A pilot produces evidence rather than anecdote only if four things are fixed in writing beforehand, namely a success metric, a measured baseline, an end date and a named honest reporter.
You have chosen the department. The next mistake is assuming the pilot will speak for itself. It will not. Run it without structure and what you get back, however genuinely positive, is a manager's impression: people liked it, it saved time, everyone wants to keep going. That is a story, and a sceptical board member is right to discount a story, however sincerely it is told.
The difference between a pilot that changes minds and one that gets politely ignored is almost entirely decided before the pilot starts, not after.
It looks like a summary built from memory rather than a record: "it felt faster." It looks like screenshots of the best outputs, selected after the fact, with nothing to show what a typical output looked like. It looks like no defined finish line, so the pilot simply continues until someone asks for an update, at which point the answer is whatever mood the team happens to be in that week. None of this is dishonest. It is just not evidence, and treating it as evidence in front of a board is where credibility gets spent for nothing.
The fix is not more enthusiasm or a longer pilot. It is four decisions made in advance, in writing, before the first task runs.
A pilot without a metric decided in advance produces a story. A pilot with one produces evidence.
All four of these cost real time before anything visibly starts: writing the metric, measuring the baseline, agreeing an end date, naming an honest reporter. It feels like bureaucracy layered onto something that was supposed to be quick and informal. The honest answer is that this upfront cost is what the evidence is made of. Skip it to save a week at the start, and you pay for it later with a result nobody in the room believes, at exactly the moment you needed them to.
Before the pilot starts, write the success metric, the baseline, and the end date on one page; if that page cannot be produced in an afternoon, the pilot is not ready to start yet.
No external statistic cited; this article presents an internal pilot-design method rather than third-party evidence.
If this is the question on your desk, a thirty-minute call tells you whether the service fits, or that you do not need us yet.