AI concepts

To appreciate the AI Service, it is worth noting some important concepts in general AI terminology. Refer to the table below for descriptions of key AI concepts.

AI term	Description
LLM	A Large Language Model (LLM) such as OpenAI's GPT, is used by AI agents to interpret prompts or generate human-readable responses.​​
Embedding	Embedding refers to the numerical representations of text that captures its semantic meaning. This enables searches based on similarity even when different words or phrases are used. This allows LLMs to retrieve and generate more accurate, context-aware results.
RAG	Retrieval-Augmented Generation (RAG) is a technique that combines external information retrieval with a LLM to produce more accurate and up-to-date responses. It uses embedding to perform similarity searches over external sources, retrieving the most relevant context. This context is then fed into the LLM along with the query to generate a more informed answer.
Lang Graph	This is a framework that enables the creation of stateful, multi-step workflows for language models. This allows for complex reasoning, tool use, and memory, across conversational steps.