A technique that improves AI answers by first retrieving relevant documents and then giving them to a language model as context.
Large language models only know what was in their training data, which can be outdated or lack private information. Retrieval-augmented generation (RAG) adds a search step. When a user asks a question, the system finds relevant passages from a knowledge base, such as company documents or blog articles, and includes them in the prompt so the model can answer based on that material.
A typical RAG pipeline splits documents into chunks, converts them into embeddings, stores them in a vector database, and retrieves the closest matches for each query. RAG can reduce hallucinations, keep answers current without retraining, and allow citations to sources. Its quality depends heavily on good document preparation and retrieval accuracy.
The maximum amount of text, measured in tokens, that a language model can consider at one time, including the prompt and its reply.
EmbeddingA list of numbers (a vector) that represents a word, sentence, image, or other item so that similar items have similar vectors.
GroundingConnecting an AI model's answers to verified sources or real data so its responses are based on facts rather than guesswork.
Large Language ModelAn AI model trained on vast amounts of text to understand and generate human-like language, code, and other content.
Semantic SearchSearch that finds results based on the meaning of a query rather than exact keyword matches.
Vector DatabaseA database designed to store embeddings and quickly find the items most similar to a given query vector.
What answer engine optimization means, how AI assistants pick the sources they cite, and the practical steps that help your content get chosen without chasing gimmicks.
AI is reshaping tech roles, but strong fundamentals still win jobs. Here are the skills employers are prioritising in 2026 and a practical plan to build them.
Why AI tools confidently make things up, real cases where it cost companies, and the practical checks your team can apply before trusting AI output.