AI visibility is how often AI assistants name your brand when somebody asks a question in your category, and which sources they cite instead of you. It is measurable across repeated runs. NovaLab measures it once for $400, or continuously from $6,500 plus $3,000 a month.
What AI visibility measures
A search engine returns an ordered list. An assistant returns a written answer that names some brands and not others. AI visibility is the measurement of which brands get named, and it has four parts.
- Appearance. Whether your brand is named in the answer at all.
- Position. Where you sit in the answer. First recommendation, one of six options, or a footnote after the real answer.
- Framing. The words the assistant uses about you. Being named as the budget option and being named as the specialist are both appearances and they are not the same result.
- Citations. Which sources were used, by domain and by specific page. This is the actionable part, because it names the exact pages holding the position you want.
Those four are captured per run. A vendor reporting a single visibility score has compressed all four into one number, which discards the only part you can act on.
Why one check tells you nothing
Assistants are non-deterministic. Ask the same question twice and you can get two different answers with two different brand lists and two different sets of cited sources. Neither answer is wrong. They are two samples from a distribution.
A single run tells you what happened once. To know whether you appear in a category you need the same prompt run repeatedly, on a fixed schedule, with every run recorded. Then the question stops being whether you appeared and becomes how often, which is a number that can move and be compared.
The same logic applies to prompts. One question is one question. A category has hundreds of them, covering what the category is, how to choose, what things cost, which providers to consider, and direct comparisons between named competitors. Visibility on one is not visibility on the category.
Every assistant answers differently
Each assistant retrieves from its own sources and cites differently, so they have to be measured separately. The set worth measuring in most categories is ChatGPT, Claude, Gemini, Perplexity, Microsoft Copilot and Google AI Overviews.
A brand that appears in Perplexity can be absent from ChatGPT on the same question. Averaging the two into one figure hides exactly the finding you needed. NovaLab reports the split rather than the average.
What you can and cannot control
Controllable
- Whether crawlers can reach your pages at all, which is a robots and hosting question before it is a content one.
- Whether your entity is unambiguous, so your brand, products and people resolve to one record rather than several.
- Whether your pages state facts plainly enough to be quoted as a fragment without the surrounding page.
- Whether the pages currently being cited instead of yours have something yours does not.
Not controllable
- Which sources an assistant chooses to retrieve from on any given run.
- When an assistant changes retrieval behaviour, which happens without notice.
- Whether a given answer names three brands or thirty.
- Any promise that work will put you in an answer. NovaLab does not make that claim and neither should anyone else.
How NovaLab measures it
| Option | What it measures | Price |
|---|---|---|
| Visibility Snapshot | A prompt panel run once across the major assistants, showing where you appear today and who is cited instead. Part of AI and Search Visibility, credited against what follows. | $400 |
| Category Visibility Intelligence | Several hundred category prompts on a fixed schedule across six assistants, multiple runs each, reported in a live dashboard. | $6,500 then $3,000 / month |
The Snapshot is a photograph. Category Visibility Intelligence is a time series. Most companies should take the photograph first, because it costs $400 and answers whether the larger measurement is justified yet. Full detail is on the AI and Search Visibility page and the Category Visibility Intelligence page.
Questions people ask
What is AI visibility?
AI visibility is how often an AI assistant names your brand when somebody asks a question in your category, where you sit in the answer, how you are described, and which sources get cited in your place. It is a measurable property of a category rather than a score a vendor assigns you, and it differs from one assistant to the next.
How is AI visibility different from SEO rankings?
A ranking is one ordered list for one query on one search engine. An assistant answer is generated text that may name two brands or twenty, may cite sources or none, and may differ on the next run of the same question. Ranking is a position. AI visibility is a distribution across repeated runs.
Can you check AI visibility yourself for free?
You can ask an assistant a few questions about your category and read the answers. That tells you what happened on those runs. It does not tell you how often it happens, whether it is changing, or which pages are being cited instead of yours. The gap between a spot check and a measurement is repetition and record keeping.
Why does the same prompt give different answers?
Assistants are non-deterministic. The same question can return different wording, a different brand list and a different set of cited sources on consecutive runs. That is why NovaLab runs each prompt multiple times on a schedule. One run is an anecdote. Enough runs make a distribution you can compare week to week.
Which AI assistants should be measured?
The ones your buyers actually use, which in most categories means ChatGPT, Claude, Gemini, Perplexity, Microsoft Copilot and Google AI Overviews. Each is queried separately because they retrieve from different sources and cite differently. A brand that appears in Perplexity can be absent from ChatGPT on the same question, and averaging the two hides the problem.
Can an agency get my brand into AI assistant answers?
No, and anybody promising it is selling something they do not control. Assistants choose their own sources and change retrieval behaviour without notice. What is controllable is the evidence: whether your entity is unambiguous, whether your pages state facts plainly, whether crawlers can reach them, and which competitor pages currently hold the citation.
What does measuring AI visibility cost?
NovaLab has two entry points. The $400 Visibility Snapshot runs a small prompt panel once across the major assistants and reports where you appear and who is cited instead. Category Visibility Intelligence costs $6,500 to build and calibrate the panel, then $3,000 a month for scheduled runs and a live dashboard.
Does traditional SEO work still matter for AI visibility?
Yes, and largely the same work. Crawler access, clean entity and schema markup, plain factual pages and citable source material all feed both. NovaLab sells search engine and AI assistant visibility as one service for that reason. The difference is measurement, because an assistant answer cannot be read off a rank tracker.
Where to go next
Measurement is one half. The work that acts on it is the other. AI and Search Visibility covers the entity, schema and citation surface work, and what does SEO cost sets out every NovaLab price in one place. Every service and its method sits on the services overview.
Start with a Visibility Snapshot
$400. A prompt panel run across the major AI assistants showing where you appear today and which pages are cited in your place, plus a technical crawl and a top-20 query gap list against two named competitors. The fee is credited against what follows.
Book the Visibility SnapshotLast updated 14 September 2026 · NovaLab Digital · Dubai, UAE