Mobility
ANWB
RAG knowledge assistant for member services
An internal knowledge assistant that surfaces the right passage from the knowledge base instantly, with source attribution for verification.
Results achieved
Shorter handling time
Consistent answers
Less cognitive load
The challenge
The right instruction was usually in the internal knowledge system, but finding it cost time: inventing search terms, scanning multiple documents and consulting colleagues in case of doubt.
Our approach
- 1.Inventory of common information needs and sources
- 2.RAG pipeline on the existing knowledge base
- 3.Source attribution on every answer for verification
- 4.Iterative improvement based on staff feedback
The solution
A RAG bot where staff ask questions in natural language and get the relevant passage from the knowledge base, with sources, instantly.
PythonRAGLLMsVector search