Project Donne

Case Study: Project Donne
Client: Orust Kommun, EU Energy Islands
Industry: Governmental
Services Provided: Vision, Enterprise Architecture

Overview

In 2024, Project Donne set out to make an important contribution in sustainability and technology by envisioning an Energy Transition Operating System specifically designed for small islands. The project, inspired by John Donne’s notion that “No man is an island, entire of itself,” aims to integrate and innovate energy systems for small island communities, starting with Orust. Spearheaded by Helhet AI and supported by the EU-funded projects “Energy Islands” and NESOI, the initiative blueprinted its initial operational capability for Orust and identified a backlog for future development.

Technical Achievement: Retrieval Augmented Generative (RAG) AI Architecture

A valuable lesson from Project Donne is the application of Retrieval Augmented Generative (RAG) AI architecture applied to using chatbots in community engagement. This advanced AI system promises a significant evolution current chatbots.

The RAG architecture combines the generative capabilities of LLMs and the precision of retrieval-based systems. Unlike traditional LLMs that generate responses based on a fixed dataset they were trained on, RAG AI dynamically retrieves information from a live external database, ensuring that the chatbot’s responses are both contextually relevant and up-to-date. This ability is particularly beneficial for a small island communities where local, real-time data can significantly influence decisions.

Advantages and Community Benefits

The RAG-powered chatbot offers several advantages:

Contextual Awareness: By integrating real-time data retrieval, the chatbot should provide responses tailored to current events and specific community needs, essential for a community transitioning to 100% renewable energy.
Accuracy and Reliability: The retrieval component could allow the chatbot to pull accurate information, reducing the risk of outdated or irrelevant data that could lead to misinformed decisions.
Scalability: The RAG architecture can easily integrate new data sources and information, supporting the community’s evolving needs without requiring complete retraining of the AI model.

Project Impact

The high level design of the RAG AI chatbot is suitable for any small islands aiming for energy independence and sustainability. The chatbot can become a vital tool in engaging the community, disseminating information, and gathering feedback, thereby fostering an informed and involved public. The reports prepared by Project Donne, outlining the architecture and its benefits, serves as a valuable blueprint for similar initiatives.

Thanks to the visionary leadership of Project Manager Patrik Andersson and Enterprise Architect Neil Williams, along with the continuous support and insights from the entire Project Donne team, the project highlighted the practical benefits of innovative AI applications in critical environmental endeavors.