Beyond Schema Markup: What Makes a SaaS Site Citable by AI

A SaaS site becomes citable by AI engines when their crawlers are allowed in and can read the main content in your server's HTML. They also need a clear page per buyer question, plain links to it, and a way to tell whose site it is.

Google's guidance on AI features, last updated December 10, 2025, says there is no special schema.org structured data you need to add to appear in AI Overviews or AI Mode.

We checked the OpenAI, Anthropic and Google crawler names and every Google rule on this page against those companies' own documentation in October 2026.

Is schema markup enough to get a SaaS site cited by AI engines?

Not by itself. Google describes structured data as explicit clues about the meaning of a page, and says markup enables a search feature without guaranteeing it.

For AI Overviews and AI Mode, Google says there are no additional requirements or special optimizations needed. On schema, it advises that structured data should match the page's visible text.

On SaaS sites, markup and text drift apart in ordinary ways:

• Stale FAQ markup: markup added through a tag manager goes stale after a pricing change.
• Invisible plans: Product markup lists plans that only a client-side widget displays.

Keep markup valid and matched to the text, since it still makes pages eligible for rich results in Google. Our guide to schema markup for AEO covers which types and properties to use. The rest of this page covers the structure underneath: citation architecture.

What has to be true before an AI engine can cite a page?

Its crawler has to be allowed to fetch the page. For Google, the page must also be indexed and eligible to show in Search with a snippet.

Let each crawler in

Google lists allowing crawling "in robots.txt, and by any CDN or hosting infrastructure" among the fundamentals for its AI features. A bot-protection rule at the CDN or firewall can turn away a crawler that robots.txt allows.

OpenAI and Anthropic each run separate bots for training, for search and for fetches a user requests. The OpenAI rows below come from OpenAI's crawler documentation.

The Anthropic rows come from Anthropic's crawler help article, and the Google rows from Google's list of common crawlers.

Google-Extended is not a crawler: Google says it has no user agent of its own and is used only as a robots.txt control.

Treat training and search separately

For SaaS legal teams, training and search are separate switches:

• OpenAI: its settings are independent, so a site can allow OAI-SearchBot and disallow GPTBot.
• Anthropic: it documents ClaudeBot and Claude-SearchBot as separate bots with separate effects.

Check every subdomain

Check every host, including docs, help and app subdomains. Anthropic asks site owners to set robots.txt rules for every subdomain they want to opt out.

This cuts both ways: a blanket disallow on a docs or help subdomain hides your setup guides from search bots. Our robots.txt guide for AI crawlers has the full crawler audit.

Do AI crawlers render JavaScript on a SaaS site?

Plan as if they see only the HTML your server sends.

What the research found

In Vercel and MERJ's study published December 17, 2024, none of the major AI crawlers rendered JavaScript, including OpenAI's OAI-SearchBot, ChatGPT-User and GPTBot, Anthropic's ClaudeBot and PerplexityBot.

The ChatGPT and Claude crawlers fetched JavaScript files (11.50% and 23.84% of their requests) without executing them. Gemini was the exception, because it uses Googlebot's infrastructure for full rendering. The study, The rise of the AI crawler, dates from December 2024, so test your own pages.

Google's JavaScript SEO basics say server-side or pre-rendering is still a good idea because not all bots can run JavaScript. Google adds that content missing from the rendered HTML cannot be indexed.

Where SaaS sites usually break

• Pricing tables: plans and prices filled by an API call.
• Feature tabs: text that loads only on click.
• Integration directories: lists filtered in the browser.
• Docs sites: documentation built as a single-page app.

Vercel's advice is to server-render main content such as product information and documentation, meta information and navigation.

How to test a template

1. Fetch the raw HTML. Request the template with curl, not a browser.
2. Search for key text. Look for a plan name, a price and a feature sentence.
3. Fix what is missing. Whatever is missing there is missing for every crawler that does not render.

How should a SaaS site be organized so each page can be cited?

Give each real buyer question its own page, answer it in text, and group those pages under hubs.

Plan pages around query fan-out

Google says AI Overviews and AI Mode may use "query fan-out", issuing multiple related searches across subtopics and data sources. A question about which expense tool suits a growing finance team can become searches about approval rules, integrations, pricing and alternatives. Each search needs a page that answers its subtopic clearly.

Google also asks that important content be available in textual form, so do not lock the answer inside a video or an image.

To find subtopics, OnlyAEO asks ChatGPT and Claude the buyer's question before writing, reads every page they cite in full, and turns each search they ran into a section of the article.

Publishing more posts is not the fix. Posts that each partly answer "how do I choose an expense tool" spread the answer thin, while a single page that answers it fully, under a hub, gives an engine a clear source.

A before-and-after example

This is an illustrative example, not a client: a fictional expense-approval product at example.com.

For on-page format, see what content structure gets cited by AI assistants.

How do internal links help AI engines find and cite SaaS pages?

Internal links are how crawlers find your pages, and anchor text tells them what each page is about.

Google says it uses links as a relevancy signal and to find new pages to crawl, and that every page you care about should be linked from another page on your site.

Both come from Google's guidance on crawlable links. Its AI features guidance lists internal links among the fundamentals that still apply.

JavaScript front ends are where this breaks:

• Use real links: Google can generally crawl a link only if it is an <a> element with an href attribute, so buttons with click handlers are invisible as links.
• Avoid URL fragments: in single-page apps, Google says not to load different content through URL fragments and to use the History API instead. Docs that all live behind /docs#something look like a single page.
• Write descriptive anchors: Google says better anchor text helps people and Google understand the linked page. "Approval rules for multi-entity finance teams" says what a page answers, and "learn more" does not.
• Build hubs around questions: the features hub links to each feature page, feature pages link to the integrations and use cases they support, and docs link back.

How do you build a SaaS brand entity that AI engines recognize?

Describe the company the same way everywhere, and connect those places with Organization markup on your home page. Google says this markup can help it disambiguate your organization in search results.

Organization markup

Google's Organization structured data documentation spells out the properties that matter here:

• The url property: helps Google uniquely identify the organization.
• The sameAs property: lists your profiles on other sites, such as social media or review sites.
• The name: Google asks you to use the same name that you use for your site name.

The full property list is in Google's Organization markup guide.

List your G2 or Capterra profile in sameAs, and keep its company name, category and description consistent with your own site. When a buyer compares vendors, that profile may be among the pages an engine reads about you. Our piece on G2 reviews and AI recommendations covers the review-site side.

Company and product names

A company name that differs from the product name, or a renamed product, makes this harder. An engine that sees the company name on LinkedIn and the product name on review sites has to guess they are related.

• State the relationship: say it in plain text on your home and about pages ("Example Spend is the spend-management product from Example Labs").
• Pick your name: use the name you want engines to use everywhere.
• Explain old names: keep old product names on a page that explains the change.

For the step-by-step entity build, read our guide to building a brand entity.

What makes the content on those pages worth citing?

Original information that a reader can check.

Google's guidance on helpful, reliable, people-first content asks whether content provides original information, reporting, research or analysis. It also asks whether content earns trust through clear sourcing, evidence of expertise and background about the author or site.

Its link guidance adds that citing your sources can help establish trustworthiness.

SaaS teams sit on unpublished original material: real plan limits, working setup steps, integration edge cases, changelog dates and the answers support gives every week. Publish it under the name of someone who knows the product, with an updated date that moves only when the page changes.

The GEO paper on arXiv, accepted to KDD 2024, reports that GEO methods can boost visibility by up to 40% in generative engine responses. It is a research result on the engines and queries the authors tested, so measure your own site. Our step-by-step guide to measuring AI visibility shows how.

Can a content optimization tool replace technical AEO for a SaaS site?

No. A content intelligence platform grades and briefs the writing, while technical AEO makes sure AI crawlers can fetch, read, find and attribute your pages.

• Content intelligence platform: grades drafts and briefs writers, measuring content quality against search ranking signals.
• Technical AEO: fixes crawler rules, rendering, page structure, internal links, entity and markup, and measures which pages AI crawlers read and where AI engines cite or mention the brand.

A page can score well in a content tool and still be invisible to a crawler that reads only the server's HTML.

How do you audit a SaaS site's citation architecture?

Check each area host by host and template by template. If an integration page fails, every page built on that template fails the same way.

1. Crawl access. Check robots.txt rules for each search and training bot on every host, plus CDN and firewall bot rules.
2. Rendering. Compare raw HTML with the browser view for each template.
3. Page architecture. Map each buyer question to a page that states the answer in text.
4. Links and entity. Check for <a href> links with descriptive anchors to every important page, Organization markup with url and sameAs, and the same name and description on every profile.
5. Structured data and sources. Confirm valid markup that matches visible text, named authors, real updated dates and linked sources.

Rerun the audit after a redesign, a CDN change or a docs migration, since any of those can quietly undo crawl access or rendering.

How do you choose a partner for this, and where does OnlyAEO fit?

Choose the partner that can show you, page by page, what AI crawlers fetch from your site and what they miss. It should be plain about what it cannot promise.

What to ask a partner

• Crawler access: checked per engine and per host, docs subdomains and CDN rules included.
• Rendering: raw HTML compared with the rendered page, briefed to your engineers.
• Page planning: pages planned from buyer questions, including which to merge.
• Off-site work: review profiles and listings, beyond your own site.
• Bot visibility: a view of which of your pages AI bots read.
• Honest limits: no promise of rankings or citations.

Where OnlyAEO fits

OnlyAEO, an AEO and GEO agency, works across ChatGPT, Claude, Gemini, DeepSeek, Perplexity and Grok. Our managed service includes:

• A website health check so every AI engine can read your site.
• Researched, fact-checked articles published on your site in your voice and design.
• Help getting mentioned on the trusted sites and listings AI reads.
• AI crawler tracking that logs which pages AI bots read, so you can check whether AI bots ever fetch your docs subdomain or pricing page.

Every article is written from a Context Graph, a single record of your company's facts, voice, buyers and the claims it can and cannot make. The full process is on our how it works page.

One client, FastTrackr AI, started at 0.1% AI visibility (#338 of 458 brands); about 100 days later it was the #3 most-cited brand in its category and the #4 most-cited source. The full story is in our case studies.

We went from being essentially invisible in our category to the #3 most-cited brand inside one quarter.

There is no results guarantee: we do not promise specific rankings or citations. The first step is a free AI visibility audit of how ChatGPT, Claude, Gemini and DeepSeek see your brand, at no cost and back within 48 hours.

Frequently asked questions

Will blocking Google-Extended remove our site from Google's AI Overviews?

No. Google says Google-Extended does not affect a site's inclusion in Google Search and is not used as a ranking signal.

Can broken structured data hurt our Google rankings?

Not directly. Google says a structured data manual action removes a page's eligibility for rich results but does not affect how the page ranks in Google web search.

Can we see which pages AI bots read on our site?

Yes. OnlyAEO's tracking recognises 193 AI, search and ads bots, logs which pages AI bots read and captures AI-referred visits that GA4 misses, with a 10-minute install.

How long before structural fixes show up in AI answers?

In OnlyAEO's client work, the first measurable results usually arrive after 60 days or more. We do not guarantee rankings or citations.