An AI architecture that combines retrieval of relevant documents with language generation, improving accuracy and grounding responses in real data.
Retrieval-Augmented Generation (RAG) is a technique that enhances Large Language Models (LLMs) like GPT-4 by connecting them to external, private data sources.
LLMs hallucinate and don't know about your specific business data. RAG retrieves relevant facts from your documentation/database first, then feeds them to the AI to generate an accurate answer.
Browse related definitions in the same glossary category.
AI Grounding
Techniques to anchor AI outputs in verified sources and facts, reducing hallucinations and improving reliability.
AI Hallucination
A phenomenon where a large language model generates false or illogical information but presents it as fact.
Citation Confidence
A measure of how accurately an AI model attributes information to its original source.
Content Provenance (C2PA)
Technical standards for verifying the origin and authenticity of digital content, increasingly important for AI-generated media.
Embeddings
Numerical representations of text, images, or other data that capture semantic meaning, enabling similarity search and machine learning.
Fine-Tuning
Adapting a pre-trained AI model to a specific task or domain by training it on additional specialised data.
Related GEO services, audits, and frameworks that operationalise Retrieval-Augmented Generation (RAG) in commercial execution.
Understanding "Retrieval-Augmented Generation (RAG)" is just the first step. Our team at TwoSquares specializes in technical SEO and digital strategy, helping brands turn complex concepts into measurable growth.