The four prompt types behind a useful AI visibility programme

Not all tracked prompts measure the same thing. A prompt that names your brand tests what an AI engine knows about you. A prompt that describes a buyer's problem without naming a supplier tests whether the engine discovers you. Treating those results as interchangeable removes the context that makes AI visibility useful.

BrandHalo organises monitoring prompts into four topics: Category, Brand, Persona and Comparison. In this series, we use Generic as the plain-language description for Category prompts because they ask about the market without seeding your brand.

1. Generic prompts measure discovery

Example: "What are the best enterprise brand management platforms?"

Generic prompts test whether an answer engine independently identifies your brand as relevant. They are usually the strongest signal of unprompted discovery because the question supplies a need or category, not your name.

Use them to answer:

• Does our brand enter the consideration set when buyers have not selected a supplier?
• Which competitors are recommended for our priority categories?
• Which sources influence the answer?

Generic prompts have the lowest inherent brand-seeding effect, but they still need care. A vague category may be too broad to reflect your market, while an overly specific one may quietly favour a particular product.

2. Brand prompts measure knowledge and representation

Example: "What does BrandHalo do?"

Brand prompts tell you what the engine says when someone asks directly about your organisation. They are valuable for checking factual accuracy, positioning, sentiment, customer-fit descriptions and whether outdated claims continue to appear.

Use them to answer:

• Does the engine describe our brand accurately?
• Are our approved positioning and differentiators understood?
• Are there recurring misconceptions or stale facts to correct?

A high Brand score should not be interpreted as high discovery. Your name was already in the question. It is better understood as a representation signal.

3. Persona prompts measure relevance in context

Example: "What tools should a CMO use to govern brand consistency across regions?"

Persona prompts introduce a role, goal, pain point, industry or operating context. They test whether an answer engine connects your brand with a specific audience need. For BrandHalo customers, these prompts can be grounded in the personas and target audiences already defined in the brand profile.

Use them to answer:

• Are we visible for the people we are trying to reach?
• Which audience needs are strongly associated with our brand?
• Where does the engine recommend a competitor for an important use case?

Persona selection is itself a source of bias. A set dominated by executive buyers may miss how practitioners search, while a set based only on internal persona assumptions may not reflect real customer language.

4. Comparison prompts measure competitive positioning

Example: "BrandHalo vs Frontify for brand governance"

Comparison prompts show how answer engines distinguish your brand from alternatives. They can reveal attributed strengths, perceived weaknesses, missing proof points and the sources shaping competitive narratives.

Use them to answer:

• Which competitors are most often associated with us?
• How does the engine explain the difference?
• Which claims or capabilities need clearer supporting content?

Because your brand may be supplied in the question, these prompts can produce a high mention rate even when generic discovery is weak. Keep the Comparison result visible as a distinct positioning measure.

Read the categories as a diagnostic

The value comes from the pattern across all four topics. Strong Brand visibility with weak Generic visibility suggests the engine knows you when asked but does not reliably recommend you unprompted. Strong Comparison visibility with weak Persona visibility suggests you are present in vendor evaluations but not strongly associated with the audience problems you want to own.

There is no universally correct result across the four categories. The useful question is whether the pattern aligns with your strategy. Part three turns these categories into a balanced, documented prompt portfolio.