The Uncomfortable Questions · An AI FAQ for leaders

Question #30 · Economics

What’s one AI investment you’d make differently with hindsight?

The short answer

Most hindsight lessons I hear are about people, not technology.

Moving before the team understood why, or measuring the wrong thing. It’s the most generous question a leader can answer in public, because someone reading this is about to make the same call.

Norm E. Murray

The economics

Why it happens

AI investments are decided on technology, but they succeed or fail on people and measures, which is where the regrets collect.

01

Hindsight makes it look obvious

Hindsight bias makes past outcomes look predictable once we know them. The useful lesson is not that we should have known, but which signals were visible at the time and ignored.

02

Adoption runs at human speed

Everett Rogers’ research on the diffusion of innovations showed that adoption depends on how people perceive a new tool, not just on what it can do. Rollouts that skip the why stall, however good the technology.

03

Money already spent keeps spending

The sunk cost fallacy makes it hard to change course once money is committed. Leaders keep funding an AI project because of what it already cost, not because of what it will return.

04

The metric becomes the mission

Researchers call it surrogation: managers start treating the metric as if it were the strategy itself. An AI program measured on usage or hours can look successful while missing what it was bought to do.

The bottom line: The expensive lessons are rarely about the model. They are about the people and the measures around it.

Your move

What to do about it

  1. Run a pre-mortem.Before the next AI investment, imagine it has failed a year from now and list the most likely reasons.
  2. Explain the why first.Make sure the people affected understand the reason for the change before the tool arrives.
  3. Pick the measure up front.Agree on the decision or outcome the investment should improve, not just usage or time saved.
  4. Set a kill point.Decide in advance what result would make you stop or redirect the investment.
  5. Pass the lesson on.Write down what you would do differently and share it with the next team making a similar call.