Soundings
Our Perspectives and Insights
AI Made Calling BS the Technical Product Manager’s Real Job
Rise of the Technical Product Manager made a simple case: when AI made the engineering nearly instant, the surface between what the business wants and what engineering builds became the rate-limiting step, and whoever owns that surface holds the most valuable seat in the company.
In that piece, negotiation and validation bookended the skillset list that makes this role work. The real negotiation is between the business that wants and the engineering that builds, and its hardest move is telling either side it is wrong. That is the negotiation combined with validation. You need to know, in the moment, which side to defend and which side (as I’ve always put it bluntly) to call BS on. On a good day, it makes you the least popular person on both sides at once. None of this is new. The technical product manager has always had to call BS. What changed is that AI made it the critical part of the job.
I had the unique experience of both sides pushing at the same time. The business demanded that “everything” be built on a very short timeline, while engineering demanded a full stop to refactor “everything” first. Both were right on a longer time scale, just not the immediate one. So I called BS on both, pushed for the customer to get what they needed now, and laid the stepping stones for what both sides wanted later.
You Work for the Outcome, Not for Either Side
Most people handed a seat in the middle default to being a messenger between two angry mobs, carrying each side’s demands to the other untouched and owning neither. It feels safe, but it is a reliable way to be useless. A real negotiator is not loyal to either party. They are loyal to the outcome that has to hold up after everyone goes home, and the closest analogy of this job is not product clerk, it is the general contractor.
A contractor does not just carry the client’s wishes to the crew. They defend the crew from a client who wants a load-bearing wall gone by Friday. They push the crew when a two-day job is quietly turning into two weeks. And they protect the client from their own worst ideas, because a contractor answers to exactly one thing: the finished house. That is the whole posture of this role. You are loyal to the outcome, not to the loudest voice on either side, and that loyalty is what earns you the standing to call BS on both of them.
Defend the Build
A good technical product manager defends engineering. When a competitor ships something and the business wants the same thing by Friday, someone has to say to the business that the date is a fantasy and here is what is actually possible. Engineering often cannot make that case themselves without sounding defensive or territorial. You can, and it is your job to. That is not being difficult. That is the negotiator taking the arrow so the team can do the work.
Defend the side that is right. Call BS on the side that is not.
And you defend the outcome from engineering too, because the loyalty is to the house, not the crew. When six weeks is padding, or the elegant rebuild is solving a problem nobody actually has, you are the one who says so, because you understand the work well enough to know the difference. This used to be self-correcting. Building was slow and expensive, so a bad engineering direction eventually ran into its own cost. AI took that friction away. It says yes instantly and hands back something clean, confident, and finished regardless of a right or wrong answer. Standing between a business that wants it Friday and a team that needs three weeks requires negotiation and validation: you cannot negotiate honestly for a build you have not checked.
Defend the Customer
Now turn around, because the same negotiation has to point the other way. Plenty of bad direction never comes up from engineering. It comes down from the business as noise dressed up as strategy: a competitor’s move, an executive’s rough meeting, a strong feeling masquerading as a decision, with no evidence at all about what users actually do. You are the voice of the customer in that room, and the customer is not in the room. If you do not negotiate on their behalf, nobody does.
You are the voice of the customer in that room, and the customer is not in the room.
Quibi is the monument to nobody playing that role. Katzenberg and Whitman raised 1.75 billion dollars on a strong feeling that people wanted premium, short-form shows on their phones. They launched in April 2020, and were gone in about six months, because no one made them prove a single user wanted it before they built it. When building was slow that kind of mistake mostly died slowly and cheaply, but AI builds the knee-jerk instantly and at full confidence, so the voice of the customer has to get louder now, not quieter.
Richard Feynman said it best, in a 1974 talk about the ways smart people fool themselves:
“The first principle is that you must not fool yourself, and you are the easiest person to fool.”
That is why the seat needs a negotiator and not a process. The engineer can fool themselves that elegant means right. The business can fool itself that certain means true. You are the one in the middle whose entire job is to not let either happen.
How to Call BS Without Becoming the Enemy
Calling BS is a negotiation, not a verdict. The goal is not to win the point, but to protect the outcome and keep both sides at the table tomorrow. Five moves:
- Restate the ask as an outcome before anyone builds anything. Put the requirement back in your own words, aimed at what it does for the customer: not ‘they want this feature,’ but ‘they need this so they can do X.’ The translation happens in that restatement, and most mistranslations die right there, for free.
- Ask what user behavior proves it. For a business demand, that is the whole game: feelings are not evidence, and what people say they want is not what they do. For a build, ask what result would prove it wrong, then go look for it.
- Defend the team out loud, not just in private. When the pressure on engineering is unfair, say so where the business can hear it. Trust is built the day you take the arrow instead of passing it along.
- Treat confident and fast as a flag, not a finish. Speed plus certainty is the texture of a good answer and a bad one alike. It is the tell, never the proof, whether it comes from the AI or the corner office.
- Be willing to kill it out loud, up or down. Say “this is wrong” to the star engineer and to the executive alike. A negotiator who only ever pushes one side is not negotiating. They are taking sides.
The Bottom Line
A messenger carries the note and answers for nothing in it. A general contractor answers for the house. The best negotiators are not the ones who win every point, they are the ones both sides keep trusting precisely because they will tell either side no. That is the seat, and AI revealed the scarcity. When everyone can build fast, the rare person is not the one who pours the wall. It is the one who can stand between the person who wants it and the person who builds it, tell either of them “that is wrong,” and redirect them to the right path and still have both at the table for the next decision. Defend the side that is right. Call BS on the side that is not. And never forget you work for the outcome, not for the side you happen to be standing on.
A Pilot & Rutter principle.