# How to Get Your SaaS Cited by ChatGPT, Claude, and Perplexity # How to Get Your SaaS Cited by ChatGPT, Claude, and Perplexity > **Quick answer:** To **get cited by ChatGPT, Claude, and Perplexity**, structure each page so a single passage answers the question completely on its own. That means a 40–60 word direct answer at the top, question-based headings answered in the first sentence, specific data and named sources, extractable formats (tables, step lists, FAQ blocks), and `Article`/`FAQPage`/`HowTo` schema. AI assistants quote clean, self-contained, verifiable passages — vague prose doesn't get lifted. **Key takeaways** - **Write the quotable sentence.** Each section should contain one passage an assistant could lift verbatim and be correct. - **Be specific.** Named sources, concrete numbers, and short quotations increase citation odds; generic claims don't. - **Structure for extraction.** Question headings, tables, step lists, and FAQ blocks. - **Mark it up.** Schema helps engines parse and attribute your content. - **Test it.** Run scheduled prompts across assistants and track whether your domain appears. Being ranked and being *cited* are now two different outcomes. A prospect asking Claude "what's the best way to connect Google Ads to an AI assistant?" gets a short answer naming two or three sources — and the vendors not named simply aren't in the consideration set. This guide covers the specific, tactical changes that make your B2B SaaS content citable, and how to check whether it's working. For the strategic overview, start with our [GEO/AEO playbook for B2B SaaS](https://www.growthspreeofficial.com/blogs/geo-aeo-b2b-saas). ## How do AI assistants choose which sources to cite? Generative assistants retrieve candidate passages, then synthesize an answer and attribute the passages they used. That means citation is passage-level, not page-level: a page can rank well and never be cited if no single passage cleanly answers the question. Research on the field (Aggarwal et al., "GEO: Generative Engine Optimization," Princeton, 2024) found that adding citations, quotations, and statistics to content measurably increased how often generative engines surfaced a source. The practical read: **make individual passages self-contained, specific, and verifiable.** ## What makes a passage citable? A citable passage answers the question completely without needing the paragraph before it. Compare: | Not citable | Citable | |---|---| | "There are several ways to approach this, and the right one depends on context." | "A GA4 MCP server connects Google Analytics 4 to an AI assistant so you can query analytics in plain English." | | "Many teams see strong results." | "Google's official Google Ads MCP server is read-only: it runs GAQL queries but cannot change bids or pause campaigns." | | Vague, hedged, context-dependent | Direct, specific, self-contained | The test: could an assistant quote this one sentence in an answer and be accurate? If not, rewrite it. ## How do you structure a page to get cited? 1. **Open with a direct answer.** Put a 40–60 word, self-contained answer immediately under the H1, before any preamble. 2. **Use question-based headings.** Phrase each H2 the way a person asks it, then answer in the very first sentence beneath. 3. **Add specific data and named sources.** Concrete figures and attributed claims are more quotable than generalizations. 4. **Use extractable formats.** Tables for comparisons, numbered lists for processes, an FAQ block for common questions. 5. **Add schema markup.** `Article`, `FAQPage`, and `HowTo` help engines parse and attribute your content correctly. 6. **Define entities plainly.** "An X is a Y that does Z" gives the model a definition to attribute to you. > **Field note:** The single highest-leverage edit on an existing page is moving the answer to the top. Most B2B posts bury the definition three paragraphs deep behind a narrative intro. Assistants rarely reach it. Lifting a clean, standalone answer into the first block — without changing anything else — is usually the fastest citation win available. ## Which formats get cited most often? - **Definitions** — "What is X?" answered in one clean sentence. - **Comparisons** — tables that state the trade-off explicitly. - **Step-by-step processes** — numbered, discrete, self-contained steps. - **FAQ pairs** — a question and a complete, standalone answer. - **Specific figures with attribution** — a number plus where it came from. ## How do you test and measure AI citations? You can't optimize what you don't observe. Build a simple, repeatable check: 1. **Write 10–20 category prompts** your buyers would actually ask ("how do I connect Google Ads to Claude?", "best MCP servers for marketers"). 2. **Run them across assistants** — ChatGPT, Claude, Perplexity — on a fixed schedule, and log whether your domain is cited. 3. **Track AI-referral traffic** in analytics. A [GA4 MCP server](https://www.growthspreeofficial.com/blogs/ga4-mcp-server) makes it a one-prompt question: "how much traffic and conversion came from AI-assistant referrers this month?" 4. **Watch your search footprint** with a [Search Console MCP](https://www.growthspreeofficial.com/blogs/search-console-mcp), since AI-influenced behavior shifts which queries you appear for. Because assistants vary in how they retrieve and cite, results differ by model — see [Claude vs. ChatGPT for marketing workflows](https://www.growthspreeofficial.com/blogs/claude-vs-chatgpt-marketing) for how they differ in practice. ## What doesn't work? Keyword stuffing, thin AI-generated filler, and "GEO" tactics that are just SEO relabeled. Assistants synthesize from passages they can verify; padding a page with keywords doesn't create a quotable sentence, and unsupported claims are less likely to be surfaced than attributed ones. Nor is there a way to force a citation — you earn it by being the clearest, most specific answer available. ## Frequently Asked Questions ### Q1. How do you get cited by ChatGPT? Structure each page so a single passage answers the question on its own: a 40–60 word direct answer at the top, question-based headings answered in the first sentence, specific data with named sources, extractable formats like tables and FAQ blocks, and schema markup. ### Q2. Why isn't my page cited by AI assistants even though it ranks? Because citation is passage-level, not page-level. If no single passage on your page answers the question cleanly and self-containedly, an assistant has nothing to quote. Move a clear, standalone answer to the top of the page. ### Q3. Does schema markup help you get cited? It helps engines parse and attribute your content correctly. Article, FAQPage, and HowTo schema make your definitions, questions, and steps machine-readable, which supports both rich results and AI extraction. ### Q4. How do I know if AI assistants are citing my SaaS? Run a fixed set of buyer-style prompts across ChatGPT, Claude, and Perplexity on a schedule and log whether your domain appears. Pair that with AI-referral traffic and conversions in your analytics. ### Q5. Can you pay to be cited by AI assistants? No. Citations are earned through being the clearest, most specific, most verifiable answer to the question. There's no paid placement for organic AI citations. **Sources & further reading** - Aggarwal et al., "GEO: Generative Engine Optimization" (Princeton, 2024). - Google — structured data and rich results documentation, Google Search Central. - Schema.org — Article, FAQPage, and HowTo type references. --- *Related guides: [GEO/AEO for B2B SaaS: The 2026 Playbook](https://www.growthspreeofficial.com/blogs/geo-aeo-b2b-saas) · [Claude vs. ChatGPT for Marketing Workflows](https://www.growthspreeofficial.com/blogs/claude-vs-chatgpt-marketing) · [Search Console MCP](https://www.growthspreeofficial.com/blogs/search-console-mcp) · [GA4 MCP Server](https://www.growthspreeofficial.com/blogs/ga4-mcp-server).*