Katonic Knowledge Engine.
Step · 02
A member of staff asks a question in plain language.
Step · 03
The agent retrieves the passages that bear on the question.
Step · 04
It writes an answer from those passages only and cites each document it used.
Step · 05
Every question, source and answer is logged for review.
Reads from · connectors
The agent connects to SharePoint, Confluence, Google Drive and Notion through connectors in the Katonic platform.
Acts through · MCP servers
Nothing. It answers and cites. It does not create, change or send anything in any system.
What you set up
- Connect each source and choose the sites, spaces or folders to include.
- Write test questions with known answers.
- Set guardrails for subjects the agent should decline.
Where it runs
The agent, its knowledge store and the model run on infrastructure you own or control. You choose the model and serve it yourself, so questions, documents and answers are not sent to a public AI service.
How you get it today
This page is labelled Use case: the agent is built in Katonic Studio. You describe the job, connect what is named above, add test questions and release the agent once it passes them. When the Catalog holds a tested template for this job, this page will carry the Template label and show how to take it.
Are our documents sent to a public AI service?
No. The documents are indexed on your own infrastructure, and the model that reads them is one you serve.
How do we know an answer is right?
Every answer cites the documents it came from. Before release, each version is scored against your own test questions.
Which sources can it read?
This page describes SharePoint, Confluence, Google Drive and Notion. Other document stores, such as OneDrive and Box, connect in the same way.
Can it update a page or raise a ticket?
No. This agent reads and answers. Acting in another system is a separate build with its own approval steps.