Flipping the order so design and applied AI prototype before the handoff, putting the hardest decision next to something real.
Prototyping used to come after the decision. Product would size the opportunity, write the PRD, hand it over, and then design would make it real. We’re flipping that order, and so far it’s the most promising change to how we work together this year.
The new sequence is roughly this: design and our applied-AI folks prototype something real first, before the formal handoff, and they go all in on it. They put a working artifact on the table that engineering and product can actually react to, instead of a paragraph everyone interprets a little differently. Then the conversation happens around the thing. Product’s job shifts toward the value judgment, the market sizing, the revenue case, the question of whether this is worth doing at all, and away from prescribing the solution path. Design and engineering iterate on the how, often without product in the loop for every turn.
What I like about this is that it puts the highest-leverage conversation, is this even the right thing, next to a concrete artifact instead of an abstract brief.
A doc forces everyone to imagine the same thing and hope they’re imagining it the same way. A prototype just shows you.
When it’s good, you’ve saved weeks. When it’s wrong, you find out in a room with something you can point at, which is a much better place to be wrong. The honest caveat is that this only works when the right people are actually there to prototype together. The first time we tried it, the applied-AI partners weren’t in the room, and the questions we most needed to answer just sat open. So the model is real, but it has a prerequisite: the people who can make the thing have to be at the table from the start, not invited to react after the fact.
The teachable part
Put the made thing before the decision. Whether something is worth doing is easier to answer next to a prototype than a paragraph.