Category Visibility
Category Visibility is a Discovery Optimization research methodology that answers one question: when someone asks about your category, do you come up? Not "do you rank for your own brand name." Anyone surfaces for their own name. The real test is whether an answer engine recommends you to a prospect who does not know you exist yet and asks a plain category question.
Where Search Performance Analysis reads the signals you already own, Category Visibility probes a signal you do not control, what the LLMs say, by asking them the questions your buyers ask.
First, what is a "category"?#
Your category is the space you compete in, described the way buyers describe it, not the way you describe yourself. It is the job to be done, not your brand or your feature list. "Project management software for engineering teams" is a category. "Acme Tracker, the AI-native planning workspace" is not, that is your brand. Buyers who do not know you yet search and ask in category terms, so that is the language the answer engines have to associate with you.
If you are not sure how to phrase yours, that is fine, and it is the first thing to fix (see "Before you run it" below).
When to use it#
This is the default first session in Rampify Discovery, and the right starting point for almost everyone:
- You want a baseline for how visible you are in AI answers, before any optimization.
- Your brand is not a household name in your space yet, so buyers describe what they need in category terms instead of asking for you by name.
- You want to know whether the answer engines even know you belong in the conversation.
It needs almost nothing to run, no competitor list and no deep profile, which is why it is first.
How the method works#
The session runs from a persona, a stand-in for a real prospect in your category, and the prompts are written from that point of view. Your brand is deliberately kept out of frame so the test measures organic surfacing, not recall.
The methodology generates five questions a prospect might ask a coding agent or ChatGPT before they know any brands, mixed across how specific they are: one broad, two medium, and two narrow and qualified. None of them name your brand or domain. That is the whole point. They are written the way people actually talk to an assistant, with context and follow-up, not as one-line search queries.
Each question is put to the answer engine in search-enabled mode, so you are measuring what a real user gets today, training knowledge plus live retrieval, not just what the model memorized. This is what makes it a measurement of the live web, not a trivia quiz.
For each answer, the session records whether your brand came up, in what company, and how it was framed. Five prompts give you a read across the category, not a single anecdote.
A worked example#
Think back to the last time you researched buying something through an AI assistant. You did not type a terse keyword. You opened with a plain question and then added context until the assistant narrowed things down for you.
Say your category is project management for engineering teams, the space where Linear, GitHub Projects and Issues, Jira, Asana, and Notion all compete. A real prospect does not search "best PM tool." They have a conversation that looks more like this:
Broad: "What are engineering teams using these days to plan work and track issues?"
Medium, with context: "I run a 15-person engineering team. We basically live in GitHub and Slack, and we are tired of keeping the roadmap in spreadsheets. What project management tools actually fit a developer-first workflow?"
Narrow and qualified: "We are a Series A startup, around 25 engineers, already on GitHub. We want something fast and keyboard-driven that engineers will not resist, with a clean API and solid issue tracking. We do not need heavyweight enterprise PM. What should we shortlist?"
Notice what is happening. None of these name a brand, and each one carries the kind of context (team size, existing stack, requirements, what they are not looking for) that an assistant uses to narrow the field. That context is the difference between AI search and a traditional one-shot query, and it is exactly what Category Visibility reproduces.
Now the test: when the answer engine responds, are you in the list? If you sell a developer project management tool and the assistant names Linear and Jira but never you, you have a category-visibility gap. And the three altitudes tell you different things:
- Missing from the broad question means a top-of-funnel awareness gap. New buyers start here.
- Missing from the medium question means you are not connected to the workflow and constraints buyers describe.
- Missing from the narrow question, even when your product is a great fit for those exact requirements, is the most painful and the most fixable: the engine does not yet know what you are specifically good at.
If you ask whether your brand is good for X, every model will say yes. It is being agreeable, not informative. Category Visibility only means something because the prompts never mention you. Surfacing in a no-brand category answer is earned. Surfacing in a brand-named one is just politeness.
How to read the results#
- You surface on broad and narrow alike. Strong category presence. Move on to sharper methodologies like Comparison Readiness or Scenario Surfacing (see the guide hub).
- You surface only on narrow, qualified questions. You are known for a specific niche but invisible in the general category. The opportunity is broad-category content and authority.
- You do not surface at all. The answer engines do not yet associate you with your category. This is the most common early-stage result, and the most actionable. It means the foundational content and third-party signals that LLMs read are not there yet.
A gap here is not a verdict. It is a worklist.
What to do about gaps#
When you are missing from category answers, the levers are the same ones that make content extractable and authoritative for LLMs:
- Publish clear, category-level content that states plainly what you do and who it is for (see AI Visibility for how answer engines read a page).
- Earn the third-party mentions and structured, extractable signals models lean on.
- Re-run the session over time to watch the gap close. The same five questions make a clean before-and-after.
Before you run it: your category lives in your Business Profile#
The questions are generated from your Business Profile, specifically the category you operate in. If that is thin or missing, the prompts will be vague and the results will be too. Flesh it out first. From your AI tool you can say "Enrich my business profile" and Rampify will auto-fill your category and positioning from your site's public pages, so the session has a sharp category to probe.
Run it in Rampify#
Category Visibility runs as a Discovery session. It takes about five minutes and needs no setup beyond a connected project. Once your business profile has a category, use the Connect Rampify button at the top of this page to start it. The agent generates the persona-shaped questions, probes the answer engines with search, records where you surface, and turns the gaps into specs you can act on.
See if the answer engines name you
Run a Category Visibility session in about five minutes and get a baseline for how you show up in ChatGPT, Claude, and Perplexity, before your buyers know your name.
Start free