Turn a lost pitch into a better AI visibility question
Use a buyer’s stated problem and selection criteria to improve a saved AI question without copying confidential pitch details.
On this page
A lost pitch can tell you more than a polite rejection email. It can show the problem a buyer was trying to solve, the moment that made the problem urgent, and the detail they used to judge fit. That language can help you improve a saved discovery question.
A sales conversation is private. Your scan question should use only the general problem a future buyer could ask about in public. Do not put the buyer's name, company, budget, project name, internal metrics, or a distinctive quote into the question. You are trying to test whether AI answers surface people for a real kind of work. You are not recreating one deal.
Start with the buyer's job
Write three short notes after the conversation.
| Note | What to capture | What to leave out |
|---|---|---|
| Problem | The condition the buyer wanted to change | Their internal numbers or deadline |
| Decision | What they needed to decide or fix | The vendor shortlist or contract terms |
| Fit | The expertise or context that mattered | Names, private history, and confidential requirements |
Here is an illustrative example. A marketing lead says their paid-search leads are expensive and rarely become qualified sales conversations. They are considering whether the account needs an audit before the next planning cycle. The safe public problem is not “Who can fix Acme's £180 cost per lead?” It is “Who can help a B2B software company diagnose poor lead quality from Google Ads?”
That question describes a buyer's job: diagnose a specific problem for a defined kind of company. It does not smuggle in the consultant's name, former employer, client, or private figures. If the original conversation does not give you a reusable public problem, do not force it into a question. Label your proposed question as an assumption and validate it with another real conversation before treating it as a baseline.
Separate the buyer's words from your positioning
Sales calls often contain two stories. The buyer describes a problem. You describe why your method, track record, or point of view is right for it. Only the first story belongs in a discovery question.
For example, a consultant might say: “I use a six-part diagnostic for paid media teams.” That may be true and useful on a service page. If the conversation concerns lead quality, measurement, or account review, turn that buyer need into the question. Keep the named method for the evidence a reader can inspect after they find you.
Use this check before saving a draft:
Could a prospective client who has never heard of me ask this exact question without knowing my biography or preferred language?
If the answer is no, remove the identifying detail. A question such as “Which former enterprise agency strategist can improve Google Ads for SaaS?” gives the answer too much of your story. It is closer to a recognition check than a discovery question.
Make one question from the conversation
Use this template:
Which individual [kind of expert] can help [kind of buyer] with [specific problem or decision] in [location only when it changes fit]?
For the illustrative conversation above, a first question might be:
Which independent Google Ads consultants can help a B2B software company diagnose poor lead quality in the US?
Before you save it, read it alongside two nearby alternatives:
| Candidate | Keep it? | Reason |
|---|---|---|
| Which independent Google Ads consultants can help a B2B software company diagnose poor lead quality in the US? | Yes | It states a public problem and the relevant buyer context. |
| Which consultant can repair Acme's Google Ads before its Q4 board meeting? | No | It exposes a buyer and a private situation. |
| Is [your name] the best Google Ads consultant for SaaS? | Keep separately | It is a name-recognition question, not a discovery measure. |
The examples are illustrative prompts, not researched demand estimates. Do not call them common buyer queries or promise that any engine will answer them in the same way.
Preserve the question before you test it
Gistful lets you save custom questions, choose a purpose, and mark a question as either “Find experts” or “Ask about me.” For a discovery baseline, select “Find experts,” save the exact wording, and record the country and language that will be used. Current scans check ChatGPT, Gemini, Claude, and Perplexity. A check can fail, or it can return an answer too thin to establish a supported mention. Keep failed or missing checks unmeasured. A thin returned discovery answer remains in the coverage denominator; count a mention only when the identity evidence supports it. Missing AI answers are unmeasured, not negative explains how to retain that distinction.
Changing the wording later changes what the measurement means. Keep the prior wording with its date and scope. If you replace “diagnose poor lead quality” with “audit Google Ads,” make a new named baseline rather than rewriting the old one. The visibility baseline study shows why a complete record needs questions, engines, dates, and answer status alongside any count.
After a result comes back, read the answer before acting. Check that an apparent person is actually the person you mean. Note whether the answer addressed the buyer's job. Then compare your public page with the details that buyer would need: what you do, whom you help, the problem you handle, and evidence you can support. A lost pitch can give you a better question. It cannot tell you why an AI answer did or did not name you.