AI visibility checker

What does the AI visibility checker measure?

The AI visibility checker measures how easily machines can reach and read one page of your site. It reads your robots.txt rules for six AI search and assistant crawlers, your structured data, how much of the page is in the HTML before JavaScript runs, whether the business is clearly named, llms.txt, and whether questions on the page get direct answers.

It is check 7 of the 8 in Convertica's free CRO audit, named "AI visibility and agent readiness". The app loads your page in a real browser at phone and desktop size, fetches the page's raw HTML, your robots.txt and your llms.txt, and measures these parts:

An AI review then reads the page text and screenshots and answers one question: could an AI agent state, from this page alone, what is sold, who it is for, the price or how to get it, the policies and the name of the business? The review is told never to claim that any AI assistant or search feature cites or shows the page, because the check has no data for that.

These parts are Convertica's own view of what makes a page easy for a machine to read. They are not a list of ranking factors. Google, for example, says there are no additional requirements to appear in AI Overviews or AI Mode beyond being indexed and eligible to be shown with a snippet, and that no special structured data or AI text file is needed.

The same run scores seven other checks on the same page: first impressions, calls to action, forms and friction, trust and credibility, mobile experience, page performance and SEO fundamentals. The landing page analyzer is the same engine pointed at a landing page.

How is the AI visibility score calculated?

The AI visibility and agent readiness score is a number out of 100 for the one page you entered. The measured parts in the table make up 70% of it and the AI review 30%. Convertica labels 80 or more good, 50 to 79 needs work, and below 50 a problem.

• Crawler access counts search and assistant crawlers only. The score uses the share of those six crawlers that may fetch the page.
• Blocked training crawlers do not lower the measured score. A rule that blocks a training crawler such as GPTBot or ClaudeBot is reported as a separate, low impact finding, because that is a business choice.
• It feeds one overall score. The overall score in your report is the plain average of the checks that could be scored, and every check weighs the same.

The parts, their weights and the pass marks are Convertica's choices, not taken from research or an industry benchmark. A high score means fewer known gaps on the page. It does not predict that any AI engine will mention, cite or recommend you.

What this AI visibility checker does not do

This AI visibility checker reads your site and nothing else. It does not send a question to ChatGPT, Perplexity, Gemini, Claude or Google, so it cannot tell you whether those engines mention or cite you, how often, for which questions, or which competitors they name. Finding that out means asking the engines, by hand or with a citation tracking product.

The two answer different questions. Start with the readiness check: a page that AI crawlers cannot fetch or read gives an engine nothing to quote, whatever a tracker reports. Tracking is worth adding once nothing on your own site is in the way.

Other limits worth knowing:

• It cannot see blocks outside robots.txt. A firewall, CDN or bot protection rule can refuse an AI crawler that robots.txt allows. The audit fetches your pages as itself, not as GPTBot or ClaudeBot.
• It cannot confirm that crawlers visit. That is in your server or CDN logs.
• It reads one page per run. It is not a crawl of the whole site, and a homepage and a product page can score differently, so check the page types that matter to you.
• It is not a ranking check. It says nothing about where you rank in Google or how much traffic AI search sends you.

Which AI crawlers should you allow in robots.txt?

Allow the AI crawlers that match what you want from each company. To be shown in an AI search product, allow its search crawler, such as OAI-SearchBot, Claude-SearchBot or PerplexityBot. For Google's AI Overviews the crawler is Googlebot itself. Training crawlers such as GPTBot and ClaudeBot are a separate choice about how your content is used.

The table lists the 20 crawlers and control tokens the check on this page reports on, from nine operators. Each is grouped by what its operator's own documentation says it is for, and no further.

Sources, each operator's own page:

• OpenAI: Overview of OpenAI Crawlers
• Anthropic: Does Anthropic crawl data from the web, and how can site owners block the crawler?
• Perplexity: Perplexity Crawlers
• Google: Google's common crawlers
• Google: AI features and your website
• Apple: About Applebot
• DuckDuckGo: Is DuckAssistBot related to DuckDuckGo?
• Mistral AI: Mistral crawlers
• Amazon: About Amazonbot
• Common Crawl: CCBot

Operators add, rename and retire crawlers, so read the linked page before you rely on a rule. Four things in the table are easy to get wrong:

• Google's AI Overviews are not a separate crawler. Google says robots.txt rules for Googlebot are the control for how a site is crawled for Search, AI features included. Google-Extended only governs Gemini training and grounding, and Google says it does not affect inclusion or ranking in Search.
• Training and search are separate settings. OpenAI, Anthropic and Mistral each document one crawler for training and another for search. Blocking the first does not block the second, unless one rule catches both.
• Some tokens are not crawlers. Google-Extended and Applebot-Extended never request a page. They are labels a robots.txt file uses to say how content fetched by Googlebot or Applebot may be used.
• User-triggered fetchers may not follow robots.txt. OpenAI says the rules may not apply to ChatGPT-User, Perplexity says Perplexity-User generally ignores them, and Amazon says Amzn-User may not follow all of them, because a person started the request.

One common setup, shown to illustrate the syntax and not as a recommendation, keeps AI search crawlers in and training crawlers out. Each named group repeats the rules from the * group, because a crawler with its own group does not read *:

User-agent: *
Disallow: /cart/

# AI search: same rules as everyone else
User-agent: OAI-SearchBot
User-agent: Claude-SearchBot
User-agent: PerplexityBot
Disallow: /cart/

# AI training: not wanted
User-agent: GPTBot
User-agent: ClaudeBot
User-agent: Google-Extended
Disallow: /

How to improve your AI visibility after the check

Improve AI visibility in the order the check exposes problems: remove what blocks AI crawlers, put the facts in the HTML, state who you are and what the page offers in valid markup, then answer the questions buyers ask in plain sentences. These steps make a page easier to read and quote. None of them guarantees a citation.

1. Unblock the crawlers you want. Fix robots.txt, then ask whoever runs your CDN or firewall whether its bot rules treat AI crawlers the same way.
2. Put the facts in the HTML. Price, availability, delivery, returns and contact details should be in the page source as text, not added by scripts or shown only in images.
3. Name the business in markup. Add Organization markup with your name, logo and sameAs links to your official profiles.
4. Use markup that fits the page. Product with an Offer and a price on product pages; Service, SoftwareApplication, Article or FAQPage where they describe the page. Keep it identical to what a visitor sees.
5. Answer questions directly. Under a heading phrased as a question, start with a short answer that makes sense on its own, then add the detail.
6. Add llms.txt if it is cheap for you. It is optional. The llms.txt guide covers what it is and what it does not do, and the llms.txt generator builds and validates the file.

What other sites say about you also shapes what AI engines repeat, and no check of your own site can measure that. Convertica's generative engine optimization services page explains how that work is done and measured.

Being read is half of it. When an AI answer sends a person, or an AI agent, to your page, the page still has to convert, which is why the same run scores your first screen, calls to action, forms, trust, mobile experience and speed. See AI conversion rate optimization for how the two fit together, and agentic commerce for what changes when the visitor is an agent that buys.