Product Thinking with AI: The Idea Matters More Than the Code

Here's something nobody tells you about vibe coding: the AI will build whatever you ask for.

That sounds like a feature. It's actually a trap.

The "Sure, I Can Do That" Problem

Ask Claude to build you a task management app and it will. Ask it to add notifications, recurring tasks, categories, priorities, labels, integrations, and a calendar view — it will do all of that too.

Three weeks later you have 4,000 lines of code, a dozen half-finished features, and something that technically runs but doesn't actually solve any problem well.

I know because I've done it. Multiple times.

If you don't know exactly what you're building — and more importantly, why — you'll end up with a beautiful mess. The AI is an incredible collaborator, but it's not a product manager. That's still your job.

What Product Thinking Actually Means

I spent twenty years in financial services and mortgage tech as a product person. The job was never "think of features." It was:

These questions don't change because AI can code faster. If anything, they matter more now. When building is cheap, building the wrong thing is the expensive part.

How This Actually Works

When I built Signal — an executive function support tool for a neurodivergent college student — I didn't start by asking Claude to build me an app.

I started with the problem: a young adult in my life has autism and ADHD. They miss assignments because they don't process syllabi the way neurotypical students do. Reminders don't work because they get ignored. Traditional planners overwhelm instead of help.

The product questions:

That last question took ten minutes to answer. Yes, it could. So then I knew: the core idea was viable. Everything else — the token rewards, the escalating reminders, the interface — came after.

Without the product thinking, I would have built a generic college planner. There are hundreds of those. None of them work for him.

Before You Open the Terminal

Every project I start now begins with a conversation — not with an AI, with myself. Sometimes I literally write it down:

Only after I can answer these do I start talking to Claude. And when I do, I share these answers first. The AI builds better when it understands the why.

AI Makes Product Thinking More Important, Not Less

There's a fantasy that AI will replace product thinking. That you can just describe what you want and it appears, fully formed and useful.

The reality is the opposite.

When anyone can build anything, the differentiator is knowing what to build. The people who will thrive with AI tools aren't the best coders — they're the ones who understand problems deeply enough to aim the AI at something real.

Twenty years of sitting in rooms asking "but what problem are we actually solving?" turned out to be the perfect preparation for vibe coding. I didn't see that coming.