--- title: "Free AI Readiness Audit: See Your Site Like ChatGPT" description: "Free AI readiness audit for B2B websites: what ChatGPT, Claude and Perplexity can read, which AI crawlers you block, and the three fixes that matter." canonical: "https://webrevolution.si/audit/" updated: "2026-10-06" words: 1278 jsKb: 8.2 --- Free · about 60 seconds · no email for the score # Free AI Readiness Audit: See Your Website the Way ChatGPT Sees It I read your site the way GPTBot, ClaudeBot and PerplexityBot do: raw HTML, no JavaScript. 25 checks give you a Super Intelligence (SI) Readiness Score: whether AI can find you, cite you and send you clients, and what to fix first. **The live audit needs JavaScript**, because it streams results while it runs. Without it, the 10-point checklist below covers the same ground by hand. *Form · GET /audit/ · fields: Your website; Company name; Main service and market; The question a customer would ask an AI; One competitor’s website · button: “Run the free audit”* No email needed for the score. One site per scan. A re-scan within 24 hours shows the cached result, and your report link stays live for 12 months. **Add context optional: I read it before we talk; a competitor gets you a side-by-side** Compared side by side in the full report. The audit identifies itself as [WebRevolutionAudit/1.0](https://webrevolution.si/audit/bot/) and reads only public pages, the way an AI crawler would. ## agent log · audit 1. human://you Enter a URL. Nothing runs until you do. 2. researcher reads robots.txt, your homepage HTML, sitemap, llms.txt and up to 5 more pages 3. tester knocks with 8 AI crawler user-agents and notes what your server answers 4. architect counts the words a crawler gets without JavaScript 5. fact-checker looks for entity data, authors, dates, prices and proof 6. watcher times the server and weighs the page 7. score Found · Cited · Chosen, out of 100. Deterministic: same site, same score. What the score measures ## Found. Cited. Chosen. 25 checks, 100 points. Plus agent readiness. Every check below runs on every scan, with the same rules for every site, and adds up to your SI-Readiness Score. No AI model grades your site: same site, same score. ### Found 40 pts Can AI crawlers reach and read you? - AI search crawlers allowed in robots.txt 8 - AI training crawlers in robots.txt 4 - Server response to AI user-agents 8 - Words readable without JavaScript 7 - Content in the HTML, not built by JavaScript 4 - Clear headings in the HTML 3 - XML sitemap with dates 3 - Canonical, noindex and language 3 ### Cited 35 pts Do assistants trust you enough to quote you? - Organization structured data 6 - Consistent name and sameAs profiles 4 - Named authors and visible dates 5 - Answer-first content 6 - Visible pricing 4 - Numbers that prove results 4 - Recently updated 4 - llms.txt (low impact) 2 ### Chosen 25 pts Do the people AI sends become clients? - Server response time 2 - HTML size 2 - JavaScript weight 2 - Real-visitor speed (Chrome UX Report) 5 - Contact or booking path in the HTML 2 - A clear next step near the top 2 - Forms people and agents can fill in 4 - Consent banner and third-party scripts 3 - Actions an agent can take 3 ### Agent-ready 10 pts Can an AI agent act on your site? Scored separately. - WebMCP tool attributes 2 - MCP or agent-card discovery 2 - Machine-readable offers and prices 2 - Markdown versions and llms.txt 2 - Form semantics 1 - API description (OpenAPI / api-catalog) 1 0–39**Invisible to AI** 40–59**Readable** 60–79**Citable** 80–100**SI-ready** Three failures make everything else irrelevant, so they cap the score at 39: a homepage with almost no readable text, a homepage marked noindex, or a robots.txt that keeps every AI search crawler out. Do it yourself ## The 10-point checklist Run it by hand in an afternoon. No tool needed. These are the checks that move the score most, in the order I’d do them. 1. **Open yoursite.com/robots.txt** Look for “Disallow: /” under GPTBot, OAI-SearchBot, ClaudeBot, Claude-SearchBot, PerplexityBot or “User-agent: *”. A blanket block keeps AI search out. 2. **View the source of your homepage** Press Ctrl+U (Cmd+Option+U on a Mac). Can you read your headline and services in the HTML? An empty
means AI crawlers see nothing. 3. **Check your CDN’s bot settings** On Cloudflare: Security → Bots. Keep AI search crawlers on Allow, switch off “Managed robots.txt” unless you meant it, and check WAF rules that match user-agents. 4. **Open yoursite.com/sitemap.xml** It should list every page you want found, each with a real date, and robots.txt should point to it. 5. **One H1, question-shaped H2s** Each page needs one H1 that says what you do and for whom, then H2s phrased the way buyers ask, each answered in the first two sentences. 6. **Search the source for “application/ld+json”** You want an Organization block with your exact name, logo, URL and sameAs links to your LinkedIn and other official profiles. 7. **Put prices on the page** At least “from” prices or typical ranges, as text. Buyers ask AI what things cost; a number gets you into that answer. 8. **Name the author, show the date** Articles and service pages need a real person’s name and a visible “Updated” date. Assistants favour content they can attribute and date. 9. **A plain contact link in the first screen** A real link to contact or booking near the top, plus tel: or mailto: if you take calls or emails. Not only a JavaScript widget. 10. **Labelled forms with autocomplete** Every field has a visible label, the right type (email, tel) and autocomplete tokens (name, email, organization). People finish them faster and agents can fill them. Questions ## Before you run it **Is this the same as an SEO audit?** No. An SEO audit asks whether Google can rank you. This audit asks whether AI assistants can *reach* your site, *read* it without running JavaScript, *trust* it enough to quote it, and whether the people they send can act. Some checks overlap with SEO; the questions behind them are different. **Why does my site score low when Google ranks it fine?** Google runs JavaScript; GPTBot, ClaudeBot and PerplexityBot don’t. A site built entirely in the browser can rank on Google and still look blank to AI crawlers. CDN bot settings are the other usual reason: some providers, Cloudflare among them, can block AI crawlers depending on the zone’s age and settings. **Does the audit ask ChatGPT about my business?** No. The free score is deterministic: no AI model grades your site, so the same site gets the same score. It checks whether assistants *can* find, read and quote you. Running real buyer prompts through ChatGPT, Perplexity, Gemini and Claude is part of the [Deep Audit](https://webrevolution.si/services/ai-visibility-audit/). **My server refused the crawler probes. Is that bad?** It is worth checking. The probe sends a normal request with each crawler’s user-agent from my audit server. Real crawlers come from verified IP ranges and your CDN may let them through, but many firewall rules block by user-agent alone. Check your bot settings, then run the audit again. **Why is llms.txt worth only 2 points?** Because its impact is small today: Google ignores it and few assistants fetch it. It takes twenty minutes and costs nothing, so it is worth doing, just not worth more than that. **What do you store, and can I block the audit bot?** Results are cached for 24 hours, so a repeat scan within a day reuses them. The report link stays live for 12 months, so I keep the address you scanned and the result for that long. Your email is stored only if you ask for the full report. To keep the audit bot off your site, add `User-agent: WebRevolutionAudit` and `Disallow: /` to your robots.txt. [More about the bot](https://webrevolution.si/audit/bot/). Rather have me look at it? 20 minutes with me, not a sales rep. Bring your report; you leave with a clear next step, whether or not we work together. [Book 20 minutes](https://webrevolution.si/contact/#book) [Write to me instead](https://webrevolution.si/start/?goal=audit) [WhatsApp](https://wa.me/528123956065?text=Hi%20Vegard%2C%20I%27m%20on%20webrevolution.si%2Faudit%2F%20%E2%80%94%20) --- ## Structured facts The page's JSON-LD (schema.org), flattened to text. ### WebSite: WebRevolution.si - url: https://webrevolution.si/ - inLanguage: en - publisher: WebRevolution.si ### OnlineBusiness: WebRevolution.si - alternateName: WebRevolution; webrevolution.si; WebRevolution SI - url: https://webrevolution.si/ - logo: https://webrevolution.si/icon-512.png - slogan: One man. An army of agents. - description: One-man AI agency for B2B: AI-ready websites, AI content & distribution engines, AI agents and custom tools that AI assistants can find, cite and recommend. - founder: Vegard Wikeby - areaServed: Worldwide - knowsAbout: Generative engine optimization; AI-ready websites; AI agents for business; AI content engines; Cloudflare Workers; Conversion rate optimization; Model Context Protocol - contactPoint: ContactPoint (contactType: sales; url: https://webrevolution.si/contact/; availableLanguage: en) - sameAs: https://www.linkedin.com/in/vegardwikeby; https://x.com/vegardwikeby ### Person: Vegard Wikeby - url: https://webrevolution.si/about/ - jobTitle: Founder, WebRevolution.si - worksFor: WebRevolution.si - image: https://webrevolution.si/img/vegard-wikeby.jpg - sameAs: https://www.linkedin.com/in/vegardwikeby; https://x.com/vegardwikeby - knowsAbout: Web development; AI agents; Generative engine optimization; WordPress; E-commerce; Cloudflare ### BreadcrumbList - Home › Free AI audit ### WebApplication: Free AI Readiness Audit - alternateName: SI-Readiness Score - url: https://webrevolution.si/audit/ - description: Free AI readiness audit for B2B websites: what ChatGPT, Claude and Perplexity can read, which AI crawlers you block, and the three fixes that matter. - applicationCategory: BusinessApplication - operatingSystem: Any - browserRequirements: Requires JavaScript - isAccessibleForFree: true - offers: Offer (price: 0; priceCurrency: EUR) - provider: WebRevolution.si - creator: Vegard Wikeby - featureList: robots.txt rules for 10 AI crawlers; Server response to AI crawler user-agents; Words readable without JavaScript and JavaScript-shell detection; Organization data, authorship, freshness, pricing and proof signals; Speed, contact path and form checks; Separate agent-readiness score