The plan has a risk register, and it has a room that has already decided the plan will work. Ask what might go wrong and you get the polite risks, the ones that do not threaten the date. The decision leaves in the same shape it arrived.
Gary Klein, chief scientist of Klein Associates and author of Sources of Power, set out a different question in Performing a Project Premortem, Harvard Business Review, September 2007. A premortem is the opposite of a postmortem. It comes at the start, so the work can be changed rather than autopsied. The team is told the project has failed, and is asked what did go wrong, not what might. He ties the move to research from 1989 by Deborah J. Mitchell, Jay Russo and Nancy Pennington: imagining that an event has already occurred raised, by 30 percent, the ability to identify correctly the reasons for a future outcome.
Why the polite list stays polite
Klein's complaint about an ordinary critique is that people are asked for things that might go wrong while the plan is still a hope they are supposed to share. The dangerous reasons are the ones that sound disloyal. In the method, the leader states that the failure has already happened, spectacularly. For a few minutes people write on their own, including the reasons they would usually swallow. Then each person reads one reason, starting with the project manager, and nobody repeats a reason already given, until the lists are on the wall. The manager's job after that is to use the list to strengthen the plan.
The examples are ordinary, which is why they were easy to skip. At one very large company, a failure story was that interest in an environmental project died when the chief executive retired. On a project to put planning algorithms onto machines used in the field, someone who had been silent in the kickoff said one algorithm would not fit the laptops and would run too slowly. The people who had built it already had a shortcut they had not wanted to mention. The shortcut went in. In another organisation a senior person said the failure was that there would be no time to prepare a case before a corporate review. Nobody had mentioned the clock. The manager changed the plan.
What the agenda gives away
If the proposed day teaches a catalogue of biases and never puts a live plan under the assumption that it has already failed, you will leave with vocabulary. The register you already have is the polite version of that catalogue. Klein's distinction is the tense. Might is an invitation to be constructive. Did is a permission to be specific. If people write after the first person has spoken, you will get variations on that person's reason. The few minutes alone are the method, not a warm up.
Then look at what happens to the plan. A list that is thanked and filed is a postmortem you held early. In the sessions he describes, a sentence changed the work: a shortcut was admitted, a review date was taken seriously. If the provider cannot show you a line in your plan that is allowed to move, the exercise is a story circle.
Collective Campus is building high-performance culture and teams for the AI world. The critical thinking workshop is for a group that can bring a live decision and is allowed to change it when a failure reason survives contact with the plan. Whether a thinking label travels to a later call in the same work is in how to judge the day for transfer. A room in which the useful comment was never sayable is the subject of nobody disagreed. Where this skill sits against a literacy course is in why critical thinking beat AI literacy.
What to bring
Bring the plan you are about to approve, and the name of the person who can change a line in it. If that person only wants the risks that leave the plan intact, the failure has already been edited out of the meeting.
