Get the M2M ICP Discovery Worksheet
Know who your website should speak to and what to say.
Schema markup is a labeling system for your website. It's code, invisible to human visitors, that tells machines exactly what they're looking at:
- This is a business, and here's its name.
- This is a price.
- This is a question, and here's its answer.
That's the whole concept. No magic, no secret handshake with Google. And with AI tools now assembling answers from websites they can confidently parse, that labeling layer suddenly matters more than it has in years.
Here's my position up front, because this topic attracts snake oil: I believe schema markup is a translation layer, not a magic trick. It makes a clear message machine-legible. It cannot make a vague message worth citing. Get the order right and schema is a cheap advantage that lasts. Get it backwards and you're labeling empty boxes.
What Schema Markup Actually Is
Think of a nutrition label. The food inside the package is your content. The label translates what's inside into a standardized format anyone (or any machine) can read at a glance: calories, protein, serving size. The label doesn't make the food nutritious. It makes the nutrition legible.
Schema markup does that for web pages. It's a shared vocabulary, maintained at schema.org, that describes what your content is in a format machines parse without guessing. On most modern sites it's added as a small block of JSON-LD code in the page's head. Here's a simplified example for a local business:
{
"@context": "https://schema.org",
"@type": "ProfessionalService",
"name": "CL Creative",
"description": "Webflow web design agency for growth-stage service businesses",
"address": { "@type": "PostalAddress", "addressLocality": "Dallas", "addressRegion": "TX" },
"areaServed": "Dallas-Fort Worth",
"url": "https://www.clcreative.co"
}A human never sees this. A search engine or AI crawler reads it and knows, with certainty instead of inference:
- This is a web design business
- In Dallas
- Serving DFW
No ambiguity about whether "CL Creative" is an agency, a paint brand, or a person.
Why AI Search Raised the Stakes
Traditional search could afford to guess. It showed ten links and let you judge. An AI answer engine can't hedge the same way. When ChatGPT or Google's AI results recommend a business, they're composing a single answer from sources they can parse cleanly and trust.
The data here is worth reading carefully. Ahrefs studied pages cited by AI assistants and found they were roughly 3x more likely to carry JSON-LD structured data than the average page. Sounds like a cheat code.
Then Ahrefs ran the second half of the experiment: they added schema to 1,885 of their own pages and watched what happened to AI citations. Almost nothing. The correlation was real; the causation wasn't.
Both findings point at the same conclusion. Teams that write clear, direct answers also tend to add schema. The schema marks the clarity; it doesn't manufacture it. AI tools cite pages with quotable substance:
Research shows statistics lift AI visibility around 22% and direct quotations around 37%. Schema helps machines find and trust that substance. It can't invent it.
Now, I don't think this means schema is optional. It means schema is step two. Write the direct, answer-first content first, then label it so machines never have to guess what it is. Skipping step two leaves easy visibility on the table.
Skipping step one makes step two pointless.
The Schema Types That Matter for a Service Business
You don't need the whole schema.org vocabulary. For most service businesses, four types do nearly all the work:
- Organization / LocalBusiness. Who you are, where you are, what category you're in, how to contact you. This is your business's ID card, and it's the fix for AI tools misdescribing what you do.
- FAQPage. Marks up question-and-answer content so machines can lift each Q&A cleanly. Pair it with real FAQ sections on service and pricing pages, since that's the most quotable format for getting cited by AI tools . One note here: Google retired the expandable FAQ dropdowns in search results in May 2026, after restricting them to government and health sites back in 2023. That killed the display feature, not the markup. AI tools still parse FAQPage schema to pull clean answers, and that's the reason to use it now.
- Service. What you offer, for whom, and in what area. It connects your offerings to the "who does X in Y" questions buyers ask.
- Article. For blog posts: author, date, publisher. It builds the machine-readable record of your expertise over time.
On Webflow, you can add all of these as a code embed or in per-page custom code, no plugins required. Validate anything you add at validator.schema.org, and check how Google reads it in Search Console.
How to Get Started This Month
A sequence, in order:
- Fix the message first. If your homepage can't tell a stranger what you do and who it's for in one sentence, schema will just certify the confusion.
- Add LocalBusiness or Organization schema sitewide. One afternoon of work, permanent clarity about who you are.
- Add FAQPage schema wherever you have real Q&A content, and add that content where you don't.
- Add Article schema to your blog template so every post you publish is labeled automatically.
- Verify from the outside: run a quick AI search self-audit and see whether ChatGPT and Perplexity describe your business accurately.
That's a few hours of total effort for a durable edge, in a market where most of your competitors haven't labeled anything.
Schema is the label. Your message is the meal. Most websites are missing one of the two. And almost everyone shopping for schema help is missing the meal, not the label. Which is your site missing?
Your Marketing Isn't Bringing in Leads and You Can't Figure Out Why
Score your marketing across 7 key areas in 5 minutes and find out exactly what's holding back your leads and what to fix first.
Questions You Might Be Asking
No. Ahrefs tested exactly this by adding schema to 1,885 pages, and AI citations barely moved. Schema removes ambiguity about content that's already clear and quotable; it doesn't create citation-worthy content. Write direct answers first, then add schema so machines can parse them with confidence.
Start with LocalBusiness (or Organization) schema sitewide, then add FAQPage schema on pages with real question-and-answer content. After that, add Service schema on your offer pages and Article schema on blog posts. Those four cover nearly everything a service business needs without touching the more exotic types.
Add JSON-LD in your site-wide custom code settings for Organization/LocalBusiness schema, and use per-page custom code or embed blocks for FAQPage and Article schema. No plugins needed. Validate everything at validator.schema.org before publishing, and monitor Search Console for structured data errors.
Yes, with one caveat. Google retired FAQ dropdowns in May 2026, so FAQPage markup no longer changes how your listing looks in Google. Schema still powers rich results like review stars and business panels, which improve click-through even when rankings stay the same. The AI search benefit stacks on top of the classic one, which is part of why SEO and AEO work best as one effort.
The basic types are genuinely DIY-able if you're comfortable pasting validated code, and free generators exist for most of them. Hire help when you have many pages, custom content types, or a bigger gap upstream. When your content doesn't answer buyer questions yet, that's the work worth paying for, not the markup.
