Skip to main content

Prompt categories explained

Prompt categories explained: the five built-in categories, custom prompts, and grouping by topic and audience segment.

Why categories exist

Categories group your prompts by the kind of question they ask, so you can balance your set and read results by intent. There are five built-in categories plus a custom option.

The five built-in categories

  • Brand Awareness: questions that name your brand directly ("What is {brand}?", "Is {brand} any good?"). These show how accurately and positively assistants describe you.
  • Competitor Comparison: questions that pit you against rivals ("{brand} vs {competitor}", "alternatives to {competitor}"). These show how you fare head-to-head.
  • Industry Recommendation: broad, category-level questions that don't name you and have no specific use case ("What are the best CRM tools?"). These reveal whether you come up unprompted for your whole category.
  • Problem / Solution: questions that describe a pain point and ask what solves it, without naming brands. These catch you in buying-intent moments.
  • Authority: questions asking for the best option for a specific use case, buyer, or constraint ("best CRM for a small team on a budget"), without naming you. These show whether you're the go-to in concrete situations, not just the broad category.

Custom prompts

Use Custom when you don't want to pick a type. A custom prompt still runs in your scan, and LumaRank assigns it the closest category automatically on its next scan, so it stops being an odd one out and joins the rest of your results by intent. If you'd rather lock in a specific type, just choose one: a category you set yourself is always kept as-is.

A balanced starting point

Most brands want plenty of unbranded prompts (Industry, Problem/Solution, Authority) alongside Brand Awareness and Competitor Comparison, because unprompted mentions are the hardest and most telling. The AI generator proposes a balanced spread you can adjust.

Topics: group prompts by subject

Categories tell you the kind of question (its intent). Topics tell you the subject it is about: the themes you care about, like pricing, integrations, or security. They are a second, independent tag on each prompt, and they are optional.

For example, these two prompts are both Industry Recommendation questions, but about different subjects:

  • "What is the best chat platform for WhatsApp support?" is about WhatsApp.
  • "Which customer support tools are the most secure?" is about security.

Topics let you read your results by subject instead of only by question type, so you can answer "where do I show up, and where don't I?" at the level of the things your buyers actually care about.

How to use them:

  1. Add up to five topics in your project settings (for example Pricing, Integrations, Security).
  2. Generate prompts, and each one is tagged with the topic that fits best. You can also set or change a prompt's topic by hand.
  3. On the Prompts page, switch By type / By topic to regroup your set either way.
  4. After a scan, two panels break your results down by topic: Performance by topic (how often AI mentions you on each one) and Reputation by topic (how positively it describes you there).

Topics apply from the moment you add them, so prompts you have not tagged, and scans run before you had any topics, appear as Untopiced. If a project has no topics at all, the by-topic panels stay hidden, since there is nothing to compare.

Audience segments

If you sell to more than one kind of buyer (say, a small shop and an enterprise), add audience segments in your company or project settings. When you generate prompts, we vary the personas so your set reflects each audience instead of just one. Segments only shape the prompts we write for you; they don't add a separate score.