Me, Myself and (A)I - Deploying a Persona-Based Agent When Real Users Don’t Exist

What if you could talk to your product personas—and they will talk back? At LSports, where I lead Gen AI and analytics products, we ran into a classic product problem: we had bold, innovative features, but no real-time users to validate them. Waiting for organic feedback was too slow, too late. So I built an AI-powered persona agent—a conversational tool that lets any product manager, even without coding skills, test ideas, ask questions, and simulate feedback loops with realistic, persona-driven responses. In this session, I’ll share how I transformed this gap into an opportunity, prototyping the agent with Streamlit and LangChain, using AI- powered assistance - ChatGPT, then scaling it into a polished React + FastAPI tool that fits seamlessly into team workflows. In this talk you’ll get a clear understanding of how AI agents can empower product teams to explore new features with speed and confidence, even in zero-user scenarios. We’ll look at how to go from raw idea to working prototype fast, and how to evolve that into a tool your entire team can benefit from. This talk isn’t just theory—it’s a live demo, an open-source launchpad, and a call to action for PMs to stop waiting for validation and start simulating it. This isn’t another AI hype fest. It’s a hands-on story about building with intent, breaking down barriers, and proving that with the right tools, anyone on the product side can become a builder. And a bonus surprise! You can fork this Repo and use it yourself.

Speakers