the deck shows the pilot was a success. accuracy is up. engagement is up. the team that built it has been promoted. the next phase is funded. the only thing nobody is asking is whether anything that the company actually does is different than it was eighteen months ago.
the pilot didn't fail. it just didn't change anything.
decoration is the most expensive kind of ai project because it survives. it has good numbers. it has happy users (the ones who built it). it has roadmap. it has executive sponsorship. it has everything except the one thing that justified the budget — a workflow that runs differently because of it.
there are three signs that an ai pilot is decoration. each one is a question you can answer in five minutes about any pilot in your portfolio.
- 01 the operations test name the workflow that stopped being done the old way last month because of this pilot. if you can't — or if the answer is some version of "people are using it more" — the pilot is decoration. activity isn't change.
- 02 the metric test point to the kpi this pilot is supposed to move. now look at the kpi six months before launch and six months after. if the line is flat or you can't find the line at all because nobody owned the kpi to begin with, the pilot is decoration. measurement after the fact is interpretation.
- 03 the absence test imagine the pilot is shut off tomorrow for a week. who would notice? if the answer is "the team that built it, and the people presenting it," the pilot is decoration. real systems leave gaps when they're gone.
decoration survives because nobody who funded it wants to be the one who killed it. the board signed off. the cfo budgeted. the head of innovation gave the keynote. the only people who would actually call it out — the operations team that's still doing the work by hand — were never asked. by the time it's obvious the pilot isn't moving anything, the budget has already moved on to the next pilot. the cycle continues.
the financial cost of decoration pilots is rarely the worst part. the worst part is what they teach the organization. teams learn that ai is something that lives in slide decks. operations learns that whatever the head of innovation is talking about will probably not affect them. the company builds an immune system against ai — not because ai is bad, but because the version of ai they've experienced was theater.
the move isn't to fund fewer pilots. the move is to refuse to fund any pilot that can't pass the three tests at the proposal stage. name the workflow. name the kpi. name what would be missing if the pilot were gone. a pilot that can't answer those three before it starts won't be able to answer them after, either.
a pilot that nobody would miss never existed.