The Uncomfortable Questions · An AI FAQ for leaders

Question #29 · Economics

When AI makes a task ten times cheaper, does your organization do it ten times more, or do it better?

The short answer

When something gets cheaper, we often use more of it, not less.

Economists have a name for this. Cheap reports produce more reports. Cheap analysis produces more analysis nobody reads. The leaders who capture value decide in advance what to do with the capacity AI frees up. Otherwise it fills itself.

Norm E. Murray

The economics

Why it happens

Lower cost changes behavior before anyone decides it should, and the default behavior is more volume.

01

Cheaper means more

William Stanley Jevons noticed in the 1860s that more efficient steam engines raised Britain’s coal use instead of cutting it. The Jevons paradox applies to knowledge work too: make a report cheap and the organization orders more reports.

02

Work expands to fill the space

C. Northcote Parkinson observed that work expands to fill the time available for it. Hours AI frees up do not sit idle; they fill with whatever is nearest, which is usually more of the same.

03

Attention does not scale

Herbert Simon warned that a wealth of information creates a poverty of attention. AI multiplies output, but leaders still have the same hours to read it, so more analysis often means less of it gets used.

04

Volume is easy to count

The McNamara fallacy is managing by what is easy to measure and ignoring what is not. Reports produced and tasks completed are easy to count, so they become the proof of AI value while decision quality goes unmeasured.

The bottom line: Cheaper work is only an advantage if someone decides what the savings are for.

Your move

What to do about it

  1. Decide where the capacity goes before you deploy.For each AI rollout, write down what the freed time or budget will be used for. If the answer is blank, the default is more volume.
  2. Cap the volume.Set a ceiling on recurring reports, decks and dashboards. When AI makes a new one, an old one retires.
  3. Measure better, not more.Track whether decisions, forecasts or customer outcomes improve, not how much output the tools produce.
  4. Reinvest on purpose.Point the freed capacity at work AI cannot do well: customers, hard judgment calls and developing people.
  5. Audit what nobody reads.Once a quarter, find the reports and analyses with no reader and no decision attached, and stop producing them.