A team adds FAQPage schema to its blog, waits a few weeks, and checks whether ChatGPT or Perplexity is citing the site. Sometimes it works. More often, the FAQ answers get pulled but the step-by-step guide next to them, the pricing page, and the about page stay invisible to AI answers. The reason is usually not that the content is bad. It is that FAQPage is only one of several schema types AI crawlers read, and the rest of the site has none.
Schema markup is structured data, written as JSON-LD, that tells a machine what a piece of content is before it tries to read the words around it. Google's structured data documentation and the schema.org vocabulary define dozens of types, but only a handful show up often enough on real pages to matter for AI search: FAQPage, HowTo, Article, Organization, and BreadcrumbList. Each one answers a different question for the crawler, and each one earns a different kind of citation.
Which Schema Types Do AI Search Engines Actually Read?
AI crawlers such as GPTBot, ClaudeBot, PerplexityBot, and Google-Extended fetch the raw HTML, including the JSON-LD block, before any rendering happens. That means whatever schema is present in the page source is what the crawler sees - not what a human sees after JavaScript runs.
| Schema type | What it declares | Best for |
|---|---|---|
| FAQPage | A list of question/answer pairs on the page | Getting a direct quote-and-answer citation |
| HowTo | An ordered sequence of steps with a named outcome | Step-by-step queries like "how do I fix X" |
| Article | Headline, author, publish date, and body of a piece of content | Establishing the page as a dated, attributable source |
| Organization | The entity behind the site: name, logo, contact point | Being named correctly when an AI answer cites the brand |
| BreadcrumbList | The page's position in the site hierarchy | Helping a crawler understand what a page belongs to |
A site that ships FAQPage on every post but never uses HowTo is answering "what is X" questions well and "how do I do X" questions not at all, even when the article contains a perfectly good numbered list.
How Does HowTo Schema Help You Get Cited in Step-by-Step Answers?
HowTo schema requires a name for the overall task and at least two ordered HowToStep items, each with its own short text. When a reader asks an AI engine a process question, the engine favors a source where the steps are already segmented and labeled, because it can lift them without having to parse a paragraph to find the sequence.
The mistake most sites make is writing a clean numbered list in the visible HTML but never wrapping it in HowTo JSON-LD. The crawler still sees an <ol>, and can often extract it anyway, but explicit HowTo markup removes the guesswork and is more likely to survive the trip into an AI Overview or a Perplexity answer card intact.
- Identify pages that already contain a numbered process, not just a list of facts.
- Write a one-sentence
namethat states the outcome, such as "How to add FAQ schema to a blog post." - Break the existing steps into individual
HowToStepobjects with a shorttextfield each. - Keep the visible
<ol>in the page body matching the schema one-to-one, so the two never drift apart. - Validate the JSON-LD parses before publishing, since one malformed brace drops the entire block.
Checking this by hand means opening dev tools on every page, copying the JSON-LD, and running it through a validator one page at a time. That gets slow past a few dozen URLs, which is the point where an automated audit that parses JSON-LD, headings, and AI crawler permissions in one pass starts saving real time instead of being a nice-to-have.
Should You Use Organization Schema If You Already Have FAQ Schema?
Yes, and the two do different jobs. FAQPage tells the crawler about the content on one page. Organization schema tells it who the site belongs to: the legal or brand name, the logo, and a contact point. When an AI engine cites a source and needs to name the publisher in its answer, it pulls that name from Organization schema if it exists, or guesses from the domain and page title if it does not.
A site with strong FAQ content but no Organization markup can still get its facts quoted, misattributed to the wrong brand name, or attributed generically as "a website." Adding Organization schema once, sitewide, closes that gap for every page at the same time, which makes it one of the highest-leverage single additions on this list.
What Happens If Your Schema Has a Syntax Error?
Most JSON-LD parsers, including the ones AI crawlers rely on, stop at the first invalid character and discard the entire block rather than salvaging the valid parts. A missing comma or an unescaped quotation mark in one FAQ answer can silently zero out the FAQPage, HowTo, and Organization schema on that page, even though two of the three were written correctly.
This is the failure mode that is hardest to catch by reading the rendered page, because broken schema produces no visible symptom. The page looks fine to a human visitor and to a quick glance in the browser. The only way to know it is broken is to parse the JSON-LD directly, which is why manual schema audits tend to miss it until traffic from AI referrals unexpectedly drops.
Most teams could build a script to check this themselves. In practice, most don't, because it means writing and maintaining a parser, a crawler-permission checker, and a Core Web Vitals pull as three separate tools. AskEO runs all three in one pass and returns the result in about 30 seconds, which is the argument for using a ready-made scan instead of assembling the equivalent from scratch.
Is It Worth the Effort to Add More Schema Types?
The honest objection here is time: adding HowTo, Article, and Organization schema across an existing site is real work, and a team with limited engineering hours might reasonably ask whether FAQPage alone is good enough. It is not good enough for step-by-step or entity-attribution queries, but it is a legitimate starting point. The return on the remaining schema types is not uniform - Organization schema is a one-time, sitewide fix, while HowTo has to be added page by page as process content is published. Prioritize Organization first, then HowTo on the pages that already contain numbered steps, and treat Article schema as a lower-priority pass for older evergreen posts.
How Do You Confirm the Schema Is Actually Working?
Three checks confirm schema is live and readable: the JSON-LD parses without error, the AI crawler user agents (GPTBot, ClaudeBot, PerplexityBot, Google-Extended, Applebot-Extended) are allowed in robots.txt, and the specific schema type appears where the content type calls for it. A site can pass the first two checks and still fail the third if HowTo content is marked up as generic Article instead.
Run an audit, fix what it flags, and re-run it after the fix ships rather than assuming the change worked. Create a free AskEO account and scan your homepage - it checks your schema, crawler access, and headings in about 30 seconds and tells you exactly which type is missing.
Frequently Asked Questions
Does FAQPage schema still matter if I add HowTo and Organization too?
Yes. FAQPage, HowTo, and Organization schema each cover a different content shape and can all exist on the same page. Adding HowTo or Organization schema does not replace FAQPage; it fills in the gaps FAQPage leaves for process questions and brand attribution.
Can I add multiple schema types to one page?
Yes. A single page can carry FAQPage, HowTo, and Article schema at once inside one JSON-LD @graph array, as long as each object has its own valid @type and the JSON itself parses without error.
Does HowTo schema require images for each step?
No. An image per HowToStep is optional in the schema.org specification. A name and text field for each step is enough for an AI crawler to extract the sequence; images help human search results more than AI citations.
How do I know if an AI crawler is blocked from reading my schema?
Check robots.txt for explicit Allow or Disallow rules against GPTBot, ClaudeBot, PerplexityBot, Google-Extended, and Applebot-Extended. If a crawler is disallowed, it will not fetch the page at all, so the schema underneath never gets read regardless of how well it is written.
Will adding schema markup guarantee an AI citation?
No. Schema markup makes content easier for an AI crawler to parse and attribute correctly, but citation also depends on content quality, crawler access, and whether the answer engine judges the source authoritative for that query. Schema removes a technical barrier; it does not replace substance.