Architecture
RAG (Retrieval-Augmented Generation)
Technique connecting an LLM to your own data for accurate answers.
Glossary entry
Definition context
Reference notes
RAG (Retrieval-Augmented Generation)
Technique connecting an LLM to your own data for accurate answers.
Architecture
Category
2
Examples
3
Use cases
Technique connecting an LLM to your own data for accurate answers.
Retrieval-Augmented Generation (RAG) is an architecture that enhances LLMs by allowing them to query an external knowledge base before answering. Instead of relying solely on what it ‘memorized’ during training (which may be outdated), RAG searches your PDFs, databases, or web, and uses that fresh, verified information to generate the response.
How Interlinked uses it
Retrieval in Interlinked is grounded in structured, editable sources: your product catalog, branches, business hours, policies, and workflow definition. When a customer asks about availability, price, or a policy, the agent retrieves from the configuration you own in AI Config — not from a vector store built behind your back. Updating the answer is as simple as editing the field.