Corporate life has acquired a new species.
It appears in innovation labs, strategy presentations and carefully choreographed demonstrations. It writes a convincing email, summarises a document and occasionally answers a question nobody thought to ask.
Everyone agrees it is impressive.
Then the demonstration ends, and nobody knows where it reports on Monday morning.
Welcome to the enterprise AI pilot: enthusiastically recruited, extensively introduced and completely unemployed.
Most pilots begin with a capability: “This model can analyse customer conversations” or “This assistant can support our sales team.”
But a capability is not a job.
A job means having something useful to do. It means colleagues notice when you do not turn up. Ideally, it also means replacing some work, not creating more of it.
Perhaps AI pilots should face the same questions as everyone else:
“What exactly do you do here?”
- If the answer is “increase productivity,” ask again.
- If the answer is “enable transformation,” alert Human Resources.
- And if nobody would notice whether it worked last Tuesday, it probably does not have a job.
That may explain why so many pilots remain pilots. They are safer there. In production, someone might expect them to produce something.
Meanwhile, employees are asked to use the new AI tool while continuing every existing report, process, meeting and approval.
That is not augmentation.
It is a second job.
So congratulate the team. Admire the demonstration. Ask about the model.
Then lean across the table and ask the pilot itself:
What exactly do you do here?