The Uncomfortable Questions · An AI FAQ for leaders

Question #28 · Economics

Is your AI investment buying capability, or buying reassurance?

The short answer

Reassurance spending feels productive and rarely is.

A lot of AI spend is defensive: the fear of being seen as behind. The test is simple. If the board stopped asking about AI tomorrow, which of these investments would you still make? Those are the real ones.

Norm E. Murray

The economics

Why it happens

Under uncertainty, organizations spend to look prepared, and looking prepared is easier to buy than being prepared.

01

Everyone copies everyone

Sociologists Paul DiMaggio and Walter Powell called it mimetic isomorphism: when the future is unclear, organizations imitate peers they see as successful. Much AI spend is bought because competitors bought it, not because anyone defined what it should change.

02

Spending becomes the signal

Michael Spence showed how costly signals tell outsiders what they can’t see directly. A large AI budget signals seriousness to boards and investors, whether or not it builds anything the business uses.

03

Theater is cheaper than security

Bruce Schneier coined security theater for measures that make people feel safer without making them safer. AI theater works the same way: pilots, task forces and announcements that calm the room and leave the operating model unchanged.

04

Sunk costs keep the theater running

Once money is spent, leaders keep funding the program to justify what was already spent. The sunk cost fallacy turns a defensive purchase into a permanent line item.

The bottom line: An investment you only make because someone is watching buys comfort, not capability.

Your move

What to do about it

  1. Run the board test on every AI investment.List each one and ask whether you would still fund it if the board stopped asking about AI tomorrow. Mark the ones that fail.
  2. Name the decision each investment improves.Capability changes a specific decision, process or customer outcome. If nobody can name it, the spend is reassurance.
  3. Fund outcomes, not announcements.Tie each budget to what the business does differently, not to pilots launched or licenses bought.
  4. Cut the reassurance spend.Treat money already spent as gone. Redirect what remains to the few investments that pass the test.
  5. Change what you report upward.Show the board capabilities built, not activity, so the pressure to look busy with AI stops setting the budget.