The Uncomfortable Questions · An AI FAQ for leaders

Question #25 · Economics

Are you measuring AI by hours saved or by decisions improved? Only one of those changes your strategy.

The short answer

Hours saved is the easiest number to report and the least meaningful.

Saved hours usually get absorbed into more meetings and more email. The return that compounds comes from better decisions: pricing, capital allocation, hiring. If your AI business case only counts time, it’s measuring the wrong asset.

Norm E. Murray

The economics

Why it happens

Time is easy to count and easy to defend, so it becomes the default proof of value even when it proves very little.

01

Looking under the streetlight

The streetlight effect describes searching where it is easiest to look rather than where the answer is. Hours saved sit under the streetlight: easy to measure, easy to report, and rarely where AI creates its real value.

02

The productivity paradox

Economist Robert Solow observed in 1987 that computers showed up everywhere except in the productivity statistics. Time saved at the desk does not automatically become value for the business, and AI can repeat the pattern.

03

Decisions compound, hours do not

A saved hour is a one-time gain. A better pricing, capital or hiring decision keeps paying back for years, which is why decision quality is the asset worth measuring.

04

What you measure becomes the strategy

Robert Kaplan and David Norton’s balanced scorecard work made the case that what an organization measures shapes what it pursues. Measure AI only in time and the strategy becomes efficiency; measure decisions and it becomes advantage.

The bottom line: Hours saved tell you AI is busy. Decisions improved tell you it is worth it.

Your move

What to do about it

  1. Name the decisions.List the three to five decisions where better judgment would move results most, such as pricing, capital allocation and hiring.
  2. Set a baseline.Record how those decisions are made and how they turn out today, before AI is involved.
  3. Measure the decision, not the time.Compare forecast accuracy, win rates or hiring outcomes where AI is used and where it is not.
  4. Assign the saved hours.If time is saved, decide in advance where it goes, or it will be absorbed.
  5. Rewrite the business case.Require every AI proposal to state which decision it improves and how you will know.