Kapari
The bench, taken apart

A good decision can end badly. So what do you grade?

You can do everything right and lose. You can do everything wrong and win. Three researchers drew a rule from that: you do not grade a decision on its outcome, you grade it on how it was made.

iWhat Kapari returns here is arithmetic, not an opinion. And it says what it cannot do.
What the research says

Outcomes depend on luck. Method does not.

In 2019 Daniel Kahneman, Dan Lovallo and Olivier Sibony published a protocol for major decisions. Their starting point is simple: grading a decision on its outcome is like grading a poker player on a single hand.

Two questions do most of the work. Which opposing views were voiced? And which major uncertainties were genuinely explored?

Those two questions almost never get asked. In a meeting, disagreement is socially expensive, and naming an uncertainty slows everyone down. The protocol exists to force what politeness prevents.

How it is coded

What got contradicted, and what nobody carried.

At the end of every run, Kapari counts two things from the panel data. The substantive disagreements that surfaced. And the blind spots, meaning what no voice defended.

These are counts, not sentences written by an AI. No language model touches this.

And here is the part rarely written on a website: Kapari cannot answer the researchers' second question. It does not measure which uncertainties were explored, because no part of the product captures that. So the block says what it knows, and nothing more. That is more modest than the original protocol, and it is true.

What this changes for you

On decision day, you know what was said against it.

When you take a decision to a board, the first serious question is rarely "are you right". It is "who argued the other way, and what did you answer".

You walk in with the list. The objections that surfaced, and the camp that was not in the room.

A blind spot is not a flaw in the panel. It is information: on announcement day, that particular driver will have nobody to carry it, or nobody saw it coming.

The source

The work this choice rests on.

Kahneman, D., Lovallo, D. and Sibony, O. (2019), "A Structured Approach to Strategic Decisions", MIT Sloan Management Review. The same authors worked on the inconsistency of human judgment, which is why Kapari's verdict is computed rather than written.

The bench, taken apart

More from this series

Our tool tells you when not to trust it

Why thirty voices beat three experts