AI changed how software gets built. A feature that took a quarter can take a week. A clear customer need can be in production the same day it’s heard. But deciding what’s worth building still happens the old way: a few customer calls, a spreadsheet of requests, and whoever argues best in the planning meeting. Doing it properly takes weeks, and engineering won’t wait that long.
So teams can now build the wrong things faster.
Agentic product discovery closes that gap: decisions as fast as building, still backed by evidence.
What it is
In agentic product discovery, agents work through customer signals around the clock, so every decision starts from what’s known.
Every call, ticket, chat, and usage pattern is read as it comes in. All signals are connected to the plans you are already tracking, and new plans are created when a clear pattern appears. Every plan includes who needs it and how much, who doesn’t, what is still unknown, and what is expected to change if you build it. Your new ideas can be researched in minutes.
Agents do the groundwork. People set the direction, talk to customers and make decisions.
What we believe
- 1
Start with what you already have.
No one has perfect customer input to begin with. Most B2B teams already talk to customers every day, so start from there. The quality improves as you go, once people know which questions matter and you fill the gaps on purpose.
- 2
Who doesn’t need it matters as much as who does.
Twenty customers asking for the same thing doesn’t tell you much on its own. Counting signals is easy. What matters is understanding the problem behind them, what it costs each customer and how much revenue is at stake. It matters just as much who says the problem doesn’t apply to them.
- 3
A good “no” counts as much as a good “yes”.
Most ideas shouldn’t be built, and most requests shouldn’t become features. Everything you ship has to be supported and maintained for years. Saying no or not now is a decision in its own right. Write down why, so the next person doesn’t have the same discussion from scratch.
- 4
Every launch is a bet.
Before anything gets built, there should be a clear, written expectation of what will change and for whom. It’s the hardest part of product work, which is why it so often gets skipped. When you start, you’ll miss most of your targets. That’s how you learn, and over time you start hitting the mark.
- 5
People decide.
Agents can read everything and recommend what to do. But they don’t own the strategy or talk to customers, and they don’t carry the risk. Decisions stay with people, and that only works if every claim can be traced back to the original customer evidence.
- 6
Improve the system, not just the decision.
Every decision teaches you something, whether it worked out or not. Feed that back into how you work. Over time you ask better questions and weigh evidence better. Do it together and every team gets the benefit.
How it works
What changes for product teams
Priorities you can defend.
Every priority comes with who needs it, who doesn’t and why. When sales ask why their deal-breaker isn’t on the roadmap, or the CEO asks why something else comes first, the answer is already there.
Good habits stick.
Agree once that every plan needs a clear expectation, or that every idea gets checked against your principles, and it happens every time. Agents don’t get busy, skip steps or forget what the team decided.
Product work you can be proud of.
Most product people didn’t get into this to ship whatever got asked for last. Working from the full picture lets you build the thing that actually moves the product forward.
Your time goes where it counts.
The reading and sorting is taken care of. That leaves your time for the conversations that fill real gaps, and for deciding which unknowns are worth the risk.
What stays human
Choosing the direction.
People set the strategy, the goals and the principles the agents work from. Agents can point out where a plan drifts from them, but they don’t decide where the product goes.
Talking to customers.
Agents read every conversation, but the ones that fill real gaps need a person. Earning trust, asking the follow-up question and reading the room stay with the team.
Deciding what’s worth building.
Agents compare the options and show what’s known and what isn’t. Which risks are worth taking is a judgment people make, and own.
That’s what product people are for. Agents make sure nothing gets in the way.