How to Audit Your Website for AI Search Visibility
Paste any URL into the free FindUtils GEO Analyzer and you'll get a 25-signal audit covering AI crawler access (16 bots including GPTBot, ClaudeBot, PerplexityBot), structured data, llms.txt, listicle format, and E-E-A-T signals — plus a citability score that estimates how likely ChatGPT, Claude, Perplexity, and Google AI Overviews are to quote your page. This guide walks through the full audit process, explains every check, and shows you exactly what to fix first.
If you only have 10 minutes, skip to the Minimum Viable GEO Fixes section at the end — it's the three changes that move citability more than everything else combined.
Why GEO (Generative Engine Optimization) Matters in 2026
AI search is where the next decade of organic traffic comes from, and the signals are not the same ones traditional SEO optimizes for.
- ChatGPT search cites its sources inline. Missing from the citation = missing from the answer = missing from the click.
- Perplexity builds every answer from 3–8 cited URLs. Sites with stacked schemas and listicle structure dominate.
- Google AI Overviews appear above the organic results for 40%+ of informational queries. Overviews pull from structured FAQ content and numbered lists.
- Claude, Gemini, and Bing Copilot follow the same patterns. What works for one usually works for all.
Traditional SEO tools measure keyword rankings. GEO measures citation-readiness — a different signal set that AI engines publicly document and rank by.
What the GEO Analyzer Actually Checks
The FindUtils GEO Analyzer runs 25+ signals grouped into four categories. Each signal is weighted by severity: critical (10 pts), medium (5 pts), minor (2 pts).
| Category | What it covers | Why AI engines care |
|---|---|---|
| AI crawler access | Access for 16 major AI bots (GPTBot, OAI-SearchBot, ClaudeBot, PerplexityBot, Google-Extended, Applebot-Extended, Bytespider, CCBot, etc.) | If a bot cannot read your page, the model cannot quote, link, or recommend you. |
| GEO signals | Listicle format, schema stacking (Article + ItemList + FAQPage), llms.txt presence + validity, author E-E-A-T | The formats AI models demonstrably cite most often. |
| Content | JSON-LD structured data, heading hierarchy, word count, semantic HTML, alt text, Open Graph | Content depth and machine-readability — AI parses headings + schema before text. |
| Technical | HTTPS, meta title/description, H1 presence, mobile viewport, canonical, robots.txt, sitemap.xml, robots meta | The SEO foundation AI engines still rank on — Bing alignment is 87%+ for ChatGPT citations. |
The overall score is a weighted average; the AI citability score is a separate composite that weights GEO signals more heavily and deducts points per blocked AI crawler.
Step-by-Step: Running Your First GEO Audit
Step 1: Paste Your URL
Go to the FindUtils GEO Analyzer and paste the URL of any page you want to audit — your homepage, a pillar article, a product page. The tool is free, browser-based, and requires no signup. Analysis takes about 3 seconds.
Never tested GEO scoring before? Click one of the example chips (Stripe, GitHub, Vercel, Anthropic) to see what a good score looks like before running your own site.
Step 2: Read the Score Gauges
The report shows five gauges at the top:
- Overall score — weighted average across all categories, with a letter grade (A+ to F)
- AI citability — how likely AI models are to quote this page (0–100)
- Technical, Content, GEO signals — category breakdowns so you know where to invest
An overall score of 85+ earns an A grade. Below 70 means you have real gaps that AI search is actively penalizing.
Step 3: Fix Priority Issues First
Scroll to the Priority fixes section. Every failing critical or medium check is listed with:
- The specific issue (what's missing or broken)
- A paste-ready code snippet (HTML, JSON-LD, or Markdown you can drop in)
- A one-line note explaining why it matters
Click Copy AI prompt to get a complete prompt pre-loaded with every fix, ready to paste into Claude, ChatGPT, or Cursor. The Apply with Claude button opens claude.ai with the prompt filled in — you can have an AI apply every fix to your codebase in minutes.
Step 4: Check the AI Crawler Grid
Scroll to AI crawler access. You'll see a grid of 16 AI bots, each marked allowed (green) or blocked (red). If any are red, your robots.txt has a rule blocking them — often accidentally via a blanket User-agent: * + Disallow: /.
The FindUtils Robots.txt Generator creates a correct robots.txt that allows all major AI crawlers by default. Paste your current file in and see what needs changing.
Step 5: Compare Against Competitors
Switch to the Compare tab and enter up to 4 URLs. The GEO Analyzer runs the full audit on each and produces a side-by-side table with category winners highlighted. Use this to benchmark against sites that ChatGPT and Perplexity currently cite for your target queries — often their scores tell you exactly where to invest.
Step 6: Export the Report
Three one-click export options:
- Copy as Markdown — self-contained report with scores, fixes (with code snippets), AI crawler matrix, all checks, and a ready-to-use AI prompt at the bottom. Paste into any editor, issue tracker, or AI chat.
- Export JSON — machine-readable report for CI pipelines, dashboards, or Zapier workflows.
- Copy share link — shareable URL (with
?url=...) so teammates can re-run the audit without re-typing.
The Five Highest-ROI GEO Fixes
Not every failing check is equally important. Based on what AI engines publicly document (and what competitor analysis consistently shows), these five moves have the biggest impact on citability:
1. Unblock AI crawlers in robots.txt
This is the highest-ROI fix of all. If GPTBot, ClaudeBot, or PerplexityBot is blocked, you are invisible to that AI engine — period. Use the FindUtils Robots.txt Generator to build a correct file.
User-agent: GPTBot Allow: / User-agent: ClaudeBot Allow: / User-agent: PerplexityBot Allow: / User-agent: Google-Extended Allow: / # Default User-agent: * Allow: / Sitemap: https://example.com/sitemap.xml
2. Ship an llms.txt file
llms.txt is the emerging AI-search sitemap — a Markdown file at /llms.txt listing your key pages. It's a minor signal on its own but correlates strongly with citation-ready content because sites that bother with it also tend to stack schema and write listicle-friendly content.
# Your Brand > One-sentence description of what you do. ## Docs - [Getting Started](https://example.com/docs/start): Setup walkthrough - [API Reference](https://example.com/docs/api): Complete API docs ## Guides - [How to X](https://example.com/guides/x): Solves specific problem - [Best Practices](https://example.com/guides/best): Production recommendations
Place at your site root. The FindUtils GEO Analyzer checks for presence and basic validity on every audit.
3. Stack Article + ItemList + FAQPage schemas
Pages with all three schemas get cited roughly 3× more by AI engines than pages with just one. The JSON-LD Generator creates each block, and the Structured Data Tester validates the combined output.
<script type="application/ld+json">
{
"@context": "https://schema.org",
"@graph": [
{
"@type": "Article",
"headline": "Article title",
"author": { "@type": "Person", "name": "Author Name" },
"datePublished": "2026-04-20"
},
{
"@type": "ItemList",
"itemListElement": [
{ "@type": "ListItem", "position": 1, "name": "Item 1", "url": "..." },
{ "@type": "ListItem", "position": 2, "name": "Item 2", "url": "..." }
]
},
{
"@type": "FAQPage",
"mainEntity": [
{
"@type": "Question",
"name": "Your question?",
"acceptedAnswer": { "@type": "Answer", "text": "Direct answer." }
}
]
}
]
}
</script>The FindUtils FAQ Schema Generator builds the FAQPage block from a plain list of Q/A pairs — the fastest path to adding this if you already have an FAQ section on the page.
4. Convert content into listicle format
Numbered headings, ordered lists, and comparison tables are the #1 format AI engines cite. Based on Cloudflare's 2.6B-citation analysis, list and comparison content represents 25.37% of all AI citations. Every important page on your site should include at least one:
- Top N heading structure (
## Top 10 X for 2026with### 1. Name — description) - Ordered list with self-contained items
- Comparison table (always wins)
Check heading hierarchy with the FindUtils Heading Structure Analyzer to make sure your listicle format is crawlable.
5. Add a visible author byline and author in JSON-LD
E-E-A-T (Experience, Expertise, Authoritativeness, Trustworthiness) is a direct AI-ranking signal, especially for YMYL (your money, your life) content. A visible byline plus an author property in Article JSON-LD pointing to a Person page is the minimum bar.
"author": { "@type": "Person", "name": "Author Full Name", "url": "https://example.com/authors/name", "sameAs": [ "https://twitter.com/handle", "https://linkedin.com/in/handle" ] }
The author's sameAs array lets AI engines cross-reference their identity across platforms — strong signal that the byline is real, not AI-generated.
Common GEO Audit Mistakes
Mistake 1: Chasing overall score instead of citability
The overall score rewards traditional SEO foundations. The AI citability score is what determines whether you get cited. A page can score 85 overall but 45 on citability if you nailed meta tags but skipped schema stacking. Prioritize the citability number.
Mistake 2: Unblocking some AI bots, missing others
Most robots.txt files block bots via a catch-all User-agent: * rule, then add an explicit Allow for Googlebot. That blocks every AI crawler. Audit with the GEO Analyzer's full 16-bot grid — not just GPTBot and ClaudeBot, but also OAI-SearchBot (ChatGPT search, different from GPTBot), Google-Extended (Gemini), and Applebot-Extended (Apple Intelligence).
Mistake 3: Adding JSON-LD without author
Article schema without an author property is a net negative signal in 2026 — it suggests AI-generated content to ranking models. Always include an author with @type: Person, name, url, and sameAs.
Mistake 4: Deploying llms.txt in HTML format
llms.txt must be served as text/plain or text/markdown, not HTML. Some CMSes accidentally wrap it in a layout. Check with the GEO Analyzer's llms.txt validator — it confirms both presence and format.
Mistake 5: Writing for humans only
AI engines extract answer capsules — 40–60 word direct answers at the top of each H2 section. If your pages start H2 sections with "In this section we'll explore..." instead of a direct factual answer, you won't get extracted. Lead every H2 with a 2–3 sentence direct answer, then expand.
GEO Tools vs Traditional SEO Tools
AI search visibility needs different tooling than keyword tracking. Here's how the category stacks up:
| Feature | FindUtils GEO Analyzer (Free) | Stackoptic.com | geodaddy-cli | Ahrefs / SEMrush |
|---|---|---|---|---|
| AI bot access grid | 16 bots | 8 bots | 6 bots | No |
| llms.txt validator | Yes | Yes | No | No |
| Schema stacking check | Yes | Partial | Yes | Partial |
| Citability score | Yes (weighted) | Partial | Score only | No |
| Compare mode | 4 URLs | No | 2+ URLs | Yes |
| AI-ready fix snippets | Yes (paste-ready) | No | Text only | No |
| Copy as Markdown | Yes | No | Yes (JSON/text) | No |
| Mobile-friendly UI | Yes | Yes | CLI only | Yes |
| Price | Free, no signup | Free tier | Free (install) | 449/mo |
| Shareable report URL | Yes | Yes | No | Yes |
The FindUtils GEO Analyzer is the only tool in this comparison that gives you paste-ready code snippets for every failing check plus a one-click "Apply with Claude" button — shortening the cycle from audit to shipped fix from days to minutes.
Minimum Viable GEO Fixes
If you only ship three changes this week, ship these:
- Unblock every major AI crawler in robots.txt (GPTBot, OAI-SearchBot, ClaudeBot, PerplexityBot, Google-Extended minimum). Use the FindUtils Robots.txt Generator and replace your current file.
- Add FAQPage JSON-LD to your top 10 pages. Use the FAQ Schema Generator. Google AI Overviews and ChatGPT both quote FAQ answers verbatim, and this is the lowest-effort schema win.
- Publish /llms.txt with your most important 10–20 pages grouped by section. It's 50 lines of Markdown and moves your citability score immediately.
Everything else in the audit is meaningful but incremental. These three move the needle.
Tools Used in This Guide
- GEO Analyzer — Score any URL for AI search visibility across 25+ signals
- Robots.txt Generator — Create a correct robots.txt that allows all major AI crawlers
- Structured Data Tester — Validate JSON-LD structured data and check rich-result eligibility
- JSON-LD Generator — Build Article, Product, Organization, and other schema blocks from a form
- Schema.org Generator — Browse all schema types with examples and copy-ready JSON
- FAQ Schema Generator — Turn a Q/A list into valid FAQPage JSON-LD
- Meta Tag Generator — Generate title, description, Open Graph, and Twitter card meta tags
- XML Sitemap Generator — Build a valid XML sitemap from a URL list
- Heading Structure Analyzer — Check heading hierarchy for listicle and schema compatibility
- AI Agent Starter Guide — The companion setup guide for making your site fully agent-ready
FAQ
Q1: What is GEO (Generative Engine Optimization)? A: GEO is the practice of optimizing content so generative AI models (ChatGPT, Perplexity, Gemini, Claude, Google AI Overviews) are more likely to cite it when answering user questions. Unlike traditional SEO which targets blue-link rankings, GEO targets inclusion in AI-generated answers. The free FindUtils GEO Analyzer audits 25+ GEO signals in about 3 seconds.
Q2: Is the GEO Analyzer really free? A: Yes. The FindUtils GEO Analyzer is completely free with no signup, no usage limits, and no paywalled features. It runs in your browser, calls a free public API, and supports compare mode (up to 4 URLs) at the same rate. You can export reports as JSON or Markdown without paying anything.
Q3: What's the best free AI search visibility checker in 2026? A: The FindUtils GEO Analyzer offers the most comprehensive free AI SEO audit available. It checks 16 AI bots (versus 6–8 on most competitors), validates llms.txt, scores schema stacking, and includes paste-ready code snippets for every fix — plus a one-click Apply with Claude button. Runner-up free options include stackoptic.com and geodaddy-cli, but both check fewer signals.
Q4: Do I really need to allow AI bots? Won't they steal my content? A: If you want your content cited in AI answers (and the traffic that follows), AI crawlers must be allowed to read your pages. If a bot is blocked, the model cannot quote, link, or recommend you. Each bot can be allowed independently — you can allow GPTBot and OAI-SearchBot for ChatGPT search visibility while blocking training-only crawlers like CCBot if that matches your preference.
Q5: How often should I run a GEO audit? A: At least monthly for pillar pages and quarterly for the full site. AI models update their citation behavior continuously, and your robots.txt or schema can drift with every deploy. The FindUtils GEO Analyzer makes it fast enough to run weekly as part of your release checklist.
Q6: Is llms.txt an official standard? A: llms.txt is an emerging specification, not a W3C standard. It was proposed at llmstxt.org in 2024 and has been adopted by a growing list of sites including Anthropic, Mintlify, and FindUtils. AI engines have started using it informally. It costs 50 lines of Markdown to publish, so the ROI is strongly positive even if the spec evolves.
Q7: How accurate is the citability score? A: The citability score is a weighted composite of the signals AI engines publicly document as ranking factors — crawler access, structured data, content depth, listicle format, and freshness. Treat it as a directional prioritization tool, not a guarantee. A score of 85+ correlates strongly with actual citation rates observed in analyses of ChatGPT search and Perplexity results.
Q8: Does the GEO Analyzer check Core Web Vitals? A: Not in v1. Measuring real browser performance requires a headless Chromium session (~150MB download) which would make the tool slow and expensive. We recommend running Google PageSpeed Insights alongside the GEO audit — the signals don't overlap, and Core Web Vitals remain a separate ranking input for Google AI Overviews.
Next Steps
- Run a full audit on your homepage with the FindUtils GEO Analyzer
- Follow the companion AI Agent Starter Guide to make your site fully agent-ready
- Benchmark against competitors using the compare mode
- Ship the Minimum Viable GEO Fixes this week — three changes, one afternoon, measurable citability improvement