7. What did not change

§ 7.1The chair across the table#

§Someone still has to sit across from a person at four in the afternoon, in the quiet between shifts, and ask what the hardest thing about the day is. And then wait. And then, when the answer is "nothing really," wait again, because the second answer is the one that matters.

§No model does this. A model can transcribe it, summarize it, draft the next question. It cannot be the person the other one decides to trust with the real answer. That decision is made about a face, in a room, and it is the source of everything the rest of the discipline processes.

§In the field sprint the customer interviews were strong and the operations interviews were weak, and the prototype was identical in both. The difference was who was in the chair and how long they could stand the silence. Gary Klein's work on where insight comes from points at the same place: it comes from a person noticing a contradiction that does not fit, and noticing is not a task you can delegate to something that was not there.

§ 7.2The sentence to the room#

§Someone still has to say, to a room that voted for payments, that payments go last.

§The product lead said it in one sentence at the close of the field sprint: prioritize the profile over payments to test the real value proposition, then payments, essential for the business case but not necessarily for the customer. The room had entered with payments as pain number one. Twelve strangers had said otherwise. The sentence let them win.

§That is not a synthesis task or a facilitation task. It is a decision that costs the person saying it something, the discomfort of overruling a room and, sometimes, themselves, and the willingness to pay that cost is the second thing AI cannot reach. The fourth whitepaper called it letting the vote lose. It is the whole job at the moment of commitment.

§ 7.3The short list of what AI did not do#

§In the field sprint, AI wrote copy, drew screens, generated code, produced every document from one brief, and turned a requirements file into a navigable app in two days.

§It did not interview. It did not recruit twelve people in a week. It did not notice that two operations staff barely touched the laptop, or that the reason was the room and not the product. It did not vote. It did not choose to put payments last. It did not sit through the silence, and it did not bear the cost of the sentence.

§ 7.4Where the time goes now#

§If a team can ship a navigable prototype in two days, it has two weeks it did not have before, every time it wants to test an idea. The question this whole essay has circled is what to do with them.

§Not more prototypes. The field sprint is clear: the features added because they were cheap were the features that failed. Spend the time instead on the three things that got harder, not easier, which are the three ends of the discipline the collection is built from.

§Understanding. Recruit from day one. Ask about past behavior, not future intent. Synthesize the same day. Sit in the chair across from the person and wait.

§Evidence. Only build what came from a question. Put it in the hands of strangers. Sort every assumption into validated, open or discarded, with the quote that moved it.

§Commitment. Let the evidence overrule the vote. Close, which means leaving things out. Say the sentence to the room, and write down why, so the reasoning survives the decision.

§ 7.5The six whitepapers, in order#

§This collection is written in the order a team moves through the work, and it can be entered wherever a team is stuck.

§Product Foundation is the map: what product is at a startup, who holds it, what the owner is accountable for. Open it if nobody owns product yet, or if everybody thinks they do.

§Product Discovery is the judgment: the methodologies that shaped product, what each got right and wrong at startup scale. Open it if you have a process and keep building the wrong thing anyway.

§Product Sprint is the practice: from a first conversation to a written position, insight and then evidence. Open it if you are starting discovery this week.

§Product Strategy is the commitment: from evidence to a ladder, bets, and a first unit of work. Open it if you have evidence and cannot agree on what it means.

§Business Design is the model: desirable, feasible, viable, and what has to be true. Open it if people love the first version and you are not sure it closes as a business.

§Product Experimentation is the loop: validation, then growth, one discipline at two stages. Open it if you have shipped and are not sure what to test next.

§ 7.6The last thing#

§Building got cheap. Thinking didn't.

§That is not a complaint about AI. It is the best thing that has happened to product work in twenty years, because it took away the excuse. There is no longer a two-week gate between an idea and a stranger's hands. There is only the question of whether you wrote the question down, whether you found the stranger, whether you sat through the silence, and whether you let what you heard win.

§That was always the work. The cost of building was hiding it, and now the cost is gone and the work is standing in plain view. Everything in this collection, the six methods and this essay, is about the part that did not get cheap, because the part that got cheap will take care of itself.

§That sentence is the whole collection. A team went and found out. What it found overruled what it wanted. And someone had the discipline to say so out loud, and the humility to let it stand. No tool will ever do that for you. It is the one thing that was always, and remains, entirely yours.