Prominara
Features

AI Visibility Audit

Run an AI visibility audit on any URL. See what we check, how findings are scored, which fixes matter most, and how crawl discovery and sitemaps work.

Last updated

The AI Visibility Audit analyzes any URL and provides a comprehensive score with actionable recommendations for improvement.

Running an Audit

  1. Enter a URL in the Quick Scan widget or from a site's page
  2. Wait 10-30 seconds for analysis
  3. Review your score and recommendations

What We Analyze

Our audit examines 40+ signals across four categories:

  • Technical Readiness (30%) - AI search crawler access, noindex and snippet controls, main content in the HTML without JavaScript, redirects and status codes, crawlable links, response time
  • Content Structure (30%) - Answer in the first 30% of the page, self-contained citable passages, section length, and on product and pricing pages whether price, audience and inclusions are stated in the text
  • Entity & Topics (25%) - Specific headings, definitions, figures and quotes attributed to a named source
  • Authority Signals (15%) - A named author on blog posts, publication and update dates

Some conditions cap the overall score regardless of content: a page set to noindex, or one that blocks an AI search crawler in robots.txt or at the CDN, is capped at 40, and a page whose main content only appears after JavaScript runs is capped at 50. A nosnippet or max-snippet:0 caps only the Google AI Overviews / AI Mode scores at 40, since Google and Bing are the engines that document honoring it. When a site refuses our direct request, we read a browser-rendered copy and mark the raw-HTML check as not verified instead of capping it.

Some checks are reported for information and never raise the score, because there is no evidence they affect AI citations: llms.txt, FAQ blocks, tables outside comparison pages, word count, lists, paragraph length, internal links, images, named entities, schema markup, trust pages, semantic HTML, alt and anchor text, meta description and Open Graph tags, page weight, URL structure and mobile optimization. Schema markup still helps search engines read the page accurately; it has not been shown to raise AI citations.

Site-level Findings

Some problems don't belong to a single page — a missing or broken sitemap, a robots.txt rule blocking a large share of your site, or broken internal links. These are surfaced on the Site Overview alongside per-page issues and subtract a bounded penalty from the site score (capped at 15 points total), so a site whose sitemap index 404s is no longer scored identically to one with a perfect sitemap.

  • No sitemap.xml found - Critical. AI engines that can't read a sitemap only find pages reachable by following links.
  • Sitemap unreadable / partial - Reported when your sitemap exists but references files that fail to load, with the count of broken references.
  • robots.txt blocking discovered pages - Counts how many pages we discovered that your robots.txt disallows. Called out as important above 25%; informational below.
  • robots.txt problems - robots.txt that returns an error, cannot be reached, has lines crawlers cannot parse, or does not point to your sitemap with a Sitemap: line.
  • Sitemap URLs that are not indexable pages - Sitemap entries we crawled that redirect, return an error, carry noindex, or point their canonical to another page.
  • Unreliable lastmod dates - Sitemap <lastmod> values all set by a deploy rather than by content changes, or older than the page's own update date.
  • Duplicate titles - Pages sharing the same title tag.
  • Broken internal links - Links in your main content that point to pages returning errors.
  • Core Web Vitals - Slow field data for real mobile users, from the Chrome UX Report, when Google has data for your site.
  • No llms.txt found - For information only; it does not affect your score.
Custom-URL scans (where you paste a list of pages instead of running discovery) don't probe sitemap / robots / llms.txt, so site-level findings are omitted for them by design.

Crawl Discovery

The Site Crawl tab separates three counts so you can see how much of your site each plan actually inspects: Discovered is every URL we found (sitemap + internal links + Search Console + AI citations), Audited in depth is the slice your plan paid to score in detail, and Not audited is discovered pages that fell outside the deep-audit quota. Discovery is cheap and uncapped up to a technical ceiling; the deep-audit quota only decides how many pages get scored.

Sitemap Generator

The sitemap generator reads your full URL inventory rather than only the audited slice — the old generator claimed to be "every page" while listing the 10, 50 or 200 pages the plan audited. Excluded URLs are counted with a reason (redirect, HTTP error, noindex, robots-blocked, off-host, duplicate, or citation-only) so you can tell why a discovered URL didn't reach the file. Tracking parameters like ?utm_source=…are stripped before URLs are emitted, so a page cited by AI as /contacto?utm_source=openai ships as its canonical /contacto.<lastmod> is stamped only when the site itself declared one or a content hash actually changed — never fabricated from the crawl date.

Understanding Results

Overall Score

A 0-100 score indicating AI citation readiness. 70+ is considered good.

Category Breakdown

See how you score in each of the four categories to identify specific areas for improvement.

Findings

Open findings are listed by severity, then by priority and impact. Checks you already pass collapse into a single summary row.

  • Critical - Blocking AI discovery (crawler access, content in the HTML, snippet controls, redirects and status); fix first
  • Important - Significant impact on citation likelihood
  • Suggested - Meaningful but lower-priority improvements

Each finding also carries an expected impact (high, medium, low) and an effort estimate (quick, moderate, significant).

What a Finding Shows

Expand a finding and you get the target, not just the diagnosis:

  • On this page now - The measured value plus the exact text or markup we read from your page
  • Optimal - The structure that should replace it, written against this page's own headings, comparison subjects, and URL path
  • How to fix - A numbered checklist for the change
  • Snippet - Where the fix is a block of code, a copy-pasteable JSON-LD, HTML, or robots.txt block, pre-filled with what the scan already read from your page
Values we could not read from your page are left as explicit placeholders (for example [category] or YYYY-MM-DD). Replace them before shipping — markup that contradicts the visible page costs trust instead of building it.

Generated Assets

From a finding in your dashboard, one button opens the generator that produces the missing asset — or opens the AI Optimizer when the fix is a content change:

  • llms.txt - AI-readable summary of your page/site (optional; it does not affect your score)
  • Schema markup - Structured data for the page type
  • robots.txt - AI crawler access rules
  • Sitemap - XML sitemap from your crawled pages

Scan History

All scans are saved to your site's history. Track score changes over time and see which optimizations made the biggest impact.

Was this page useful?