What is AI search visibility and why does it matter?

AI search visibility is the degree to which an AI system; such as ChatGPT, Perplexity, Claude or Google Gemini Overviews; can parse, trust and cite your website content in generated answers. A page that scores poorly for AI search visibility is unlikely to appear as a source or citation in AI generated responses, regardless of its traditional search ranking. AI search visibility depends on structured data quality, heading clarity, content depth, metadata completeness and machine readable formatting.

How does AiVIS.biz calculate a search visibility score?

AiVIS.biz runs a multi-stage audit pipeline: it crawls the target URL with a real browser, extracts structural signals (headings, schema, meta tags, body content, links), runs technical checks, then passes the evidence to a primary AI model for scoring. On Signal and Score Fix tiers, a second AI model peer-critiques the findings and a third validates the final result. The 0-100 CITE LEDGER score is a weighted composite across seven dimensions: Schema and Structured Data (20%), Content Depth (18%), Technical Trust (15%), Meta Tags and Open Graph (15%), AI Readability (12%), Heading Structure (10%), and Security and Trust (10%). Each finding is tied back to real page evidence.

What is the difference between AiVIS.biz and traditional query and answer engine tools?

Traditional SEO tools optimize for search engine crawlers and ranking signals like backlinks or keyword density. AiVIS.biz focuses specifically on answer engine readiness; whether AI systems can extract, trust and cite your content in generated responses. AiVIS.biz evaluates machine readability, entity clarity, schema coverage and citability rather than link authority. Every recommendation is evidence-linked, meaning it references the specific crawled signal that triggered it, not a generic checklist. The platform also tracks brand presence across citation sources and discussion platforms: Answer Engine Optimization (AEO).

What sources does brand mention tracking monitor?

Brand mention tracking scans 19 public sources: Reddit, Hacker News, Mastodon, DuckDuckGo site search, Bing site search, Google News RSS, GitHub repositories, Quora, Product Hunt, Stack Overflow, Wikipedia, Dev.to, Medium, YouTube, Lobsters, Bluesky, Twitter/X, Lemmy and GitHub Discussions. Available on the Alignment tier and above.

What plans does AiVIS.biz offer and what is included?

AiVIS.biz offers five tiers. Observer (free) includes 3 scans per month powered by open-weight AI models. Starter ($15/month) includes 15 scans per month with full recommendations, implementation code, PDF exports, and shareable links. Alignment (Core, $49/month) includes 60 scans per month with paid AI models, competitor tracking, brand mention scanning, reverse engineering tools, and WebMCP access. Signal (Pro, $149/month) includes 200 scans per month with a triple-check pipeline, citation testing, niche ranking, scheduled citation tests, team workspaces with 25 seats, full MCP/API access, and white-label reports. Score Fix (AutoFix PR, $299 one-time) uses higher capacity models and generates automated GitHub PRs with 250 credits.

How AI Models Read Websites — Technical Breakdown | AiVIS.biz

Understanding exactly how AI reads your site is the most valuable thing you can do before optimizing anything. The process is completely different from what a human sees.

Step-by-step: how AI models access and parse a webpage

Step 1 — HTTP request: The AI crawler sends an HTTP GET request with a specific user-agent (GPTBot, ClaudeBot, PerplexityBot). Your server responds with the HTML document.

Step 2 — HTML parsing: The crawler parses the raw HTML. It looks for content in the <body>, structured data in <script type='application/ld+json'>, metadata in <head>, and heading hierarchy throughout.

Step 3 — Content extraction: Text is extracted from semantic elements: <h1>, <h2>, <p>, <article>, <section>. Navigation, footers, and ads are typically de-weighted or discarded.

Step 4 — Representation: The extracted content is compressed into a latent representation (for training data) or structured fragments (for real-time retrieval).

Step 5 — Attribution: The crawler records the source URL, domain, and any entity metadata from JSON-LD.

What AI models cannot do when reading websites

Execute JavaScript. Most AI crawlers do not run client-side JavaScript. SPA content that appears after JS execution is invisible.

Authenticate. Content behind login or cookie consent walls is inaccessible.

Process images without alt text. Images without alt attributes deliver no text signal.

Interpret layout. AI crawlers extract text content; visual hierarchy and design are irrelevant.

Frequently Asked Questions

Do all AI crawlers read websites the same way?: Roughly the same. The core pipeline (HTTP request → HTML parse → text extraction → representation) is shared. Differences are in JavaScript execution capability (some newer crawlers handle JS), crawl frequency, and attribution weighting.
Can I see what an AI crawler sees on my page?: Use 'View Source' in your browser to see the raw HTML — that is the closest approximation of what an AI crawler receives. AiVIS.biz replicates the crawl experience more precisely using a headless browser with AI crawler user-agents.