A paid project comes with edges built in. There is a document, a price, and a list of what is included. Somebody has to sign it, so somebody reads it.
A free pilot has none of that by default. It usually starts with goodwill and a sentence like "let's just try it for a few weeks." I like that spirit. I have also watched it turn into open-ended work that nobody planned and nobody could gracefully end.
Because no money changes hands, every new request feels small. Could it also do this? Could one more person get access? Saying no feels petty when the whole thing is a favour. So you say yes, and then yes again, and a month later you are running a project with no scope and no end date.
These days I try to give a free pilot the same edges as paid work, just lighter.
An end date that is on the calendar before it starts.
One or two outcomes we are trying to see, written in plain words. Not a feature list. Something like "the team can log a call in under a minute."
A short list of what is in, and a second list called "possible later." The second list matters more than I expected. It lets me say "good idea, that goes on the later list" in place of a flat no.
A rough cap on my own hours, so I notice when I pass it.
And an agreed moment at the end where we both decide what happens next. That can be a real project, a pause, or a friendly goodbye.
None of this is about protecting myself from the other person. It protects the pilot. A trial with a clear question can fail honestly or succeed honestly. A trial that drifts teaches nobody anything, and it tends to end with both sides a little disappointed and not sure why.
I still say yes to free work sometimes. I just write it down first.