Look for behavior, not compliments
Friends and potential users can call an idea interesting without intending to use it. Hypothetical questions produce polite answers; past behavior produces better information.
Ask when the problem last happened, how they solved it, what it cost and why they did not adopt an alternative.
Choose the cheapest useful test
A page with a clear promise, a clickable prototype or a manually delivered service can test demand before full automation.
The test should require commitment: a qualified contact, a demo booking, data, time spent in a trial or a preorder.
Decide in advance what changes your mind
Set a threshold for conversations, requests or paid tests that justifies development. Also define what would make you change segment or stop.
Without criteria, every weak signal becomes confirmation. Validation buys clarity at low cost.
Involve a developer at the useful moment
A freelancer can support discovery through feasibility, risk review and prototypes. They should not turn every question into weeks of code.
Once you have a specific user, recurring need and promising behavior, MVP scope becomes smaller and estimates become more credible.