AI Search Optimisation:
How to Get Cited by ChatGPT, Perplexity
and Google AI Overviews

Three channels, three architectures

"AI search" is frequently discussed as a single phenomenon, but ChatGPT, Perplexity and Google AI Overviews have meaningfully different technical architectures — and those differences change which optimisation tactics matter most for each.

ChatGPT (with browsing)

When a user runs a web search in ChatGPT, the model queries Bing's search index via Microsoft's API. Pages that rank well in Bing — or that are indexed by Bing at all — are candidates for retrieval. ChatGPT then synthesises an answer from retrieved content and cites the sources it drew from in a sidebar or inline.

This means: if your pages are not in Bing's index, ChatGPT cannot cite them regardless of your Google rankings. Bing Webmaster Tools verification and active sitemap submission are prerequisites for this channel.

Perplexity

Perplexity operates its own crawler, PerplexityBot, which indexes pages continuously and independently of both Google and Bing. Perplexity's source selection algorithm weights answer proximity, content completeness, and domain credibility signals more heavily than traditional keyword matching.

The practical implication: Perplexity can surface pages that do not rank in traditional Google search, but only if those pages are not blocking PerplexityBot in robots.txt and load quickly enough to be crawled regularly.

Google AI Overviews

Google AI Overviews (formerly SGE) sources content from Google's own search index. Pages that rank in the top 10 for a query are not automatically included in AI Overviews — Google selects sources based on additional signals including direct-answer content, Schema markup, and what Google's systems identify as the most authoritative response to a specific question.

The starting point here is solid: if your pages rank in Google, they are candidates for AI Overviews. The optimisation layer is about answer structure and Schema markup on top of existing Google SEO.

Tactic 1: Answer-first content structure (shared across all three channels)

The single tactic with the highest cross-channel leverage is answer-first content structure. All three AI search systems reward pages that place a direct, self-contained answer close to the top of the page, under a question-shaped H2 heading, before expanding into detail.

The format:

• H2 or H3 phrased as a question ("How does [topic] work?" or "What is [term]?")
• A capsule of 2–4 sentences directly answering that question, before the explanatory paragraphs
• The key claim bolded within the capsule

A page with answer-first structure at both the page level and the section level gives AI search systems multiple extraction points: the full-page answer for broad queries, and section-level answers for specific sub-questions. For the foundational explanation of this approach, see our article on how to get your business cited by ChatGPT and Perplexity.

Tactic 2: Schema markup — the machine-readable layer

Schema markup in JSON-LD is the second shared tactic. It provides AI search systems with structured, machine-readable signals that do not require parsing prose.

The Schema stack for AI search optimisation:

• Organization: defines your entity. At minimum: name, url, logo, contactPoint, areaServed, sameAs. The sameAs property should link to at least one external, independently verifiable profile (LinkedIn is most commonly effective). This is your entity identity card for AI systems.
• FAQPage: on any page with a questions section. Each Q&A pair becomes a discrete extraction unit — the model can answer a specific question by pulling one FAQ item rather than parsing a full article.
• Article: on blog posts. headline, datePublished, dateModified, author, publisher. Freshness signals matter to AI search, particularly for Perplexity.
• Service: on service pages. serviceType, provider, areaServed, description.

Tactic 3: Index coverage — channel-specific prerequisites

This is where AI search optimisation diverges by channel.

For ChatGPT: verify Bing index coverage. Go to Bing Webmaster Tools, add your property, submit your XML sitemap, and run an index check on your key pages. Pages missing from Bing's index are invisible to ChatGPT web search. Also ensure Bing's crawlers and OAI-SearchBot are not blocked in your robots.txt.

For Perplexity: ensure PerplexityBot is explicitly permitted in robots.txt. Check page speed: pages loading over 3 seconds are crawled significantly less frequently by Perplexity's infrastructure. Run Lighthouse on your key pages; target above 90 for Performance.

For Google AI Overviews: the prerequisite is traditional Google SEO — pages indexed in Google Search Console, healthy Core Web Vitals, no manual actions. The optimisation layer is answer structure and Schema on top of existing rankings.

For a strategic perspective on how these channels sit alongside traditional SEO, see our article on AI SEO vs traditional SEO.

Tactic 4: Topical authority through content clusters

All three AI search channels give higher citation weight to domains that cover a topic with depth. Build topical authority with:

• A primary pillar page covering the topic overview (this acts as the anchor)
• Supporting articles each covering one specific subtopic, each with its own answer structure and Schema
• Internal links between all pages using descriptive anchor text
• Consistent terminology: use the same terms for the same concepts across all pages

Tactic 5: Freshness and publication signals

Perplexity weights freshness more heavily than ChatGPT or Google AI Overviews. For time-sensitive topics, pages with a recent dateModified in Article Schema and a visible update date on the page are cited more frequently.

Practical application: review your top-performing pages every three to four months. If the content is still accurate, update the dateModified in Schema and make a substantive update to at least one section. This refreshes the freshness signal without requiring a full rewrite.

Channel-by-channel comparison: where to start

Where to start if you have limited time: implement answer capsules and FAQPage Schema on your top three pages, verify Bing Webmaster Tools and Google Search Console, and check robots.txt permits OAI-SearchBot and PerplexityBot. These five actions cover the highest-leverage ground across all three channels simultaneously.

Measurement: tracking AI search performance across all three channels

• Weekly citation audit (20 minutes): query your 5 target topics in ChatGPT, Perplexity and Google AI Overviews. Record which domains are cited, and specifically whether your brand is named by name.
• GA4 referral source breakdown: filter Traffic Acquisition to isolate sessions from chatgpt.com, perplexity.ai, claude.ai, bing.com/chat and bard.google.com.
• Google Search Console AI Overview filter: the Performance report in GSC now includes an AI Overview filter — the only automated, channel-specific measurement for Google AI Overviews.
• Bing Webmaster Tools: provides impression data for Bing-indexed pages and helps identify crawl issues affecting ChatGPT coverage.

For the full context of these tactics within the GEO framework, see our foundational article on Generative Engine Optimisation explained.