Founders often fall in love with a solution before they have confirmed the problem. A few weeks of validation can save months of development and show how the product should really look.
Talk about the problem, not the idea
Find people from your target audience and ask how they deal with the problem today: what they do, how often it happens, how much time or money it costs, what they have already tried. Avoid presenting your idea and asking "would you use it?" - most people answer politely "yes". Stories about real past behaviour are far more reliable than opinions about the future.
Check whether they are already paying
If people spend money or significant time on workarounds - spreadsheets, freelancers, several tools stitched together - the problem is real. If nobody has ever tried to solve it, it may not be painful enough to pay for.
Build a simple landing page
Describe the value of the future product in a few clear sentences, add a call to action - join the waiting list, request a demo, book a call - and bring visitors from the channels where your audience lives. The share of visitors who take action tells you more than any survey.
Deliver the service by hand
Before automating, try providing the result manually to a few clients: prepare reports in spreadsheets, match requests yourself, send recommendations by email. Clients receive real value, and you learn exactly which steps matter and which could be dropped.
Ask for a commitment
The strongest signal is money or another real commitment: a paid pilot, a pre-order, a signed letter of intent, time spent on regular calls with you. Compliments are pleasant; commitments are evidence.
The short version
Confirm the problem through stories of real behaviour, look for existing spending on workarounds, test demand with a landing page, deliver the first results manually and ask for real commitments. Then build the product the evidence points to.