Monday morning. Three features were built over the weekend. They're sitting in the test system and look good. Nobody quite knows who ordered them. Sales wants a demo. The boss has seen a competitor's release video.
This scene comes from a talk at the Just Product conference, and it's not just a problem for product teams. I can build something in an afternoon today too. Whether anyone needs it isn't answered one bit faster.
The bottleneck moves
A product manager there told how her team switched to agentic development. Her backlog emptied before her eyes, the developers kept asking what's next. Suddenly product was the bottleneck again. The team abolished sprints and estimates because nothing could be estimated meaningfully anymore. A different question became more important: what's the actual goal?
That's the experience of one person and one small team, not a general number. As a pattern, I still find it very plausible: First building was the bottleneck, now it's deciding what gets built. And after that comes the next one: the people who are supposed to adopt it.
Shipped isn't adopted
Another speaker described the shift that's needed now: away from “done means shipped”, towards “done means adopted”. Sales, rollout and retention don't come after building. They're part of building.
How you notice you're in the trap:
- The feature goes live, confetti rains in the chat, and three months later nobody uses it.
- The launch is the last line in the plan.
- Rollout is “someone else's job”.
The 2024 study by Washington State University fits here: when “artificial intelligence” appears in a product description, purchase intent drops for many people. The authors advise showing the benefit instead of the technology. People don't adopt technology. They adopt relief.
The fire hydrant
The image from the talk: the team ships feature after feature at light speed, and customers can't keep up anymore. They dismiss “What's new” every day. Sales learns about new features from customers. It's like giving someone a drink from a fire hydrant.
That applies on a small scale too. Building too many offers, freebies and courses at once creates the same effect, just more visibly. Your customers have a day job. What you build should help, not distract.
Building to learn
- Adoption is part of building. “Who adopts this, and how?” belongs at the beginning, not at the end.
- Less at once. One offer people can really adopt beats five finished ones.
- A quickly built feature is a test, not a result. Building fast is fine if you bring the discipline to learn from it.
Done is when someone has adopted it.
Take something you built that's live. How would you notice it's being adopted? Does that number or sign already exist?