The decision
An idea becomes easier to test when it names a person, a recurring task, and a visible result. “AI for small businesses” leaves too much open. “Help a small agency turn approved meeting notes into a client update” tells you whom to interview, what sample to request, and how to compare the output with existing work.
A worked example
Three agency owners all mention client communication, but their tasks differ. One writes proposals once a quarter; another sends weekly updates; a third needs invoice reminders. Treat these as separate opportunities. Weekly updates may be easier to observe quickly, but frequency alone does not establish a budget or permission to use the notes.
How to put it into practice
- Collect recent examples instead of asking people to imagine future behavior.
- Write down the current workaround and who performs it. Existing effort is more useful than a compliment about your idea.
- Separate the end user, the budget owner, and anyone who controls access to input data.
- Choose a narrow task you can deliver manually before investing in automation.
A failure to plan for
A frequent annoyance can still be too minor to buy a solution for. Ask what happens when the task is delayed or done poorly and compare that consequence with the price and switching effort of your proposed product.
Try it on your project
Write three candidate tasks using this sentence: “For [role], produce [result] from [input] when [trigger].” Find one recent example for each and choose the candidate with the strongest evidence and easiest next test.
Keep the next step small
Use the free demand scorecard or planning tools to make your assumptions explicit. The $19 launch kit brings the blueprint and seven editable worksheets together.