This session explores implementing secure, content-controlled AI assistants within institutional settings for enhanced support and learning.
This session shows the development and and integration of content-controlled AI-assistants designed to operate with pre-defined content. We will discuss approaches like Retrieval-Augmented Generation (RAG) using institutionally controlled knowledge bases, such as research papers, internal wikis, and curated support documentation, to ensure reliable and contextually appropriate responses.
The approach will be showcased on two examples: a chatbot for IT service support and an interactive learning assistant.
