The briefing looks complete. The model is confident. The slides are clean. Everybody in the room is a little relieved because the hard part appears to be over.
Then one person says the sentence nobody wants to hear.
That’s wrong.
Not “interesting.” Not “let’s circle back.” The actual judgment call. The number doesn’t hold. The assumption is off. The recommendation sounds good because it was built to sound good.
That person is becoming scarce.
A lot of shops spent the last two years getting people to use the tools. The ones pulling ahead now are the ones who know when to reject the product. Open commentary from leadership research keeps landing on the same cluster. Judgment. Discernment. The willingness to challenge an answer that arrives looking finished. The novelty wore off. Access is no longer the advantage. Knowing when the machine is flattering you is.
I’ve been in rooms where the data and the human reality were not the same map. You could see it if you watched the people instead of the deck. Someone agrees too fast. Someone else goes still. A jaw sets. In my Profitable Body Language book I kept coming back to clusters, not single moves. One nod means nothing. A nod plus a body angling away plus a voice that gets quieter is a different story. The briefing says the path is clear. The room is not buying it. The general who only reads the packet misses that.
Sun Tzu put the commander in the five constant factors for a reason. Method and discipline matter. Terrain matters. None of it saves you if the person in charge cannot tell a sound report from a pretty one. He also treated deception as the normal condition of war. That includes self-deception. A tool trained to be helpful will often rather agree with you than correct you. Researchers have been writing about that sycophantic loop in plain language. You bring a shaky theory. The system finds facts that fit. It stays quiet about the ones that don’t. You walk out more sure than you walked in. That is not analysis. That is a polished mirror.
I covered a version of this in my Redirecting Rejection book. The first frame you accept is the one you work inside. If the first frame is an AI draft that already picked a side, you spend the meeting defending a position you never actually chose. The move is to break the frame before it hardens. Ask what would have to be true for this to be wrong. Then sit there until somebody answers.
If you lead, stop rewarding the fastest packet. Reward the person who can kill a bad idea without making a speech about it. Ask where the recommendation is weakest. If the briefer cannot show you the weak joint, they did not do the work. They assembled a case.
Watch who still argues. Not the performer who objects to look sharp. The one who will slow the room down over a number that doesn’t sit right. That used to get people labeled difficult. In this environment it is one of the few signs you still have a mind in the seat.
If you are the briefer, build the habit of writing the contrary case first. Two paragraphs. What would a skeptical operator say. If you cannot write it, you do not understand your own recommendation yet. And read the thing before you stand up. If a sentence only works because the tool liked the rhythm, cut it.
Sun Tzu’s winning commander decided the outcome before the clash. That does not mean they ignored new information. It means they did not outsource the last judgment to a voice that cannot be held responsible when the field changes.
The tools will keep getting better at sounding sure. The scarce skill is still the same one. Look at the packet. Look at the room. Then say the sentence when it needs to be said.
That’s wrong.
Stay Strong, Stay Disciplined, Stay Dangerous,
–S