Building Coach Bob’s Playbook: How Market Research Lessons Emerged from My AI Avatar Endeavor

A personal AI avatar project took an unexpected turn into football — and ended up teaching me lessons I now use every day in research.

From Bob to Coach Bob

The Building Bob project began as a personal endeavor — part curiosity, part fun. He started as a browser-based avatar that blinked, talked, and answered questions with a personality shaped by GenAI. It was whimsical, a little playful, and a reminder that tone and framing matter just as much as the model behind them.

But eventually I hit a point where Bob needed to do more. And as fall rolled in and my other side passion — college football (We Are!) — took center stage, I decided to venture into a side quest: Coach Bob.

That’s where things got interesting, because I started to see parallels that extended well beyond football — straight into the research world I live in every day.

From Bob to a Playbook

To actually do something, Bob needed to:

• Call plays for his team in a predictable way
• Field a roster of players with their own strengths and weaknesses
• Face rivals with distinct styles and personalities (post-game interviews included)

At first, it seemed simple enough to extend Bob’s existing GenAI system. After all, I was already using AI prompts to shape his personality, and AI text-to-speech to give him a voice.

But the deeper I went, the clearer it became: this wasn’t just about extending Bob. Each step revealed lessons about how different AI approaches succeed or fail — lessons just as relevant to market research as to football.

Why This Matters for Research

For me, these projects aren’t just tinkering (though I do love to tinker!). They’ve been a way to explore AI hands-on and see where certain methods shine — and where they fall short.

And they mirror what I see in research every day:

• Some problems need generative creativity.
• Some need deterministic predictability.
• Some need synthetic data to add realism.
• And often, the most useful perspective comes from contrasting personas.

The bigger point? The magic isn’t in the box. It’s in choosing the right tool for the right moment — and sometimes in combining them in innovative ways.

The Road Ahead

Over the next few articles, I’ll share the lessons I learned along the way:

• Right Tool for the Job: Why deterministic play calling beat GenAI for Coach Bob.
• Synthetic Makes It Real: How Bob’s draft class of players showed me the value of synthetic data.
• Personas in Play: Why rival coaches matter, and how contrasting styles reveal more than a single “perfect” model ever could.

Together, they tell the story of how a personal side project taught me something directly relevant to my professional work: innovation doesn’t come from chasing the newest tool, but from knowing how to use the right tools together.