Most founders begin with a solution they can picture clearly. The harder, more useful work is proving that the problem behind it is real, frequent and painful enough for someone to change what they do today.
Start by naming a specific customer, not a market. Then speak to people who fit that description. Ask about the last time the problem happened, what it cost them and what they tried. Listen for workarounds such as spreadsheets, manual steps and paid fixes. They show the problem is already being solved badly.
Avoid asking whether they would use your idea. Polite interest is easy to give and tells you little.
Write down your riskiest assumption and design the smallest test that could prove it wrong. Build only after it survives.