How SaaS Companies Actually Get Cited by ChatGPT and Perplexity

November 12, 2026

By Steve Merrill, Founder of WRKNG Digital — November 12, 2026

How do SaaS companies get cited by ChatGPT and Perplexity?

SaaS companies get cited by ChatGPT and Perplexity when their docs, pricing, and comparison pages answer specific buyer questions in plain text that AI crawlers can read, and when review sites like G2 and Capterra say the same things about them. AI recommends the product it can verify. Everyone else gets skipped.

That's the whole game. The rest of this post is how to do it.

I came up in ecommerce. I built a clothing brand to about $10 million a year, and most of our AEO work at WRKNG Digital started with Shopify stores. When we started running the same prompt checks on SaaS products, I expected a totally different playbook. It wasn't. Software buyers ask AI the same kind of questions shoppers do. "What's the best CRM for a 10-person agency?" "Is Notion or Coda better for docs?" "How much does Intercom cost?" The AI answers with a short list. You're on it or you're not.

Why do AI answers skip most SaaS websites?

Most SaaS sites are built to impress a human on a demo call. AI isn't a human on a demo call.

The homepage says something like "The all-in-one platform for modern teams." A language model can't do anything with that. It doesn't say who the product is for or what it replaces. When ChatGPT goes looking for "best project management tool for construction companies," a vague hero line gives it nothing to quote.

We see the same few problems over and over:

  • Pricing hidden behind "Contact sales." If the AI can't find a price, it pulls one from a review site or a two-year-old Reddit thread. Or it picks a competitor that lists pricing openly.
  • Pages that render only in JavaScript. A lot of marketing sites and docs portals ship an empty HTML shell and fill it in with scripts. Some AI crawlers don't run those scripts. They see a blank page.
  • Bot protection blocking AI crawlers. Security teams turn on aggressive bot rules in Cloudflare and never check which bots got caught.
  • Comparison pages that read like attack ads. AI tends to discount one-sided content. A page that says you win on every row isn't believable to a person either.

None of this is hard to fix. It just never got anyone's attention, because nobody was measuring it.

Which SaaS pages actually get cited in AI answers?

Five page types do most of the work. Treat them as the core of your AI visibility.

Docs and help center pages

Docs are the most underrated asset a SaaS company owns. When someone asks Perplexity "Does Tool X integrate with HubSpot?" the best possible source is a doc page titled "HubSpot integration" that answers yes or no in the first sentence, then explains the setup. Docs are specific and get updated often. That's exactly what answer engines want.

The catch: docs often live on a separate subdomain, behind a JavaScript-heavy help center tool, sometimes with a login wall. Check that public docs return real text in the HTML.

Pricing pages

Pricing questions are some of the most common software prompts. "How much does X cost?" "Is there a free plan?" "What's the cheapest option for a small team?"

List plan names, monthly and annual prices, seat or usage limits, and who each plan fits. Write it in plain text, not baked into an image or a toggle that only loads on click. If enterprise pricing is custom, say that, and give a starting point if you can.

Comparison and alternatives pages

"X vs Y" and "Alternatives to Y" pages map straight onto how people prompt AI. The ones that get cited are honest. They say where the competitor is the better pick. That honesty is what makes the rest of the page credible enough to quote.

Integration pages

One page per integration, with a clear title and a one-line answer to "does it connect to this?" A single grid of logos doesn't cut it. Logos aren't text.

Changelogs and release notes

AI answers go stale. A public, dated changelog gives the model fresh facts to work with. When you ship a feature a competitor's comparison page says you're missing, your changelog is the proof.

How much do G2 and Capterra reviews matter for AI visibility?

A lot. Probably more than your own site for some prompts.

ChatGPT and Perplexity cross-check. When your site says you're the best help desk for ecommerce brands and G2 reviewers say the same thing in their own words, that agreement is a strong signal. When your site says one thing and the reviews say another, the AI tends to go with the reviews.

I learned this the hard way in ecommerce. You can write the perfect product page, and if the reviews and the forum threads tell a different story, the AI repeats the forum threads. SaaS works the same way. G2, Capterra, TrustRadius, the Shopify App Store or Salesforce AppExchange if you're listed there, Reddit threads in your category. All of it feeds the answer.

Practical moves:

  • Make your category, pricing, and main use cases match across your site and every review profile.
  • Ask happy customers for reviews that name a specific use case and company size. "Great tool" doesn't help anyone. "We moved our 12-person support team off Zendesk" does.
  • Reply to reviews, including the bad ones. Those replies get crawled too.

Can AI crawlers actually reach your SaaS site?

Check this first. It's the cheapest fix there is.

OpenAI runs separate crawlers for different jobs. Per OpenAI's crawler documentation, GPTBot collects training data, OAI-SearchBot powers ChatGPT search results, and ChatGPT-User fetches pages when a user asks. Perplexity documents its own bots, PerplexityBot and Perplexity-User, in its crawler guide.

Plenty of SaaS companies blocked GPTBot in 2023 out of training concerns and never looked again. That's a fair choice. But if a blanket rule also blocked OAI-SearchBot, you've removed yourself from ChatGPT search answers. Look at robots.txt and your CDN's bot settings. Both.

Then add structured data. Schema.org's SoftwareApplication type lets you mark up application category, operating system, offers, and ratings in a format machines read without guessing. Add FAQPage schema on pages built around questions.

How do you get your SaaS product cited, step by step?

Start with measurement. Everything else depends on it.

  1. Run a prompt baseline. Write 20 to 30 prompts a real buyer would type. Mix category prompts ("best onboarding software for B2B SaaS"), comparison prompts ("Appcues vs Userpilot"), and pricing prompts. Run them in ChatGPT and Perplexity. Log whether you're mentioned, cited with a link, or missing, and which URLs get cited.
  2. Open the door to AI crawlers. Fix robots.txt and bot-protection rules so the search and user-fetch bots can reach public pages.
  3. Make core pages readable without JavaScript. Server-render pricing, docs, integrations, and comparison pages. Load the page with scripts off. If it's blank, that's your problem.
  4. Publish real pricing and honest comparison pages. Plain text. Real numbers. Clear "best for" statements, including where a competitor fits better.
  5. Add structured data. SoftwareApplication with offers on product and pricing pages, FAQPage on question-driven pages.
  6. Align third-party sources. Update G2, Capterra, and marketplace listings so the story matches your site. Then push for specific, use-case reviews.
  7. Re-run the baseline monthly. Use the identical prompt set in both engines. Track what moved and which URLs started getting cited.

Read that first step again. Most teams skip it and go straight to writing content. Then they have no idea whether anything worked.

What should SaaS teams stop doing?

Stop writing top-of-funnel blog posts that could belong to any company. "10 Tips for Better Team Productivity" won't get your product named in an answer. AI already has a thousand of those.

Write the pages only you can write. Pricing logic, integration details, and an honest take on when a customer should pick you and when they shouldn't. That's the content an answer engine can't find anywhere else, so it has to cite you.

Frequently asked questions

Do SaaS companies need to pay to show up in ChatGPT or Perplexity answers?

No. Organic citations in ChatGPT and Perplexity come from what their search systems can crawl and verify. Clear pages and consistent third-party reviews matter more than budget.

Should a SaaS company block GPTBot?

GPTBot is OpenAI's training crawler. OAI-SearchBot is the one tied to ChatGPT search results. You can block training and still allow search, but blocking OAI-SearchBot keeps you out of ChatGPT search answers.

How long does it take for a SaaS company to start getting cited by AI?

Perplexity searches the live web, so fixes to crawlable pages can show up within weeks. ChatGPT search behaves similarly. Changes that depend on review volume and third-party mentions take months.

Is AEO for SaaS different from SEO?

They overlap. SEO gets a page ranked. AEO gets a specific fact from that page quoted in an answer, which rewards direct answers, plain pricing, and agreement across third-party sources.

Want to know where your SaaS product stands in AI answers?

We run the prompt baseline, check crawler access, and map exactly which pages need work. You'll see which prompts name you and which ones hand the spot to a competitor.

Get your AI visibility audit from WRKNG Digital.

More on SaaS-specific prompt patterns next week.

Back to Blog