AI Search as Digital Public Service: Lessons from Finnish Universities (Drupal + RAG in Production)

https://www.youtube.com/watch?v=D27gGn2kwzA

Summary

TL;DR — Universities and organizations often struggle with fragmented and inaccessible knowledge bases, leading to user frustration and wasted resources. By implementing an AI-powered search solution using Retrieval-Augmented Generation (RAG) with Drupal and Elasticsearch, the University of Helsinki successfully transformed its vast content into an accessible, user-friendly digital public service. This approach prioritizes user experience and provides direct answers, rather than just links, significantly improving information access for students and staff.

Key points

• Organizations possess abundant content, but the primary challenge lies in making it easily accessible and usable for end-users, not in creating more content.

• Confusion and unanswered questions are expensive, highlighting the need for efficient information retrieval as a competitive advantage.

• A RAG-based AI search system, leveraging Drupal for configuration and Elasticsearch for data storage and retrieval, can effectively aggregate and present information from disparate sources.

• Building trust is paramount; the system must provide verifiable answers with clear references, offer traditional search results alongside AI summaries, and indicate when it lacks sufficient information.

• The AI search solution can detect user language, adapt to different user segments, and provide analytics to identify content gaps and improve the system over time.

• The ultimate goal is to create an AI-powered system that not only answers questions but also anticipates user needs and continuously improves, treating clarity as essential infrastructure.

Takeaway — The most successful organizations will be those that leverage AI to make their existing knowledge clear and accessible, treating clarity as a core infrastructure rather than just a feature.