Raiffeisen: Enhancing knowledge management with generative AI

The Challenge

Generative AI has opened the door to new business applications of unprecedented quality. Large language models, like ChatGPT, have created new ways of interacting with users and handling textual data.

Raiffeisen Switzerland wanted to leverage this advanced technology to capitalise on the collective internal knowledge hidden within its extensive company-specific text database.

The aim was to create a chat solution to accurately answer questions from employees in a natural language. This would be a key step on the path towards more efficient information retrieval and processing.

The Solution

Our cloud and data science team developed an Azure-based prototype of an innovative web application called ‘Ask your documents’.

This application allows employees to ask text-based questions via an interactive chat interface. Answers are then generated automatically based on company-specific documents to ensure a fact-based response. Relevant references are also shown for verification and to provide deeper context. Large language models like GPT-4 as well as Retrieval-Augmented Generation (RAG) are used for these advanced functions.

The Results

The ‘Ask your documents’ application allows employees at Raiffeisen Switzerland to be more efficient in retrieving information and getting answers to their specific questions.

The solution not only improves hit rates, but also the general quality of the answers in comparison to traditional search methods.

Our collaboration with Raiffeisen Switzerland and the development of the ‘Ask your documents’ application has created a solid foundation for future use cases in the field of Gen AI solutions.