Knowledge-retrieval-assisted generation (RAG) (Glossary)

Knowledge-retrieval-assisted generation (RAG) belongs to the category of artificial intelligence and digital transformation.

Knowledge retrieval-assisted generation (RAG) is an innovative method in which artificial intelligence (AI) specifically accesses existing knowledge in order to generate better and more accurate answers. In contrast to conventional AI models, which answer purely from stored data, a RAG system also actively searches for relevant information in current knowledge databases or documents.

A chatbot in customer service uses this, for example: if a customer asks about the current dispatch times for a specific product, the chatbot filters through internal documents in real time and provides the most up-to-date information - instead of just relying on general, pre-stored answers. This provides the customer with reliable and up-to-date information.

Knowledge retrieval-supported generation therefore combines the strengths of AI with direct access to external data. This improves the accuracy of the answers and opens up a wide range of possible applications, for example in companies, research or customer service. RAG is therefore an important step towards intelligent and trustworthy AI applications.