The terms that come up when an AI answers your team's questions, in two sentences first.
When a small team starts using an AI to answer questions, a handful of terms keep coming up. Each one below starts with a two sentence definition you can quote, then explains what it means for a team of five to fifty people, with examples and the pages we read. We only list terms people really search for.
An AI hallucination is an answer from an AI system that sounds confident and is false, or is not supported by anything the system was given. It happens because language models produce plausible text, not looked-up facts, and many are trained and scored to prefer a confident guess over saying they do not know.
Retrieval-augmented generation, or RAG, is a way of building an AI system so that it looks up relevant passages in a set of documents and uses them when it writes an answer, instead of relying only on what the model learned in training. The idea comes from a 2020 research paper by Lewis and colleagues.
A knowledge base is an organized collection of answers and how-to information that people, or an AI assistant, can search to solve a question without asking someone. For a small team it works only if each answer has an owner, a date it was last confirmed, and a title written as the question people ask.
An approval workflow is a defined path for getting a decision signed off: who asks, who reviews, who decides, and where the result is recorded. In a small company it replaces 'ask the founder in chat' with a short, repeatable route so the same request gets the same answer.
A single source of truth is the one place where a fact is officially recorded, so that every other place points to it instead of keeping its own copy. For team answers, it means each answer lives in one approved spot, with an owner and a date, and nobody repeats it from memory.
To see the terms in practice, read what an AI hallucination is, check a bot with the free honesty test, or start from the answer library of policy templates.