Skip to main content

ENDORSE 2025 | GenAI for interoperability

· 3 min read
Wouter van den Berg (TNO)
Wouter van den Berg (TNO)
Expert semantic interoperability & Scrum master

What drives our work is the idea that semantic interoperability is not fundamentally a technical problem, but a commons governance problem. As we’ve mentioned before: 'Systems not talking actually represent people not talking.' Keeping that in mind, we are asking the question: how can Generative AI help?

That was the subject I explored in my recent presentation, "Practical Experiences with BOMOS, Semantic Treehouse, and GenAI," at the ENDORSE 2025 conference in Brussels. For those who couldn't attend, I'm happy to share the full recording of the session.

Overview​

This 15-minute talk provides an overview of our vision and practical examples so far in the field of interoperability and AI. It has three parts:

  1. Foundation: the presentation starts by framing interoperability through our socio-technical lens. I introduce the BOMOS framework as our method for organizing the "participatory" work of standardization, and Semantic Treehouse as the digital platform built to support these collaborative processes and the key roles involved.
  2. Demos for two standardization roles: the core of the talk showcases two proofs-of-concept, each designed to assist a specific role in the standardization process:
    1. For the domain expert: a demo of the AI Specification Generator, designed to kickstart the development of a new standard. In my demo I used a prompt for a "Digital Product Passport for concrete" and it quickly creates a complete, structured data model within the platform.
    2. For the functional manager: a demo of the AI Issue Co-Pilot, which assists knowledge managers in handling complex community feedback like issues, questions or new requirements. The assistant can use the rich context already within Semantic Treehouse (like specifications, previous issues, working group notes) to provide analysis and suggest next actions.
  3. Research agenda: finally, I challenge myself and others with the danger of the "horseless carriage syndrome", i.e. are we using AI to optimize old workflows? After which I summarize our research agenda:
    1. more advanced AI-augmented standardization processes
    2. autonomous data integration and negotiation
    3. and enabling evidence-based standards evolution through observability in data spaces.

I'm thankful for the energy and ideas from the other attendees and speakers.

Looking back, the experience reinforced my belief that we are on the right track.

Join the Conversation​

Conversations about AI's role in standardization are still just beginning. I believe the best solutions will come from open dialogue and community collaboration.

So, watch the full presentation above and please let us know what you think. You can join the discussion with our team and other community members on our Discord server or follow our progress on these AI features on our public roadmap. Or just send me an old-fashioned email ;)