How is it different from enterprise search?

Enterprise search is the retrieval capability across systems. A knowledge base is the curated store being searched. In practice AI platforms provide both.

Does the content have to be written for AI?

No, though structure helps. Clear ownership, dates and explicit supersession make a far bigger difference than rewriting documents in any particular style.

Knowledge Base

A knowledge base is the indexed store of a company’s documents and written decisions that an AI system draws on when answering.

A knowledge base answers "where is it written down". That is genuinely useful, and it is where most companies start, because the content already exists in wikis, drives and handbooks and only needs indexing to become reachable.

Its limit is that documents are only part of what a company knows. A knowledge base can return a policy without knowing it was superseded last quarter, or a spec without knowing the project was cancelled. Closing that gap is the job of the layer above it, which holds how things relate and which of them are still true.

Frequently Asked Questions

It is a strong start and rarely sufficient on its own. It covers written material but not the relationships, ownership and currency that decide which of it applies.