Everything here
- What retrieval means when an AI assistant answers from your files
A plain explanation of what happens behind the scenes when a chatbot says it answers from your own documents, step by step.
- A knowledge base is not training: what changes and what doesn't
Giving an AI assistant your documents doesn't teach it. It reads them when asked. Here's what that means for updates, mistakes and cost.
- How documents are split into chunks and indexed
Why an AI assistant can find one paragraph in your file yet miss the one right beside it, and what chunking and indexing mean.
- Embeddings and meaning-based search in plain English
How an assistant can match your question to a document that uses different words, and where this kind of search can still miss.
- What kinds of documents suit a knowledge base, and what to leave out
Which business files help an AI assistant answer well, and which to keep out, including private data, old versions and anything staff shouldn't see.
- Why an assistant still gets answers wrong when it has your documents
The usual reasons an AI assistant gives an old price or an invented answer even though the right file is in its library, and what to look for.
- Keeping a knowledge base current when documents change
What happens to an AI assistant's answers when you edit or delete a source file, and a simple routine to keep the library in step.
- Reading the source links an assistant shows and checking them
How to open an AI assistant's cited sources, tell whether the passage really supports the answer, and what to do when it doesn't.