Why Am I Doing This?

I live at the intersection of storytelling, art, design, and technology. I have spent years building products, making sense of the unknown, and learning how to read a room before I build for it. That last part matters more than people think.

This is a unique time to be alive. For the first time, you can iterate at insane speeds and either build something beautiful or something terrible and useless. This is especially true with agents in enterprise. Whether the AI product at hand is an agent or a product itself, it is important to understand who the consumer is going to be and why. Without this, we can go nowhere and reach no useful conclusion.

But here is the thing about speed. That same speed that lets you build fast also means more builders will ship whatever they think is right for their understanding of the problem. And for the first time, that speed will make it harder for end users and agents to get to the outcomes they actually need. The ability to build fast is beautiful and it is rare. But it is not enough.

I believe the most powerful tool a builder can have right now is the ability to truly understand the Jobs To Be Done of the person in front of them.

And that is where it gets interesting. How do you bring what feels like a mix of art and science, and quite honestly a bent of mind, into building and iterating AI products? When do you know what to get feedback on? What should you ask? What should your evals look like? How do you ensure your iterations are moving toward what your users want to achieve, rather than just shipping fast and helping no one?

I have spent years developing an intuition for what good feels like. Learning what to look for in a room, how to understand the energy of a space, how and when to lean deeper into a specific note because it is bringing out all the depth that the space is seeking. I often joke that my job is really to orchestrate the energies of the people I am working with, so that I can meet them where they are and understand the said and the unspoken of what is in the room. All of this very naturally prepared me to repeat this in my work with products.

Empathy and Evals is where I think out loud about all of this. The philosophy behind picking up real problems and approaching them with care. The methodology of comparing against what good looks like. The practical learnings from building feedback loops that actually move you closer to what your users need.

Empathy gets you to the right questions. Evals tell you whether your answers are working. Everything I write here lives somewhere between those two ideas.