RAG (Retrieval-Augmented Generation)
A technique that retrieves relevant documents before generating a response, grounding the AI in factual business data.
Retrieval-Augmented Generation solves one of the fundamental problems with LLMs: they can only draw on their training data, which knows nothing about your specific business.
The process works in three steps. First, the caller's question is converted to an embedding vector. Second, this vector is compared against your knowledge base chunks to find relevant information. Third, the retrieved text is included in the prompt, giving the model factual context.
RAG is what allows an AI voice agent to accurately answer questions about your pricing, hours, services, and policies. Without RAG, the agent would either refuse to answer or hallucinate incorrect information.