How to Check Your AI Visibility

"AI visibility" sounds abstract until you break it into things you can actually inspect. There is no dashboard equivalent to a Google Search Console ranking report — no single number a platform hands you. But the public information an AI assistant can draw on is concrete, checkable, and mostly outside the business's marketing department: they live in robots.txt, in schema markup, in review platforms, in page copy, and in directory listings.

This is the same five-signal method AEO Radar's free scan runs automatically. Here's how to check each signal by hand and what a pass or fail actually looks like.

The five checks

Signal 1 — Crawler access (the gate)

Before on-site content can appear in an answer, the relevant search crawler has to be able to retrieve it. Check the effective robots.txt policy for OAI-SearchBot, Claude-SearchBot, and PerplexityBot — the published search or retrieval crawlers for ChatGPT, Claude, and Perplexity. A blanket User-agent: * / Disallow: /, or a specific block for one of these agents, prevents that crawler from fetching your pages directly. A missing robots.txt file is not a failure: crawling is allowed by default. This is a gate on your own site only — blocking a crawler doesn't stop an AI assistant from describing your business using other public sources such as directory listings, review platforms, or press coverage, and it says nothing about whether other sites mention you. Training-only crawlers (like GPTBot or ClaudeBot) are a separate, usually irreversible privacy choice and are not treated here as a search-visibility signal.

Signal 2 — Structured schema (the clarity layer)

LocalBusiness or Organization JSON-LD markup can help an eligible AI system interpret what your business is, where it operates, and when it's open, without having to infer it from unstructured prose. That help is conditional: the markup has to be accurate and match the visible page content, and not every assistant's retrieval path actually consumes JSON-LD — its presence alone doesn't resolve ambiguity. Check by viewing your page source and searching for application/ld+json, or by running your homepage through Google's Rich Results Test. A controlled Ahrefs study (1,885 pages that added schema between August 2025 and March 2026, matched against 4,000 control pages, difference-in-differences methodology) found a statistically significant 4.6% decline in Google AI Overview citations after adding schema, and small, not-statistically-significant changes for Google AI Mode (+2.4%) and ChatGPT (+2.2%).¹ The study pooled all JSON-LD types together (Article, FAQ, Product, HowTo, Organization, etc.) rather than isolating LocalBusiness markup, and it doesn't explain the AI Overview decline. Schema is still worth having for accuracy and rich-snippet eligibility; it hasn't been shown to increase citations for any AI product tested here, so don't over-invest in it relative to the signals below.

Signal 3 — Reviews and UGC

SOCi's 2026 Local Visibility Index covered about 350,000 US locations. Recommended locations averaged 4.3 stars on ChatGPT, 4.2 on Perplexity and 3.9 on Gemini.² These are averages, not minimums, so RaveHQ treats them as context, not a pass/fail gate. Recency adds a separate customer-trust signal: BrightLocal's 2026 survey found 74% of consumers only trust reviews from the last three months, and 32% only trust reviews from the last two weeks.³

Signal 4 — Content structure (the extraction layer)

AI systems quote and cite content that directly answers a question in an extractable format — a question as a heading, followed immediately by a direct, specific answer. The 2024 GEO (Generative Engine Optimization) paper reports gains of up to 40% on its own benchmark, varying by domain.⁴ Check your own site's key pages: do they lead with vague, adjective-heavy marketing copy ("award-winning service in a welcoming environment") or with specific, checkable claims ("Invisalign Diamond Provider since 2019, 340 verified Google reviews")? The second format is what gets extracted and quoted.

Signal 5 — Directory presence (the sourcing layer)

Yext's API-derived analysis of 6.8 million citations found different source mixes by model: OpenAI model output drew 48.73% from controllable third-party listings, while Gemini drew 52.15% from first-party websites.⁵ Check whether your business has claimed, accurate profiles on relevant directories, then compare them with the sources shown in repeated direct-app answers.

What to do with the result

Once you've checked all five, compare the suspected gaps with the sources and facts shown in repeated answers. An unclaimed directory listing or vague page copy is worth correcting for accuracy and clarity. Use what you find to choose the next check.

RaveHQ Insights share our views, based on our research and the sources cited. They are general information, not legal, financial or professional advice; check what applies to your business before acting.