Knowledge base (AI)
In AI apps, a knowledge base is the collection of documents and data that an assistant or agent searches to answer questions, such as help articles, policies or product docs, kept separate from the model itself.
Also called: knowledge base, KB
By Divya Sharma · Checked Nov 15, 2026 · How we review
A model's built-in knowledge is fixed at training time and does not include your private material. A knowledge base is where you put that material so the system can look it up. In practice it means the source files plus the processed, searchable index built from them.
What is inside
A knowledge base pipeline usually covers ingestion from sources such as a wiki, Google Drive or a help desk, document parsing, chunking, embedding, and storage in a vector database or search index. At query time it retrieves relevant chunks for retrieval-augmented generation. Managed services such as Amazon Bedrock Knowledge Bases package these steps.
Example
An internal HR assistant has a knowledge base of the employee handbook, benefits PDFs and holiday calendars. When someone asks about parental leave, it retrieves the relevant handbook section and answers from it, with a citation.
Keeping it useful
The hard part is not the search, it is the content. Outdated or duplicated documents produce wrong answers that look confident. You need a sync process so changes and deletions reach the index, and access rules so people only retrieve what they are allowed to see, often via metadata filtering.
Related terms
A knowledge graph is one way to structure a knowledge base, as linked entities rather than text chunks. Agent memory is similar but is written by the agent during use, not curated in advance.
Related terms
- Retrieval-augmented generation
- Retrieval-augmented generation (RAG) is a pattern where a system first searches a collection of documents for passages relevant to a question, then gives those passages to a language model so its answer draws on them.
- Vector database
- A vector database stores embeddings alongside their source records and finds the stored vectors closest to a query vector, usually with an approximate index so search stays fast across millions of items.
- Knowledge graph
- A knowledge graph stores information as entities (people, products, places, concepts) connected by named relationships, so software can follow links between facts instead of searching through loose text.
- Agent memory
- Agent memory is the set of techniques that let an AI agent keep information beyond a single context window, such as facts about a user or past work, by storing it outside the model and loading it back when needed.
- Document parsing
- Document parsing converts files such as PDFs, scans, slides and spreadsheets into structured text that software can use, keeping headings, reading order, tables and images instead of a flat stream of characters.
- Chunking
- Chunking is the step that splits long documents into smaller passages before they are embedded and indexed, so a retrieval system can find and return only the parts of a document that answer a query.