As Co-Founder and CEO of #PoDojo, Stefan Haas helps organizations transform their product development through AI-driven customer understanding. Stefan Started in AI research at DFKI in the early 90s, working on computational linguistics and symbolic AI systems. He has spent the following 30 years in product transformation and agile coaching. Since ChatGPT's breakthrough, he's been exploring how generative AI enhances product discovery.

The Workshop

From Idea to Insight with AI Agents

Build is not the bottleneck anymore — it's discovery. With AI coding tools, teams build and ship 10x faster. That means you can work on 10x more ideas. But where do those ideas come from? How do you test them fast enough? And when you run far more experiments — how do you make sense of all that data?

Most teams stop at AI prototyping. In this workshop, we take you through the complete discovery cycle — ideate, test, and synthesize— using Claude Code as the agentic runtime, Skills to encode your domain expertise, and MCPs (Model Context Protocol) to connect your data sources.

You'll see how we're helping Bolt's UX Research team go from research question to validated insights in a single day. You'll learn how to use AI for rapid prototyping, automated user testing, and data synthesis — all orchestrated through Claude Code Skills. And you'll take your first hands-on steps building your own agentic discovery assistant.

You leave with: Your first working set of Claude Code Skills, connected MCPs, and the blueprint to scale an AI-powered discovery system with your team.

What you Will Learn

• Understand why AI-accelerated development shifts the bottleneck to discovery — and what a complete AI-powered discovery cycle looks like

• Ideate: Use Claude Code Skills as an AI thinking partner to generate and evaluate ideas in the context of your business model

• Test: Leverage AI prototyping, automated user interviews, and synthetic users to run more experiments faster — orchestrated through MCPs connecting your data sources

• Synthesize: Build Claude Code Skills that analyze qualitative and quantitative test data into actionable insights at scale

• Apply the architecture of Claude Code + Skills + MCPs as an integrated discovery system

• Exchange practical experience and learn from real customer projects

• Bonus: Follow-up support for teams who want to scale the approach using Claude Agent SDK