Business idea validators: what they can and cannot tell you

Kind 1: AI validators

You describe your idea and a large language model writes an assessment, often with a score and a list of risks. Some also search the web for competitors. They are fast, they can surface angles you had not considered, and they are good at spotting obvious gaps.

Limits: the model does not know your customers. Ask the provider how its scoring is kept consistent. Read the reasoning, not the number, and check anything factual, such as a competitor or a market figure, before you rely on it. Check too what happens to the idea text you submit.

Kind 2: market-data scores

Some tools combine public data, such as population, local business counts or search interest, into a score for an idea in a particular place. They are most useful for local, physical businesses where geography matters.

Limits: public data can lag by a year or more, and a good-looking market says nothing about whether customers will buy from you in particular.

Kind 3: fixed rubrics and scorecards

A rubric scores your description against published rules, for example whether you have named a customer, a price and a channel. Under the same rules, the same words get the same score, so you can see exactly what changed when you edit. A rubric measures clarity, not potential.

Limits: a rubric cannot judge whether the problem is real or the price is right. It can only tell you whether you have stated them clearly enough to test.

What no validator can do

• Tell you whether strangers will pay. Only asking them for money can do that.
• Replace customer conversations. Ten honest conversations beat any score.
• Account for you: your skills, time, network and appetite for risk.
• Give legal, tax or financial advice. Use GOV.UK and qualified advisers for that.

How to use a validator well

• Write the idea in five parts first: problem, customer, offer and price, channel, first sale.
• Use a tool to find the vague or missing parts, then fix them.
• Take the sharpened version to real people within a week.
• Record what they said and did, and re-score only after you have learned something new.
• Ignore any tool that promises the idea is "validated" because of a score alone.

Questions to ask before you trust a score

• What exactly does the score measure: clarity, market data or a model's opinion?
• Would the same description get the same score tomorrow?
• Is the method published, so I can see what earns the points?
• What happens to the idea text I submit?
• Does the tool say clearly what the score does not mean?

Where Solopreneur Protocol fits

Solopreneur Protocol is a fixed rubric (kind 3). It scores how clearly your idea is described across five parts, out of 100, using published rules on our server. No AI is used and your idea text is not sent to any AI model. The first score is free; Full Access is £29 one-time for unlimited scores (fair use), the Idea Vault and the Full Access Playbook. It measures clarity, not success, and it is not business, financial or legal advice.