How to make your website AI-readable
Before any AI engine can cite your business, it has to be able to read your website. This is the technical checklist for transforming a typical small business website into one AI crawlers can extract, parse, and quote in milliseconds.
Most small business websites are invisible to AI search engines. The fix is engineering, not content. This guide is the four-layer technical checklist that takes a typical website from invisible to AI-readable.
Layer 1: rendering architecture
AI crawlers do not execute JavaScript during the retrieval step. Your site must render its primary content in the initial HTML response. Acceptable architectures: Astro, Next.js with SSG or SSR, Eleventy, plain HTML, server-side rendered WordPress with caching, Webflow static export. Unacceptable: Wix without the SSR option enabled, Squarespace default templates, client-only React/Vue/Angular sites.
Test with curl https://your-site.com. If you can read your services, pricing, and providers as plain text in the response, layer 1 passes.
Layer 2: Schema.org structured data
Ship a complete Schema.org @graph as a single JSON-LD block in the page head. For a specialty practice, the minimum graph includes:
- Primary business entity (MedicalClinic, LegalService, HomeAndConstructionBusiness) with @id, name, address, phone, hours.
- Each named provider as a Person entity with hasCredential, worksFor pointing to the business.
- Each Service offered with @id, name, description, provider linked to the business.
- For medical: Drug or MedicalProcedure entities per treatment, with mechanismOfAction and indication.
- Offer entities for pricing.
- FAQPage with Question/Answer entities for top buying-intent questions.
- BreadcrumbList for navigation.
- Speakable specifications for voice-driven AI.
Layer 3: answer paragraphs
Every page that targets a query intent should open with a 40-60 word answer paragraph in this format: fact-dense, named entities, specific facts (price, location, treatment name, eligibility criteria), no qualifiers or marketing language. This is what KailxLabs calls the Answer Capsule. Place it immediately after the H1 of every page.
Layer 4: crawler permissions and manifests
Explicit robots.txt entries for AI crawlers:
User-agent: GPTBot
Allow: /
User-agent: OAI-SearchBot
Allow: /
User-agent: ChatGPT-User
Allow: /
User-agent: ClaudeBot
Allow: /
User-agent: anthropic-ai
Allow: /
User-agent: PerplexityBot
Allow: /
User-agent: Google-Extended
Allow: /
User-agent: Bingbot
Allow: /
User-agent: Applebot-Extended
Allow: /Plus llms.txt at the domain root: a markdown summary of the business under 3,000 tokens listing services, providers, pricing, and key pages. ai.txt at the domain root: a machine-readable AI usage policy.
Verification
The free 48 hour AI visibility audit runs all four layers as part of the diagnostic and identifies which are passing and which are failing on your current site.