Direct Answer
How do I get my business cited by ChatGPT and Perplexity?

To get cited by ChatGPT and Perplexity, you need three things in place: structured content (a direct answer capsule at the top of each key page), Schema markup (FAQPage and Organization Schema in JSON-LD), and topical authority (multiple interlinked pages covering your subject in depth). Both ChatGPT with browsing and PerplexityBot crawl live pages — well-structured, fast-loading pages with complete Schema are selected as citation sources ahead of unstructured pages regardless of domain age.

Why AI citation is worth pursuing

When Perplexity answers "best accounting firm for expats in Valencia" or ChatGPT answers "which AI web agency works with small businesses in Spain?", the cited source receives a referral with the user's intent already resolved. The AI has pre-qualified the lead, explained why you are relevant, and handed the user to your site with the decision nearly made.

Referral traffic from AI platforms converts differently from standard organic search. The user has received a direct recommendation — not just a link in a list — and typically arrives with a specific question to confirm rather than a broad intent to explore. For service businesses and B2B companies, this is a qualitatively different kind of inbound lead.

Important: AI citation is not a replacement for traditional SEO — it is an additional channel built on top of the same content and technical foundation. For the strategic framing of both channels together, see our comparison: AI SEO vs traditional SEO: what changed.

How ChatGPT and Perplexity select citation sources

ChatGPT (with browsing enabled) and Perplexity use different underlying architectures — ChatGPT uses Bing-indexed web content via Microsoft's search index, Perplexity uses its own crawler (PerplexityBot) — but both select citation sources based on overlapping signals:

  • Answer proximity: how close to the top of the page is the direct answer to the query? Pages that bury the answer in paragraph five are extracted less reliably than pages that lead with it.
  • Structured markup: FAQPage Schema gives crawlers machine-readable question-answer pairs; Organization Schema establishes entity identity. Both signals increase citation confidence.
  • Content completeness: does the page fully address the topic, or does it require the reader to follow multiple links to get a complete answer? Self-contained pages are preferred.
  • Domain credibility signals: existing external links, consistent NAP (name, address, phone) data across the web, and brand entity recognition in Google's Knowledge Graph all contribute.
  • Technical accessibility: pages that load slowly, are JavaScript-only rendered, or block common crawlers are indexed less frequently and cited less reliably.

Step-by-step: implementing AI citation optimisation

Add an answer capsule to every key page

Place a short, self-contained paragraph at the top of your homepage and each main service page. It should directly answer the primary intent of that page in 2-4 sentences. Write it to stand alone: if an AI system extracted only that paragraph, would it constitute a complete, accurate answer? If not, tighten it. Avoid preamble ("In this article we will discuss...") — lead with the answer. The format you are reading right now is an example of this pattern in practice.

Implement FAQPage Schema on every page with questions

Add FAQPage Schema (JSON-LD) to any page that contains a questions-and-answers section. Each question-answer pair in the Schema is directly extractable by AI crawlers without parsing prose. This is the single most efficient Schema implementation for AI citation because it pre-formats your content in the structure AI systems prefer. Validate your Schema with Google's Rich Results Test after implementation.

Complete your Organization Schema

Your Organization Schema should define at minimum: name, url, logo, contactPoint, areaServed and serviceType. This is how AI models build their understanding of what your business is and where it operates — it is your entity definition. Without it, AI systems must infer your identity from prose, which is less reliable. For local businesses, use LocalBusiness Schema with full address and opening hours.

Rewrite headings as questions

Convert your H2 and H3 headings from declarative to interrogative format where natural. Instead of "Our services", use "What services does [your business] offer?". Instead of "Why choose us", use "Why do businesses in [location] work with us?". This matches the natural language structure of queries that users type into AI assistants, and gives extraction systems clear anchors for sub-answers within your content.

Build a topical authority cluster

A single optimised page is a starting point; a cluster of interlinked pages on related subtopics is what builds citation authority over time. If your core topic is AI web design for small businesses, your cluster might include: a main service page, an article on what AEO is, an article on GEO, a comparison of AI SEO vs traditional SEO, and a guide on how to get cited by AI assistants — each page linking to the others. This signals to AI systems that your domain is a primary reference on the topic, not a single-page resource.

Ensure technical crawlability

Check that PerplexityBot and other AI crawlers are not blocked in your robots.txt. Ensure your key pages load under 2 seconds — slow pages are crawled less frequently. Avoid JavaScript-only rendering for content that should be cited: server-side rendered or static HTML is indexed more reliably by all crawlers, including AI-specific ones. Run a Lighthouse audit and target a Performance score above 90.

How to measure whether it is working

AI citation measurement is currently more manual than traditional rank tracking, but the signals are clear:

  • Weekly manual citation audit: query your 5-10 target topics in ChatGPT, Perplexity and Google AI Overviews. Record which sources are cited. Track your brand appearance rate over time. This takes 15-20 minutes weekly and is the most direct signal available.
  • GA4 referral source filter: in GA4, filter Traffic Sources to include sessions from chatgpt.com, perplexity.ai, bing.com/chat and claude.ai. Monitor session volume and conversion rate from these sources separately from standard organic Google traffic.
  • Google Search Console AI Overview data: the Performance report in GSC now includes an AI Overview filter showing impressions and clicks from citations in Google AI Overviews. This is the easiest automated measurement available for the Google channel.
  • Branded search growth: as AI citation builds awareness, branded search queries for your business name typically increase. Track this in GSC as a secondary confirmation signal.

For the full context of these signals within a broader AI SEO strategy, see our Answer Engine Optimization service page and the foundational guide on what AEO is.

Common mistakes that prevent AI citation

Most sites that fail to appear in AI citations share one or more of these patterns:

  • Burying the answer: long preamble before the direct response. AI systems extract the most answer-like block from a page — if that block is 400 words into an article, the extraction is less reliable.
  • Missing or incomplete Schema: a site with no Organization Schema is, from an AI model's perspective, an unidentified entity. Without entity definition, citation is possible but inconsistent.
  • Single-page depth on a broad topic: trying to address an entire topic cluster in one article instead of building depth across multiple interlinked pages. AI systems weight domains that demonstrate coverage breadth, not just content length.
  • Blocking AI crawlers in robots.txt: some sites that blocked aggressive scrapers in the past inadvertently blocked AI citation crawlers. PerplexityBot and Anthropic-AI bot should be permitted unless there is a specific legal reason to block them.
  • No visible publication date: freshness is a citation signal. AI systems prefer pages with a clear datePublished in Article Schema and a visible date on the page.