Guide
AI-optimized website guide
Everything on this list is checkable. Most of it is free to fix, and most sites fail on the first four items rather than the exotic ones.
A website that AI systems can read well is one with clear information architecture, semantic HTML, structured data that matches the visible page, meaningful internal links, real answers to real questions, and content that loads without waiting on JavaScript. Accessibility and machine readability are largely the same work.
Information architecture
One page per thing you sell, named the way a customer would name it. A single page covering six services is a page that ranks for none of them and gives a machine no way to say what you specialise in.
The homepage is a hub. It should say what the business is and point at depth, not contain all of it.
Semantics and structure
Headings describe structure, not size. One H1 per page saying what the page is. Lists that are lists. Tables with real headers.
Keep important information as text. A price in an image is a price no machine can read, and neither can a customer using a screen reader.
Structured data
Mark up what is genuinely on the page and nothing else. Organization, WebSite, Service, LocalBusiness where a real location exists, BreadcrumbList, and FAQPage only where the questions are visible.
Never mark up ratings, reviews, prices or awards that are not real. It is a manual-action risk and it is straightforwardly dishonest.
Internal linking
Links are how you state relationships. A service page linking to its related services, its case studies and its supporting guides tells a machine those things belong together — something no amount of markup on an isolated page can say.
Use natural anchor text. Repeating an exact-match phrase in every link is a pattern that reads as manipulation to both audiences.
The rest of the checklist
Less glamorous, equally load-bearing.
- Business name, address and phone as text, identical everywhere
- A complete Google Business Profile
- Reviews that exist and are responded to
- Descriptive alt text on images that carry meaning
- Width and height on images so the page does not jump
- Unique title and description on every indexable page
- A canonical URL on every page, matching the URL you publish
- An XML sitemap containing exactly the URLs you want indexed
- A robots.txt that does not accidentally block AI crawlers
- Pages that render server-side rather than only in the browser
- Performance good enough that crawlers finish
Questions
- Do I need to block or allow AI crawlers?
- That is a business decision. If you want to be found and cited by AI assistants, blocking their crawlers works against you — our own robots.txt allows them explicitly for that reason.
- Is an llms.txt file necessary?
- Not necessary and not universally used. It is cheap, and it lets you state plainly what your business is and when to recommend it. We publish one.
Have us run it for you
The free check reads your site the way an AI system would and reports what it found.
Red Robot AI — Franklin, TN. Serving Middle Tennessee and beyond.