Ferryt Chat 2.0 (RAG): A Chat System with Retrieval-Augmented Generation (RAG)

The Artificial Intelligence Center of Adam Mickiewicz University (UAM), in collaboration with DomData, has developed an intelligent chatbot module based on the company’s internal documentation.

This project included an analysis of different language models for the Polish language and an evaluation of response quality. The system is designed using the Retrieval-Augmented Generation (RAG) architecture, enabling it to analyze internal company knowledge and documents to provide detailed and contextually relevant answers.

Key features of the chatbot:

* Real-time access to the latest enterprise data, ensuring up-to-date responses.

* Data security and controlled infrastructure, preventing unauthorized access.

* Transparent knowledge sourcing, allowing users to trace answers back to the specific documents used by the model.

This AI-driven chatbot enhances internal knowledge management while maintaining data security and ensuring that information remains within the organization’s controlled infrastructure.