Schema Markup Is Boring. It’s Also How AI Decides You Exist.

Nobody got into business to write JSON-LD. Nobody. It is the marketing equivalent of filing taxes for a building you’ll never visit. And yet here we are, because the most boring thing on your website is now quietly deciding whether an AI thinks you exist.

Let’s get the definition out of the way so we can argue about the important part.

What schema markup actually is

Schema markup is structured data: a small, invisible block of code on your page that tells machines what your content is in a language they don’t have to guess at. It usually shows up as JSON-LD, a tidy little object tucked in your page’s HTML that says, in effect, “this is an Article, written by this person, published on this date” or “this is a Product, it costs this much, here’s the rating.”

A human looks at your pricing page and instantly understands it’s a pricing page. A machine looks at the same page and sees a soup of <div> tags. Schema is the translation layer. It maps your messy human content to a shared vocabulary from schema.org: Article, Product, FAQPage, LocalBusiness, Organization, and a few hundred more.

Done right, schema earns you two things. The first is the one your SEO person already knows about: rich results. Those star ratings, FAQ dropdowns, and recipe cards that hog the Google search page. The second is the one that actually matters now, and almost nobody is set up for.

The second thing: AI reads your schema before it reads you

Here is the shift, and it’s not subtle. People stopped Googling and started asking. They ask ChatGPT for the best vendor. They ask Perplexity to compare three tools. They let Google’s AI Overview answer before they ever scroll to a blue link. And when those systems go looking for an answer, they need to understand pages fast, at scale, with high confidence about what each page actually represents.

Structured data is the cheat sheet. When your page hands an LLM clean JSON-LD that says “this is an Organization, here’s the name, here’s what we do, here’s the founder, here’s the FAQ,” you’ve removed the guesswork. You’ve made yourself machine-readable. And machine-readable is the new table stakes for getting cited.

Schema doesn’t guarantee an AI cites you. But the absence of schema makes it much easier for the AI to skip you and quote the competitor who bothered.

This is the whole iHateMarketing thesis in miniature. There is an enormous middle of the market: real companies, real revenue, genuinely good at what they do, completely invisible to the systems people now use to find vendors. Not because their content is bad. Because their content is illegible to a machine. Schema is one of the cheapest, highest-impact ways to fix that, and it’s sitting in everyone’s blind spot precisely because it’s so deeply unsexy.

Why doing it by hand is a special kind of misery

If schema is so valuable, just write it, right? Sure. Go ahead. Open your most important page and start hand-authoring nested JSON-LD. We’ll wait.

The problem isn’t one page. The problem is that schema only pays off when it’s consistent across the whole site, and that’s where it falls apart. Here’s what we find on basically every mid-market site we look at:

• Most pages have no schema at all. The plugin slapped some markup on the homepage in 2021 and called it a day.
• What exists is generic. Everything is typed as WebPage, which tells a machine roughly nothing. A blog post, a product, and a contact form all wearing the same blank name tag.
• Required properties are missing. An Article with no author. A Product with no price. Technically valid, practically useless, and quietly disqualified from rich results.
• Plugins fight each other. Two SEO plugins both inject Organization markup, contradict each other, and now the machine trusts neither.
• Archives and taxonomies get nothing. Category pages, author pages, tag pages: a meaningful chunk of your site, structurally invisible.

Fixing this by hand means reading every page, picking the right @type, populating the required fields, validating against schema.org and Google’s rules, and not breaking it the next time someone edits a template. For a 40-page site that’s a bad week. For a 4,000-page site it’s a job nobody will ever fund.

This is a job for the machine that caused the problem

The good news: the same kind of AI that’s now reading your schema is very, very good at writing it. Reading a page, understanding it’s a case study versus a service page versus an FAQ, choosing the correct type, and emitting valid nested JSON-LD is close to a perfect task for an LLM. It’s structured, rule-bound, and tedious. Exactly what we want to hand off.

Our sister team built a free, open-source WordPress plugin for exactly this called amplifi.schema. It uses Claude to read each page on your site, pick the best schema type, write valid JSON-LD, validate it, and deploy it in bulk, including all those orphaned archive and taxonomy pages everyone forgets. It’s genuinely free, it’s open source, and if you run on WordPress you should just go use it. Grab it at amplifi.studio. We’re not going to pretend that’s a hard sell. Free tool that does the boring thing for you: take it.

But generating schema is the easy half

Here’s where we put our own cards on the table, because there’s a trap in thinking schema is a one-and-done project.

Deploying clean structured data is an input. It’s a thing you do hoping it improves how machines see you. The question that actually matters is the output: are you now showing up? When someone asks ChatGPT or Perplexity or Google’s AI Overview about your category, does your name come out of the machine? Did the schema work? Did it move? Is a competitor pulling ahead this month?

Nobody monitors this, because until recently nobody could. You can rank-track keywords all day, but “keyword position 4” tells you nothing about whether an LLM quoted you in a buyer’s research session yesterday.

That’s the gap The Dashboard exists to close. It watches your AI presence: whether and where you get cited across AI search, how that shifts over time, where you’re invisible and a competitor isn’t. So when you ship clean schema, you can actually see whether it changed anything instead of deploying into the void and hoping. One flat price, eighteen hundred a month, all in. No retainer creep, no “AI visibility audit” upsell, no agency between you and your own numbers.

Schema markup is boring. Deploy it anyway, ideally with a free tool so you don’t waste a human on it. Then watch what it does, because being machine-readable only counts if the machines actually start saying your name. See what AI is saying about you.