Schema Markup for AI Search: How ChatGPT and Google AI Overviews Actually Read Your Website

If you’ve read our overview of AI SEO for Auckland small businesses, you already know AI-driven search is real and already affecting who gets found. This post goes one level deeper into the part almost nobody explains clearly: schema markup — the actual mechanism that helps AI systems understand what your business is, without guessing.

What Schema Markup Actually Is

Schema markup (also called structured data) is a standardised code format — usually JSON-LD — added to your website’s pages that spells out facts about your business in a way machines can read directly, instead of inferring them from paragraphs of text.

Without schema, an AI system reading your homepage has to guess: is this a local business or a national one? Is that phone number current? Are those reviews real and how many are there? With schema, you tell it directly, in a structured format there’s no ambiguity about.

Here’s a simplified real example — the kind of thing that sits invisibly in a page’s code:

{
  "@type": "LocalBusiness",
  "name": "Kiwi Web Design",
  "telephone": "+64-21-039-6580",
  "address": {
    "@type": "PostalAddress",
    "addressLocality": "Grey Lynn",
    "addressRegion": "Auckland",
    "addressCountry": "NZ"
  },
  "aggregateRating": {
    "@type": "AggregateRating",
    "ratingValue": "5.0"
  }
}

That block does more for machine trust in five lines than a paragraph of marketing copy saying “we’re a trusted local Auckland agency.”

The Schema Types That Actually Matter for AEO

Not every schema type is worth your time. For a small Auckland business, these are the ones that carry real weight:

Schema Type	What It Tells AI Systems	Where to Use It
`Organization` / `LocalBusiness`	Who you are, where you’re based, how to contact you	Homepage, every key page
`ProfessionalService`	Confirms you’re a service business, not a product retailer	Homepage, service pages
`Review` / `AggregateRating`	Real, structured proof of reputation — not just a number in a headline	Homepage, testimonials page
`FAQPage`	Direct question-and-answer pairs AI systems can lift verbatim	Service pages with genuine FAQs
`Article`	Author, publish date, and topic for blog content — supports E-E-A-T	Every blog post
`BreadcrumbList`	How your pages relate to each other and your site structure	Every page

The mistake we see most often isn’t missing schema entirely — it’s schema that doesn’t match what’s actually on the page. If your Review schema claims a rating your visible page doesn’t back up, or your address in schema doesn’t match your Google Business Profile, that inconsistency undermines trust rather than building it. Structured data has to be accurate, not just present.

How This Plays Out Differently Across AI Platforms

Not every AI system reads the web the same way. A few practical differences worth knowing:

Google AI Overviews and AI Mode draw heavily on Google’s existing index and crawl data — meaning your traditional SEO fundamentals (indexability, Core Web Vitals, clean crawlability) still gate whether you’re eligible to be included at all. Schema helps once you’re in the running.
ChatGPT’s web search and Perplexity tend to favour clearly structured, directly-answering content and can cite sources more explicitly by name — which is where FAQPage schema and direct, plainly-worded answers near the top of a page earn their keep.
Microsoft Copilot (powered by Bing) leans on Bing’s own index, which is smaller than Google’s — meaning consistent basic signals (verified Bing Webmaster Tools listing, clean NAP data) matter more than they might for Google-only strategies.

The practical takeaway: no single platform-specific trick covers all of them. The fundamentals — accurate schema, clean crawlability, consistent business information — are what earns visibility across all of them at once.

A Simple Way to Check Your Own Site

You don’t need to be technical to spot the gaps. Pull up your homepage and ask:

If I strip away the design, does the raw content clearly state what this business does and where it’s based?
Do the FAQs (if you have any) answer real questions directly, in plain language, near the top?
Does every number you display — reviews, years in business, client count — match what’s actually verifiable elsewhere?

If any of those feel shaky, that’s the gap schema markup and content structure are meant to close.

Where This Fits With Everything Else

Schema markup isn’t a silver bullet on its own — it’s one piece that works alongside the broader trust picture we cover in E-E-A-T for AI search and the AI SEO service overview. Structured data tells machines what you are; E-E-A-T and consistent signals tell them whether to trust it. You need both.

If you’re not sure what’s currently implemented on your own site — or whether it’s accurate — that’s a quick, concrete thing we can check as part of technical SEO for Auckland businesses.

← Back to all articles