# Buyer-intent prompts

> Buyer-intent prompts are the questions a person asks an AI assistant while deciding what to buy: which options fit a category, how two of them compare, whether a known objection holds, and what works for a specific use case or context.

Canonical page: https://llmaudit.app/learn/buyer-intent-prompts

Buyer-intent prompts are the questions a person asks an AI assistant while deciding what to buy: which options fit a category, how two of them compare, whether a known objection holds, and what works for a specific use case or context. They are the questions worth measuring, because they are the ones where being named or skipped changes a purchase.

## Five kinds of buyer question

Most buyer questions fall into one of five shapes, and a good set of prompts covers all of them rather than repeating the first.

- Category: "what are good options for X", the shortlist question.
- Comparison: "A or B for someone who needs Y".
- Objection: "is X worth it if I already have Z", the doubt that delays the purchase.
- Use case: "what should a small team use to do X".
- Context: the buyer describes their situation instead of asking for a tool, and expects the assistant to name one.

## Branded versus open prompts

A branded prompt names the brand: "is Acme good for X". An open prompt does not. When the buyer types the brand's name, a mention in the answer only proves that the model knows the brand exists; it measures recognition. A recommendation is when the model names the brand in an open prompt, without being asked about it. The two are different facts, and counting them together inflates the headline. That is why only open prompts count toward a verdict, and branded ones are reported separately.

## How LLM Audit derives them

The audit does not start from a generic template. It reads the brand's own site, works out what it sells and to whom, and derives up to 12 questions in the five shapes above. Each question is asked to every provider in the basket, and the verdict is computed only once at least 6 open questions have an answer. The result is a map: which questions named the brand, which named a competitor, and which nobody answered clearly.

## FAQ

### Why not just ask "best X tools"?

Because buyers rarely ask that, and the answer to it is unstable between runs. One shortlist prompt tells you little about the comparison, objection and use-case questions where the purchase is actually decided.

### Should the prompts include my brand name?

Some should, so you can see how the model describes you. But those answers measure recognition, not recommendation, and they should not be counted toward the same total as open prompts.

### How many prompts are enough?

Enough to cover the five shapes for the brand's main offer. LLM Audit uses up to 12 per audit and requires at least 6 answered open questions before it states a verdict.

## Related

- [Query fan-out](https://llmaudit.app/learn/query-fan-out.md)
- [AI visibility audit](https://llmaudit.app/learn/ai-visibility-audit.md)
- [How many runs](https://llmaudit.app/learn/how-many-runs.md)
- [Query fan-out and the buyer questions that decide](https://llmaudit.app/blog/query-fan-out-buyer-questions-ai-visibility.md)
- [All definitions](https://llmaudit.app/learn.md)

---

Machine-readable index: https://llmaudit.app/llms.txt · Full context: https://llmaudit.app/llms-full.txt · All URLs: https://llmaudit.app/sitemap.xml · Developer and agent resources: https://llmaudit.app/developers.md