Categories
Artificial Intelligence Data Architecture

SEDA: Building Evolving Ontologies, Now Available in ChatGPT

Building SEDA, an agentic system for evolving industry ontologies, from data architecture practice to a ChatGPT plugin and a shared World Ontology ambition.

I wanted to build a system that uses agents to create highly usable industry ontologies, and to keep those ontologies evolving.

This became SEDA: Self Evolutionary Data Architecture.

An ontology gives a domain a shared vocabulary. It describes the concepts, their meaning and the relationships between them. My objective with SEDA is to make that knowledge useful, and to give it a structure that can evolve as our understanding of the domain grows.

From industry ontologies to a World Ontology

There is also a larger idea behind this experiment: could we share this knowledge through a World Ontology?

A single view that brings together good data architecture practices and ontologies built collectively. A way to capitalize on what we learn across industries, rather than starting again each time.

This is the direction I want to explore with SEDA.

Working headless, with a system of record

To test this way of working, I wanted to operate SEDA through a ChatGPT plugin, in headless mode.

SEDA has an MCP server that makes it usable from ChatGPT. The conversation becomes the place where I work with the system, while SEDA provides the system of record and a central harness for keeping the evolution of the ontologies coherent.

For me, the point is to bring together the flexibility of working with agents and the discipline needed to maintain a data architecture over time.

The practices behind the system

The data architecture practices I bring into SEDA are those I have accumulated over the years. In particular, I learned a great deal from John Sullivan, a mentor whose thinking has strongly influenced mine, and from the excellence in Data Management I had the opportunity to practice with my former team at ING.

That experience is part of what I want to carry into this system. The tools evolve. The knowledge we have built together can help us use them well.

Published, and ready for wider testing

The plugin is now available as SEDA: The World Ontology in ChatGPT. It is time to test it on a larger scale.

Publication confirmation for the SEDA plugin in ChatGPT
The publication of the SEDA plugin in ChatGPT.

This publication opens the next part of the experiment: seeing how the approach holds up through broader use, and learning from it.

The Build → Learn → Lead loop has never been this fast.

Discover SEDA in ChatGPT · Explore the SEDA platform

Related: Architecture methods and resources · My projects

By Yannick Huchard

Technology is my Business, Complexity is my playground, Strategy is my algorithm, precision execution, proven results.

I am devoting my life to knowledge, craft, and empowering people who want to do better for themselves and their surroundings.

About Yannick Huchard · Talks & Media

Leave a Reply