The Machine-Readable Method™
AI assistants don't browse your website the way a customer does. They retrieve it, extract its text, and cross-reference it against your profiles, your reviews, and the rest of the web. They compose answers from what they can read and confirm. The Machine-Readable Method is our engineering discipline for showing up well in that process.
Every recommendation an AI gives is built from evidence. If the evidence about your business is missing, unclear, or contradictory, you start at a disadvantage, however good your work is. The method treats your business's presence on the web as a data system, and fixes it the way engineers fix data systems: measure, correct, verify.
AIRS measures the five dimensions that shape whether an AI system can recommend you. Each dimension is scored red, amber, or green, and each maps to concrete, checkable work. AI answers vary between runs, so we repeat every prompt several times and score the pattern.
01 / 05
Can machines reach you?
AI crawlers must be able to access and index your content. We check robots.txt rules against every major AI bot, verify your content is server-rendered rather than locked behind JavaScript, and confirm your sitemap and canonical structure are clean.
Failing looks like: an AI bot blocked by a default robots rule, or a site whose text only exists after scripts run — invisible to most machine readers.
02 / 05
Are the facts right?
We verify that what machines read — your hours, services, address, phone, service area — matches reality on every surface: your site, your Google Business Profile, directories, and social profiles.
Failing looks like: an AI confidently telling a customer you're closed on Saturdays because one stale directory says so.
03 / 05
Is anything missing?
Everything that wins you a job must exist in machine-readable form: full service list, service area, credentials and licensing, years in business, reviews. If it's only in your head — or only in a photo — machines can't use it.
Failing looks like: you do emergency repairs, but no machine-readable source says so — so AI never mentions you for the most urgent, highest-value calls.
04 / 05
Can machines quote you?
Content must be easy to extract: one clear H1, logical headings, a direct answer under every question, all in visible text. Structured data should mirror that content, never replace it. Machines quote what's unambiguous.
Failing looks like: your services buried in marketing prose an AI can't safely paraphrase, so it quotes a competitor's cleaner page instead.
05 / 05
Does the web corroborate you?
AI systems trust what multiple sources confirm. We reconcile your profiles (Google, Bing Places, Yelp, BBB, trade directories), link them with sameAs, build your review footprint, and make your identity resolve as one consistent entity across the web.
Failing looks like: three slightly different business names across the web, so the AI treats you as three weak entities instead of one strong one.
No magic, no tricks that break next quarter. The setup work is concrete and inspectable:
We build on open web standards — semantic HTML, Schema.org, sitemaps — and on facts that stay true whichever platform reads them. Platforms change their interfaces; clear, consistent evidence about your business doesn't expire. That's why this is a setup you own, not a subscription you rent. No method can guarantee an AI answer, so we measure instead of promising.
See your own score first.
Run my free AI-readiness check