Sense and decide
The Key Assumptions Check
List what the plan assumes and test each
Name what would have to be true before you act on an AI recommendation.
Borrowed from the Key Assumptions Check, a structured analytic technique taught in the CIA's Tradecraft Primer. Read the original source.
One session · Minutes to start · Anyone on the team can start it
Try this first
Ask one question. Which assumption, if wrong, sinks this?
Use it when
A plan rests on unexamined assumptions.
Skip it when
Use it on decisions that matter and on AI output that arrives confident and polished. Do not slow every small call with it.
How to introduce it
Before the team acts on a recommendation, ask what has to be true for it to hold. Say why you are asking. You are strengthening the decision, not attacking whoever brought it.
How to show up
Lead it yourself the first few times. Ask what we are assuming and write each one where people can see it. Push past the first two, because the assumptions nobody says out loud are the ones most likely to be wrong.
How long it takes
Two to three minutes once it has become a habit. Longer the first several times.
What makes it hard
Teams list the safe assumptions and stop, so prompt them for what has not been said. Smooth, certain-sounding output makes people reluctant to push, so name that pull and let them catch it happening.
What it looks like when it's working
At least one unnoticed assumption comes out, and a confident recommendation drops in status once its footing shows. A neat list everyone nods at is going through the motions.
How long until it sticks
This is a thinking habit. Expect several real uses before it becomes automatic, and note that it fades if the team stops doing it.
How you know it stuck
Attach it to a trigger. Any AI-driven decision above an agreed threshold gets one, so it stops depending on you.
The idea behind it
Find the weak assumption before reality does. Whoever spots it is helping.
Where it comes from
Intelligence analysis. Name the load-bearing assumptions under a plan and ask what happens if each is wrong. Explicitly list the assumptions a judgment rests on, then stress-test each.
Why it matters for senior teams
It shapes direction and consequential decisions
Also fits: Confident claims go unchecked
Home problem: The group lands on the first answer, often the AI's