Data Strategy & Engineering for Agentic Workflows

You know that feeling that AI workflows aren't worth learning because ChatGPT, Claude, or Copilot keeps mangling context and making things up?

That's not the model failing. It's the standard chatbot setup being the wrong tool for the job. You can force it to work, but you're wasting your time.Moving beyond this begins with data strategy, which sounds boring and technical, but it isn't either. A big part of it is thinking clearly about what data you have, what it means, and how to get the right slice of it in front of an AI so the output is something you trust and you use.

Here's why this is on the MKE DMC agenda: companies are starting to draw a line between people who use AI and people who build with it. Last month, Zapier set a new minimum AI fluency bar for every new hire, published a fluency rubric (attached), and rebuilt employee onboarding around a "builder mindset." Internal AI usage is now at 100%. Several enterprises I've talked to have said the same thing in different words: that capable (using AI) isn't enough anymore.

Building repeatable workflows is what moves you out of the "capable" bucket. Data strategy is step one.Ray Grieselhuber, founder of DemandSphere, will be walking through what that looks like in practice and where to start. No agents, schemas, or SQL required.

• Networking & happy hour at 5pm
• Ray Grieselhuber talks at 6:15pm
• Time to leave at 7:30pm

Beginners welcome! You don't need to be an engineer or data scientist for this talk!