9 interviews. 1 answer. That's how we found our real ICP.

"We don't have time for customer interviews."

You've said it. I've said it. Every product team says it at some point, usually while shipping features nobody asked for at a pace that feels like progress. It's a comfortable excuse. And it costs SaaS companies thousands of dollars a day, because the alternative to talking to customers isn't "no risk," it's guessing with extra steps.

Here's what happened the last time we made the time.

The setup: 9 interviews, one question

At UX Pilot, an AI design tool for product teams where I'm Head of Product, we booked 9 in-depth interviews with Pro-plan users. All senior-level designers. All from different industries, different companies, different backgrounds. The goal was simple to state and hard to skip: understand the real characteristics, needs, and use cases of our ICP, not the one we'd assumed from the persona doc.

Five hours of conversations, total. That's the part worth sitting with. Not five weeks. Five hours.

What we heard, over and over

Here's where it gets interesting. Every single designer, all 9, shared the exact same primary use case. Not a similar theme we had to squint to see. The same one, in their own words, unprompted.

Another designer told us something even more specific: they had completely replaced their entire whiteboard-to-sketches-to-wireframes ideation phase with UX Pilot. Not augmented it. Replaced it.

So what does "the opening of the second diamond" actually mean? In the double-diamond design process, the first diamond is discovering and defining the problem. The second is exploring and narrowing solutions. Nine out of nine designers weren't hiring UX Pilot to help them understand a problem. They were hiring it to generate and explore solution variations, fast, at the exact moment they'd normally be standing at a whiteboard.

That's an ICP. Not a job title, not a company size band. A specific moment in a specific workflow where your product is the obvious answer.

From insight to action

An insight that doesn't change the product is just an interesting anecdote. So we asked the "so what" question until it hurt: if every one of our best users is hiring us for solution exploration, is that flow as fast and obvious as it should be?

It wasn't. Generating design variations and importing components existed, but they weren't easy to discover or use. We made one focused change: we made that exact workflow, the one nine out of nine users had already told us was the reason they paid, easier to find and faster to use.

Five hours of real customer conversations, one shipped change, and a measurable week-over-week lift in the metric that actually pays the bills.

How to run this yourself

You don't need a research team or a quarter of runway to do this. You need an afternoon and the discipline to actually listen instead of pitching. The version we ran:

• Pick your best users, not your newest. We interviewed Pro-plan users, people already paying, already invested, already past the point of being polite instead of honest.
• Ask what job they hired you for, not what features they want. "Walk me through the last time you used this" beats "what should we build next" every time. This is Jobs to Be Done in practice, not theory.
• Listen for the sentence that repeats. One designer describing their workflow is an anecdote. Nine describing the same moment, in different words, is your ICP.
• Ship one change that serves that exact moment. Not a roadmap. One focused change, fast, aimed at the workflow every interview pointed to.
• Measure whether it worked. We watched variation generation and activation. Whatever your version is, track it before and after, weekly, not once.

That last step is the one most teams skip, and it's the one that turns "we think that helped" into a number you can actually defend in a planning meeting.

Why this matters beyond UX Pilot

Nine interviews found the pattern. But patterns fade if you only check them once. The same "very disappointed" segment interviews surface qualitatively, a Sean Ellis survey measures quantitatively, at every user, on a schedule, so you're not relying on memory of nine conversations from six months ago to know who you're building for.

That's the combination that actually works: talk to a handful of the right people to find the shape of your ICP, then track the very disappointed segment over time to see whether you're still serving them, and whether the segment is growing or shrinking as you ship.