Using AI Assistants for Product Discovery

One interesting thing about adding an AI assistant to a product: you suddenly see how your users think.

I recently added the EcoMoat AI assistant. It’s still very early. I’m still forming my conviction about how valuable the assistant can be for investors.

But one thing became clear immediately. The assistant gives me direct access to what investors are actually trying to understand.

Not surveys.

Real questions. (Anonymous of course, I’m looking at patterns, not people.)

Things like:

 • What are the risks to AppLovin’s moat?

 • How does Novo Nordisk compare to Eli Lilly?

 • Does DocuSign actually have a moat?

 • How do AI tools threaten Adobe?

In other words, investors on EcoMoat aren’t asking: “What stock goes up tomorrow?”

They’re asking: “How strong is this business really?”

From a product perspective, this is gold.

(And yes, gold is up ~75% in the last 12 months.)

First, I can see what investors are actually trying to understand, and how they ask it. Not through surveys or interviews, but through real questions.

Second, with a simple KPI dashboard I can quickly see whether the assistant is actually helpful.

Things like:

 • follow-up rate

 • questions per session

 • repeated topics

 • deep dives into the same company

And when I dive into specific chats, I learn even more, both about how to improve the assistant and how to improve the product UX overall.

Talking directly with users is still extremely valuable. But this kind of instant behavioral feedback loop is something I’ve rarely had in product building.