AI visibility audit: score any page on 47 AEO and GEO signals
An AI visibility audit checks whether a page is easy for AI engines such as ChatGPT, Perplexity and Google AI Overviews to read, trust and cite. Site Audit is AnswerixAI's audit for individual pages. Paste any public URL and in about 30 seconds it scores the page from 0 to 100 on 47 signals for AEO (answer engine optimization) and GEO (generative engine optimization). The signals span structure, content, authority, E-E-A-T and trust. The ones a page fails or flags become fixes, many written for that page with copy-ready text.
Site Audit
Acme Coffee · 47-signal AEO / GEO rubric
- Fetching the pageHTML, robots.txt, llms.txt
- Structure13 signals80
- Content11 signals55
- Authority8 signals56
- E-E-A-T8 signals53
- Trust7 signals71
- Writing AI fix recommendationsAI
Needs work
47/47 signals evaluated · about 30 seconds
Built for marketers, executives, builders, and developers worldwide.
What does the AI visibility audit check?
Every page is scored on the same rubric. We built it from published research, such as the GEO paper by Aggarwal et al. and Google's E-E-A-T (experience, expertise, authoritativeness and trust) guidelines. Each signal passes, warns or fails. A category's score is the average of its signals, and the total weights the five categories. When we can't evaluate a signal on a page, we leave it out instead of counting it as zero. The report says how many of the 47 were checked.
How easily an AI engine can parse the page: heading depth, H1 quality and H2 coverage, and JSON-LD (structured data that labels what a page is about) presence, validity and type relevance. It also checks FAQ schema, lists, tables, image alt text, Open Graph metadata, internal links and page weight, with pages over 500 KB flagged.
Structure
13 signals · 25% of the total score
- Structural depth (GEO-SFE)high impactPassed
- JSON-LD presencehigh impactPassed
- FAQ schemamediumFailed
- H1 qualitymediumWarning
- JSON-LD type relevancemediumWarning
- H2 coveragemediumPassed
- Internal linkingmediumPassed
- JSON-LD validitymediumPassed
- Open Graph metadatamediumPassed
- Alt-textstandardWarning
- Page weightstandardPassed
- ListsstandardPassed
- TablesstandardPassed
What do you get back from each audit?
One report per URL: the score, what failed and why, and the fix, ready to hand to whoever edits the page.
Audit report
acmecoffee.com/guides/espresso-machines-under-500
Needs work
47/47 signals evaluated
Category breakdown
Weight
- Structure80/100 · 13/1325%
- Content55/100 · 11/1125%
- Authority56/100 · 8/820%
- E-E-A-T53/100 · 8/820%
- Trust71/100 · 7/710%
The report opens with the page’s score from 0 to 100 and a grade: Good from 70, Needs work from 40, Poor below that. Beside it are the passed, warning and failed counts, how many of the 47 signals were evaluated, and a score for each category. Export any completed audit as a PDF or re-run it with one click.
Where a tailored fix helps a failing or warning signal, an AI model reads the page and writes a recommendation for its topic, ranked high, medium or low. Many come with a draft and a copy button: a sharper H1, a meta description, an answer-first opening, FAQ pairs with FAQPage JSON-LD, inline definitions or sources to cite.
Expand any signal to read what the audit found on this page, what the signal measures and why it matters for AI answers, followed by the page-specific fix or the rubric’s general one. Signals are grouped by category, and an Issues only filter hides everything that passed so you can work down the list.
Ask the AnswerixAI AI Agent to audit a URL and it shows the result as a score card in the chat. You can also call run_site_audit from any MCP (Model Context Protocol) client connected to AnswerixAI. The audit runs in the background and get_site_audit returns the scores and recommendations once it completes. These audits use the same monthly allowance as the dashboard.
Which pages should you audit?
Any page that loads publicly can be audited, not just your own. Every URL is scored on the same 47 signals, so a draft, a live page, a competitor's page and a page AI engines cite can be compared category by category.
- 01
Pages before they go live
Audit a draft on a staging or preview URL that loads without a login, and fix a missing answer, schema block or byline before launch.
- 02
Pages that already matter
Find the weak headings, buried answers and unsourced claims holding back the pages you most want AI engines to use.
- 03
Competitor pages
Run a rival’s URL through the same rubric and weights, then compare its five category scores with your own page’s.
- 04
Pages AI engines cite
Audit URLs that keep appearing as sources in Citations Monitoring to see which structure and evidence they share.
Recent audits
Acme Coffee · same rubric for every URL
- acmecoffee.com/guides/espresso-machines-under-500Your pageYour pageJul 663
- staging.acmecoffee.com/guides/cold-brew-at-homePre-publishPre-publishJul 558
- democoffee.co/compare/acme-coffeeCited pageCited pageJul 574
- samplebeans.com/espresso-buying-guideCompetitorCompetitorJul 347
- acmecoffee.com/subscriptionYour pageYour pageJul 169
How do you know the fixes worked?
Ship the changes, then audit the page again. The Site Audit page charts the score of every completed audit on your primary domain over 7, 30 or 90 days or all time. The latest category breakdown sits beside it, so you can see which part of the score moved.
Re-run in one click
Every report has a re-run button that audits the same URL again and opens the new result.
One point per day
Several audits on one day collapse to the latest, so the line shows real day-to-day change.
Your primary domain
The chart follows pages on the brand’s primary domain; other audits stay in the recent list.
Score over time
acmecoffee.com
Category breakdown
- Structure80/100
- Content55/100
- Authority56/100
- E-E-A-T53/100
- Trust71/100
How does a site audit work?
Five steps take a page from a URL to a measured improvement. Only audits that complete count toward your monthly allowance: 100 on Starter and Growth, 500 on Business.
- 1
Paste a URL
Any public page. Sites that block plain requests are loaded in a headless browser.
- 2
Score
47 signals in about 30 seconds, rolled up into five weighted category scores.
- 3
Prioritize
Start with high-priority fixes and the high-impact signals that failed.
- 4
Fix
Paste the drafts into the page or send them to whoever owns it.
- 5
Re-audit
Run it again and watch the score and categories move on the chart.
Explore Related Features
Turn audit findings into AI visibility
Find the pages worth fixing, brief new ones, see which sources AI engines cite and measure how your brand compares with competitors.
Content Intelligence
Find the prompts where AI answers favor competitors, rank each gap by AI volume, visibility and intent, then brief, draft and send the fix to your workflow.
AI Content Briefs & Outlines
Turn prompts where ChatGPT, Perplexity and Google AI Overviews cite competitors into briefs with a title, outline, keywords, CTA and competitor insights.
Citations Monitoring
Find the domains and URLs ChatGPT, Perplexity, Gemini and Google AI Overviews cite for your prompts, split into your site, competitors and third parties.
Competitor Benchmarking
See how often ChatGPT, Perplexity, Gemini and Google AI Overviews mention, cite and rank each rival, then compare every prompt and platform head to head.
Customer stories
What do teams do with an AI visibility audit?
“What impressed us was being able to move from a site audit to actual fixes. Instead of receiving a generic score, our team could see prioritized AEO and GEO issues, understand the visibility gap and decide what to address first.”
FAQ
AI visibility audit FAQ
An AI visibility audit checks whether a web page is easy for AI answer engines such as ChatGPT, Perplexity and Google AI Overviews to read, understand, trust and cite. AnswerixAI Site Audit does this for any public URL, scoring the page from 0 to 100 on 47 AEO and GEO signals and turning the signals it fails into fixes.
It checks 47 signals in five categories. Structure (13) includes headings, JSON-LD and FAQ schema. Content (11) includes a direct answer, query coverage and readability. Authority (8) includes citations, statistics and an author byline. E-E-A-T (8) includes first-person evidence and author credentials. Trust (7) includes HTTPS, canonical tags and AI bot access in robots.txt.
Each signal gets a score from 0 to 1. A category's score is the average of its evaluated signals, and the total weights the categories: Structure and Content 25% each, Authority and E-E-A-T 20% each, Trust 10%. Signals that can't be evaluated are left out rather than counted as zero. Pages score Good from 70, Needs work from 40 and Poor below that.
Yes, for most content and structure issues. An AI model reads the page and writes a recommendation for each failing or warning signal where a tailored fix helps, ranked high, medium or low priority. Many include a draft to copy, such as a new H1, a meta description, an answer-first opening paragraph, FAQ pairs with FAQPage JSON-LD, or sources to cite. Other signals show the rubric's general fix.
You can audit any URL that loads publicly: your own pages, a staging or preview link that needs no login, competitor pages and pages AI engines cite. AnswerixAI fetches the page directly and switches to a headless browser when a site blocks plain requests or shows a bot challenge. Pages that sit behind a login or paywall can't be audited.
About 30 seconds for most pages. The audit runs in the background: the page is fetched, the 47 signals are scored and the AI recommendations are written, then the report opens on its own. You can leave and come back, because every audit is saved to your recent audits list with its URL, date and score until you delete it.
Starter and Growth include 100 site audits a month and Business includes 500. The allowance resets when your subscription renews, and only audits that complete count toward it. Audits started from the AnswerixAI AI Agent or an MCP client use the same allowance. Every plan starts with a 14-day free trial.
Yes. Audit the competitor's URL the same way as your own. Both pages are scored on the same 47-signal rubric with the same category weights. Their totals and their Structure, Content, Authority, E-E-A-T and Trust scores are directly comparable, and each report shows exactly which signals the other page passes or fails. Both audits stay in your recent audits list.
Yes. Re-run an audit from its report after you change the page. The Site Audit page charts the score of every completed audit on your brand's primary domain, one point per day, over 7, 30 or 90 days or all time. It also shows the change in the range and the latest category breakdown. Audits of other domains stay in your recent audits list.
Yes. Ask the AnswerixAI AI Agent to audit a URL and it shows the result as a score card in the chat. You can also call run_site_audit from any MCP client connected to AnswerixAI, then read the score, category scores and recommendations with get_site_audit once the audit completes. Each run uses one audit from your monthly allowance.
A traditional SEO audit crawls a whole site for ranking issues such as broken links and indexing. Site Audit scores one URL at a time on 47 signals that decide whether AI engines can read, trust and cite that page, such as a direct answer, JSON-LD, cited sources, E-E-A-T markers and AI bot access in robots.txt.
See How AI Engines Read Your Pages
Use AnswerixAI to grow AI Search visibility with AEO and GEO.