In this episode, Michael Fembek moves from theory to practice, exploring how useful clusters are actually built.

Building on the previous discussion about why clustering matters, this episode takes listeners into the everyday reality of managing innovation data—where Excel spreadsheets, dropdown menus, and carefully designed categories become powerful research tools. Drawing on years of experience from the Zero Project, Michael explains how thoughtful data structures can prevent inconsistencies, accommodate overlapping categories, and create datasets that remain flexible as research evolves.

Why should every cluster have its own column? What happens when projects fit into multiple categories at once? And how are AI tools like ChatGPT beginning to transform the way we discover patterns in large datasets?

Tune in to learn practical principles for building clustering systems that are robust, adaptable, and genuinely useful for understanding innovation ecosystems.

🔗 Read Disability Innovation Insights on LinkedIn

Note: This audio version is offered to support accessibility and is not intended as a full podcast production.

Human + AI Collaboration Breakdown

Writing, editing, fact-checking, and all editorial decisions done by Michael Fembek (the human).

Data analysis used during the creation of this blog post was powered by Zero Project Responsible AI.

Proofreading was supported by ChatGPT.

Audio versions of this post were generated using Eleven Labs, with an AI voice model of Michael.

The visual uses original photography from the Zero Project Conference 2025, with image editing by Alessandro Gobello.