Nobody asks a utility for a proof of concept.
You ask for one when you don’t believe the thing will work in your situation, which makes a POC a useful signal — just not the one most vendors read. It rarely means the buyer is sold. It usually means they are carrying a doubt they couldn’t settle any other way, and the doubt is almost never about the feature you demonstrated.
Will it hold up under our load. Will support pick up the phone. Does it work with the two systems nobody mentioned in the meeting. Did I forget anything.
Every one of those is a completeness question, and every one is answerable with evidence rather than a pilot: a reference doing the same thing at the same scale, a documented integration, a service level with teeth, a partner who has delivered it four times already. Get good at that and the proof of concept gets shorter. Sometimes it stops being asked for at all.
In my opinion that’s the cheapest deal-velocity lever most companies have, and it needs no engineering whatsoever. Though, there is a counter-argument, because there are companies where the POC is simply a procurement requirement and no amount of evidence replaces it.
Thoughts or comments? You can find me on Linkedin or Substack.