𝐀 𝐏𝐞𝐫𝐟𝐞𝐜𝐭 𝐀𝐧𝐬𝐰𝐞𝐫... 𝐓𝐨 𝐭𝐡𝐞 𝐖𝐫𝐨𝐧𝐠 𝐐𝐮𝐞𝐬𝐭𝐢𝐨𝐧?

The Einstein quote goes “Not everything that can be counted counts, and not everything that counts can be counted.” Yet with AI, the temptation is greater than ever to succumb to the precision of numbers.

Rory Sutherland named the trap of our moment back in 2019:

"The risk with the growing use of cheap computational power is that it encourages us to take a simple, mathematically expressible part of a complicated question, solve it to a high degree of mathematical precision, and assume we have solved the whole problem."

It’s again the story of the man on his knees under a streetlight, searching for his keys. A passerby asks if he's sure he dropped them there. "No," he says, "but this is where the light is."

Here is what changed: the blind spot itself is old. People have always trusted what is easy to count over what truly counts. Cheap computing power did not create that bias. It industrialized it. Optimizing the measurable slice now costs almost nothing, so we do it everywhere by default.

But flip the risk over. The opportunity is hiding inside it.

When precision becomes free, precision stops being the prize. Everyone runs the same optimizer on the same measurable slice, so that slice quickly becomes a commodity.

So the value moves to the part no model can reach: the work of framing the right question, deciding what truly deserves precision, and earning the trust that makes a number worth believing. As Sutherland put it, you can optimize ad spend to the decimal and still say nothing about why anyone should trust you enough to buy.

So take the metric your team optimizes hardest, and ask three questions:

- What wider question is this precise number quietly standing in for?
- What did we stop measuring because it would not fit the model?
- If a competitor ran the very same optimizer, what would still be ours?

The edge in the AI era will not belong to whoever holds the most precise answers, because those are nearly free now. It will belong to whoever unearths the questions no machine would ever think to ask.

Is there one number your team trusts a little too much?

#PrecisionFallacy #RightQuestion #DecisionMaking #AIStrategy #CriticalThinking

Contact

bruno.gentil@sherpaconsultingasia.com

© 2026. All rights reserved.