A pilot is easy to start. Someone in the business finds a tool, tries it on a real task, gets a result worth showing around. Six weeks later there is a slide with a screenshot on it. This is the part everyone is good at now.
What happens after the slide is where most of the work quietly stops.
The pattern is familiar enough that it barely needs describing: a pilot proves a concept, gets applauded in a meeting, and then sits. Never cancelled, never formally failed, just never picked back up, because picking it back up means answering questions nobody scoped for at the start. Who owns this now that it is not a project. What happens when the input data changes. Who checks the output before it reaches a customer. None of that is glamorous, and none of it shows up on a pilot slide, which is exactly why it gets skipped.
The pilots that do reach production share a boring trait: someone decided, before the pilot started, what “done” would mean: a specific measure, agreed in advance, that a real operator would check on a Tuesday six months later. Time saved against a baseline. Error rate against a threshold. A number that exists whether or not the project is exciting anymore.
The second trait is ownership that survives the pilot. A pilot run by an enthusiastic individual dies when that person moves teams, gets busy, or loses interest, which happens to everyone eventually. A pilot with a named business owner, someone whose job description already includes the outcome the pilot is meant to improve, has somewhere to live once the initial energy fades.
The third trait, and the one most often missing, is a decision about what happens next before the pilot even finishes. Scale it, kill it, or fix it and try again: three possible outcomes, and most organisations only plan for the first one. When the pilot instead produces a mediocre result, which is a common outcome, there is no process for that. So it sits. Not a failure anyone has to explain, not a success anyone can build on. Just unfinished.
None of this is a technology problem. The tools involved in a pilot that reaches production and a pilot that stalls are frequently identical. The difference is whether the organisation treated the pilot as a demonstration or as the first stage of something with a defined next stage. A demonstration proves the tool can do the thing. A pilot, properly run, proves the business can absorb the change, and that is a different question with a different answer.
The practical move is to write the ending before the beginning. Before a pilot starts: the measure it will be judged against, the person who owns it once it stops being a project, and the decision rule for what happens at the end, all three, in writing, agreed by the people who will still be in the room when the pilot finishes.
If your organisation has more than one pilot sitting in this state, that is not unusual, and it is not a sign the tools were wrong. It is usually a sign the ending was never written. Momentum Diagnostic exists to score exactly this: which of the workflows already in motion, or waiting to start, are worth carrying to a real ending.