Twenty-eight points across the work that gets local businesses cited by AI Overviews, ChatGPT, Perplexity, and Claude. Schema fundamentals, content structure for machine parsing, E-E-A-T signals, entity establishment, and the conversational-phrasing patterns that show up disproportionately in AI answers.
AI Overviews preferentially surface FAQPage-marked Q&A pairs. Use FAQPage with Question + Answer children on dedicated FAQ pages and on service pages with substantive Q&A blocks. Don't fake it on pages that don't actually answer questions — that triggers manual actions.
AI engines parse headline, datePublished, author, and articleBody to decide which pages to cite. Without article schema, your content competes blind.
Author Person entity at /about#person. Link via sameAs to LinkedIn, Wikipedia (if applicable), GitHub, professional licenses. Every BlogPosting references this Person via @id.
Tutorials with explicit numbered steps use HowTo with step array. AI engines extract these as procedural answers — high citation rate for 'how to X' queries.
knowsAbout array lists the specific topics you have authority on. Helps AI answer engines map your business to relevant queries beyond your primary category.
AI engines extract the first definitive answer they find. Bury the answer beneath an introduction and you lose to the competitor who answered immediately. Open with the answer, then explain.
How long does local SEO take? as an H2 outperforms Timeline expectations. Match the phrasing real people use in queries — that's what AI engines match against.
Use <dl> for definitions, semantic <ul>/<ol> for lists, real <table> for comparisons. AI engines extract these structured elements at higher rates than prose.
ChatGPT and Claude were trained on conversational text. Prose that reads like written advice ("if you're a local business, you probably...") gets cited more often than corporate-marketing voice.
"30-90 days for map pack movement" is citable. "Quickly" or "in a reasonable time" gets paraphrased away. AI prefers content with extractable specific claims.
Use distinct sections with parallel structure. Pros heading + bullet list, then Cons heading + bullet list. AI engines preferentially cite this format for comparison queries.
Visible author name + photo + link to /about. The Person entity in schema needs a visible counterpart on the page — both signals reinforce each other for AI engines deciding who wrote this.
AI engines weigh demonstrable expertise. "15 years in local SEO, served 100+ businesses" beats "experienced consultant". Specific past employers, certifications, and case histories increase citation rate.
Link to original research, primary sources, and Google's own documentation. AI engines verify claims by following links — content that grounds itself in citable sources gets cited more.
"In my last 50 audits", "I've watched", "the engagement we ran for X showed". First-person experience is the E in E-E-A-T. AI engines specifically look for this — they cite firsthand accounts more than aggregated content.
Content that hedges where appropriate ("this won't work for X situation", "results vary by competitive density") signals epistemic honesty. AI engines have learned to distrust uniformly confident content.
AI engines map topical authority via internal-link density. A site with 30 articles on local SEO outranks one with 3 well-written ones. Build clusters around your service pillars; interlink generously.
Link with anchors like how to optimize Google Business Profile categories, not read more. Anchor text is a major signal AI engines use to map page topics.
Pillar = comprehensive 3000+ word article on a primary topic. Links to all supporting articles. Each supporting article links back to the pillar. The hub-and-spoke structure is what AI engines parse as 'this site is an authority on X'.
AI engines deprioritize stale content. A cluster that hasn't been updated in 18 months reads as abandoned. Even small updates (refreshed examples, new sections, updated dates after real revisions) signal active authority.
Wikipedia is the gold-standard entity reference. For local businesses, the next-best signals are Crunchbase profiles, professional licensing registries, BBB pages, and local Chamber listings. AI engines cross-reference these to verify the brand is real.
Wikidata is the structured-data backbone of the open knowledge graph. Free to add a basic entry for any verifiable business. AI engines use Wikidata as a primary entity-resolution source.
Industry publications, podcast appearances, conference speaker bios, guest articles. AI engines pattern-match brand co-occurrence across sites — a brand that's only on its own site reads as low-trust.
AI engines particularly weigh person-entity signals for service businesses. The principal's LinkedIn (with work history), published articles under their byline, conference talks, and professional memberships all reinforce the brand entity.
Run your top 20 query targets through Google search (with AI Overviews enabled) and ChatGPT/Claude/Perplexity. Note which sources get cited. Track your citation rate over time.
Track every public mention of your brand across the web. AI engines aggregate mention frequency as a trust signal. Catch and respond to mentions; nudge incomplete mentions to include your URL.
Emerging convention for declaring AI-crawl preferences and surfacing key pages to LLM training pipelines. Not yet universal but cheap to deploy. See llmstxt.org for the spec.
If a page gets cited frequently, study what makes it citation-friendly and apply the pattern elsewhere. If a comprehensive page gets ignored while a thinner competitor gets cited, the issue is structure, not depth.
If you want someone to audit your AI-search readiness across schema, content structure, entity presence, and citation patterns specific to your industry, the $197 audit includes AI-search analysis as one category.