
How to Add Schema to Your Pages (So Google and AI Can Recommend You)
I want to be upfront about something before I say a single word about search: I am not an SEO expert, and I'm never going to pretend to be.
But I know it matters. And I know that if I keep waiting until I feel qualified, nothing on my site ever actually gets fixed.
A few weeks ago I wrote about the two small files that tell AI search what your site is for. This is the companion piece, and it covers more ground.
Because llms.txt only talks to AI crawlers. Schema talks to everyone. It's the standard Google and Bing have been reading for over a decade, and now the AI tools read it too.
Here's what's actually happening when something lands on one of your pages:
- A sales page and a blog post look almost identical underneath
- Nothing declares your price, your author, or your publish date
- Your FAQ section is just text sitting on a page, not questions and answers
- So the crawler makes its best guess, and its best guess becomes your search result
Schema is not an SEO tactic. It's a labeling job, and it's the one piece of SEO that has right answers.
Why I built schema into every page
I'd already fixed the AI side of this. Every domain I run has its own llms.txt file explaining what that property is for.
So I thought I was done. I wasn't.
Because those files describe the site. Nothing on the individual pages said what each page actually was. My paid offers, my free resources, and my blog posts were all just words in a box as far as anything reading them was concerned.
Schema is the fix for that. It's a small block of structured data sitting on the page that says plainly: this is an FAQ page, this is a product, here's the price, here's who wrote it.
And then I hit the part that cost me the most time. I added schema to a set of pages, felt very accomplished about it, and it was not there. My platform had stripped it out on the way in without a single error message.
Which taught me the actual lesson of this whole shortcut: adding schema and having schema are two different things.
If a crawler has to guess what your page is, it will. And its guess becomes your search result.
Here's the part I didn't expect to enjoy. This is the rare SEO task that AI is genuinely good at, because it isn't creative writing. It's structured data with correct answers. There's a right type for your page and a right shape for the markup, and a machine is better at that than I am.
You also don't need to paste your page copy in. Just hand your AI the URL and let it read the live page. It sees the real structure that way instead of words pulled out of context, and the type it picks is noticeably better for it.
One thing I noticed while doing this across my sites: the pages I'd built with AI came back cleaner every time. Real headings, real sections, actual structure. That's the same thing I teach inside Pages in Minutes, and it turns out the payoff isn't only that the page looks good. It's that machines can parse it.
If your pages came out of a drag and drop builder as a pile of nested divs, this still works. Your AI just has more guessing to do, same as the crawler.
What actually changes when your pages are labeled
❌ Before
- ❌ Crawlers guess what each page isa sales page and a blog post look the same
- ❌ No price, author, or date declarednothing structured to pull into a result
- ❌ Schema added and silently strippedyou assume it's live and never check
✅ After
- ✅ Every page declares its own typearticle, product, service, FAQ, person
- ✅ Google, Bing, and AI read the same labelone block covers both sides of search
- ✅ Verified live instead of assumedyou check before you move on
Where to start with your own schema
You need four things on hand before you begin. That's it.
- One page URLNot your whole site. The single page you'd most want an AI tool to cite when someone asks about what you do.
- An AI tool that can read a live pageSo it works from the real page instead of copy you pasted in.
- Access to wherever your page's code livesA settings panel, a code block, a template. Where it goes varies by platform.
- A way to check your workGoogle's Rich Results Test takes a URL and tells you what's actually on the page.
Let AI read the page and write the markup
Give it the URL, not your page copy. Then paste what it hands you.
Copy and paste
Read the live page at [your page URL] and tell me the most appropriate schema.org type for it (for example Article, FAQPage, Product, Service, Person, or LocalBusiness) and why that type fits better than the alternatives. Then generate complete, valid JSON-LD schema markup for that page inside a script tag. Only use information that actually appears on the page, and do not invent prices, ratings, dates, or authors. Finish with two things: a list of any recommended properties you left out because the page did not contain that information, and instructions for exactly where to paste this on [your website platform].
The takeaway
One labeled page beats a plan to label all of them. Start with the page you'd most want quoted back to you.
The full breakdown is inside The Shortcut
The prompt above gets you the markup. The rest of this shortcut is what keeps it from quietly falling off your site.
- The second prompt that checks whether your schema is actually live, and what to do when it isn't
- Where the markup goes on the platforms most of us use, including the one place it gets deleted without telling you
- The rules that stop AI inventing properties that will work against you
- How to choose which pages get labeled first when you have more than a few
- The setting that generates schema across every page at once, so you're not doing this one URL at a time
Let's be real about what schema does and does not do
Schema will not rank you. It's a label, not a lever. If your page is thin, labeling it correctly just tells search engines precisely what kind of thin page it is.
It also won't get you those fancy expandable FAQ results in Google anymore. Google stopped showing those for most sites back in 2023, and plenty of people quietly stopped bothering with schema because of it.
I think that was the wrong call. The rich result went away. The machine reading did not, and there's a lot more machine reading now than there was in 2023.
So this is a small, boring, permanent thing. Not a growth hack. Just your pages saying what they are instead of leaving it to a guess.
Common questions
What is schema markup?
Schema markup is a small block of structured data added to a web page that tells search engines and AI tools what that page is. It uses the shared vocabulary at schema.org, usually written in a format called JSON-LD, and it can declare things like the page type, the price of a product, the author of an article, or the questions and answers in an FAQ. Visitors never see it. Only machines read it.
Do I still need schema if I have an llms.txt file?
Yes, because they do different jobs. An llms.txt file sits at the root of your domain and describes your site as a whole to AI crawlers. Schema sits on an individual page and describes that specific page to Google, Bing, and AI tools alike. One is a map of the property, the other is a label on the door. Most sites benefit from having both.
Can AI write schema markup for me?
Yes, and this is one of the tasks it does better than most people would by hand. Schema is structured data with correct answers rather than creative writing, so an AI tool can read a live page, pick the right schema.org type, and produce valid JSON-LD in one pass. The one rule that matters: tell it to use only information that actually appears on the page, so it does not invent prices, ratings, dates, or authors.
Where do I put schema on my website?
It depends on your platform. Some website builders have a settings field or a toggle that adds schema for you, some let you paste a code block onto the page, and some require it to go in the site template instead. The important part is checking afterwards, because certain platforms strip script tags out of page content without showing an error, so schema you believe you added may never have reached the live page.
How do I know if my schema is working?
Run the page URL through Google's Rich Results Test. It will tell you what structured data it can find on the live page and flag anything invalid or incomplete. You can also view the page source and search for the words "application/ld+json". If nothing comes back in either place, the schema is not on the page, no matter what your website builder told you when you saved it.
Recommended Resources 🛠️
The tools behind this one.


