Built for the AI Web
A repeatable layer that makes small-business sites legible to AI search and agents—structured data, an llms.txt brief, and explicit AI-crawler rules—so ChatGPT, Claude, and Perplexity can actually find and cite them.
The Challenge
AI assistants are fast becoming how people find local businesses—but most small-business sites are invisible to them. Classic SEO does not translate: LLMs want explicit, structured, machine-readable context, and most sites give them none.
The Approach
I built a portable "generative engine optimization" stack and rolled it across a portfolio of sites. Rich Schema.org JSON-LD tuned per industry (LegalService, MedicalClinic, Physician, RealEstateAgent, FAQPage), a dynamic llms.txt route that serves a clean markdown brief of who the business is and what it offers, and explicit robots rules that welcome AI crawlers—GPTBot, ClaudeBot, PerplexityBot, Google-Extended and more. The whole layer drops into any Next.js or Astro site in an afternoon. This very site runs it too.
The Outcome
Client sites now carry the structured context AI search engines need to surface and cite them accurately—the difference between being a string of HTML and being a business an AI can recommend. One pattern, applied consistently across legal, healthcare, and real-estate clients.
Key Details
- Schema.org JSON-LD tuned per industry (LegalService, MedicalClinic, Physician, …)
- FAQPage structured data answering real customer questions
- Dynamic llms.txt brief: facts, services, and credentials in machine-readable markdown
- Explicit AI-crawler allowlisting in robots (GPTBot, ClaudeBot, PerplexityBot, …)
- Chatbase conversational assistant where it fit the client
- Repeatable across Next.js and Astro; deployed on this site and client sites