Sense and decide
Common Ground First
Everyone preps with AI beforehand, so the session starts at the same level and goes straight to judgment.
Borrowed from the Cybernetic Teammate field experiment at Procter & Gamble, by Fabrizio Dell'Acqua and colleagues (NBER Working Paper 33641). Read the original source.
One session · Minutes to start · Anyone on the team can start it
Try this first
Ask everyone to get up to speed on the background with AI before the next session.
Use it when
Sessions burn their first half getting everyone to the same starting line.
Skip it when
It works where background material exists and AI can help absorb it. It does not suit sessions whose purpose is building understanding together, and it should not replace sense-making that a team needs to do jointly.
How to introduce it
Before a working session, have everyone use AI to get to the same baseline on the material. Background, basics, context. Then spend the session on what needs human judgment. AI raises the floor, so the newcomer arrives oriented and the group skips the cost of getting everyone level.
How to show up
Set the expectation that people arrive at the baseline, and design the session for what remains. If the prep does not happen, the session slides back into catch-up.
How long it takes
Shifts preparation to before the session. Makes the session itself shorter and sharper.
What makes it hard
People skip the prep unless the session visibly assumes it, so the norm has to be real. A baseline reached privately can also hide gaps or push everyone into the same framing, so check that the shared ground holds.
What it looks like when it's working
People arrive oriented and the time together goes to judgment. Sessions still opening with long context-setting mean the prep is not happening.
How long until it sticks
Expect a few sessions before coming-prepared-to-a-shared-baseline becomes the norm.
How you know it stuck
People arrive at the shared baseline and meetings start from common ground.
The idea behind it
Meetings burn most of their time getting everyone to the same starting point, and AI can do that beforehand. What is left is the part only people can do.
Where it comes from
The Cybernetic Teammate experiment. It uses the leveling effect on purpose: the baseline rises, the newcomer walks in oriented, and the group skips the tax of catching everyone up. Before a working session, everyone uses AI to reach a shared baseline on the material, so the session itself is spent only on the parts that need human judgment.
Pairs with
Independent view first, and this pairing is non-negotiable. A shared baseline is one keystroke from shared groupthink, because everyone's model nudges them toward the same answer. On anything that's a decision, people form and record their own position before the group converges. Use this for orientation, never to pre-cook the call.
Also fits: Decisions don't close, or nobody knows who decides