Generative Engine Optimisation (GEO): A 2026 Playbook

A practical GEO playbook for 2026: which crawlers to allow, what research says gets content cited, how to structure pages, build off-site presence and measure AI citations across ChatGPT, Gemini and Perplexity.

SSquarebase
Published 1 Oct 2026 · 9 min read

Quick answer: Generative Engine Optimisation (GEO) is the work of making your brand easy for AI assistants to find, trust and cite. In 2026 that means letting AI search crawlers in, ranking well in classic search, writing pages with clear answers, facts and sources, keeping brand facts consistent across the web, and tracking citations in Bing and referral traffic from ChatGPT.

Generative Engine Optimisation, or GEO, is now a normal line item in a search plan. Buyers ask ChatGPT, Gemini, Perplexity and Google AI Overviews for shortlists and comparisons, and those tools answer with a few cited sources. If your brand is not one of them, you are not in the conversation.

This playbook covers what actually influences AI citations, based on platform documentation and published research, and gives you a checklist your team can run this quarter. It avoids tactics that nobody can verify.

Key takeaways

Each AI assistant uses its own crawler. If you block OAI-SearchBot or PerplexityBot, you cannot be cited in ChatGPT search or Perplexity.
Google says there are no special requirements for AI Overviews or AI Mode. Indexable, snippet-eligible pages with strong SEO are the entry ticket.
Research on GEO found that adding statistics, quotations and cited sources raised visibility in AI answers, while keyword stuffing did not.
Structure matters: clear headings, short answer blocks, tables and FAQs are easier for AI systems to lift.
Measure with Bing Webmaster Tools AI Performance, Google Search Console and the utm_source=chatgpt.com parameter.

What is Generative Engine Optimisation?

The term comes from a 2023 research paper, GEO: Generative Engine Optimization, by Pranjal Aggarwal and co-authors, published at the KDD 2024 conference. It describes how content creators can improve how often and how prominently their content appears inside answers written by large language model search tools.

Classic SEO aims for a ranked blue link. GEO aims for a mention or a citation inside a generated answer. The two overlap heavily, because most AI assistants retrieve live web pages before they write, and they tend to retrieve pages that already rank and load well.

How AI assistants pick sources

The broad pattern is the same across platforms. The assistant turns a question into one or more searches, fetches candidate pages, extracts the passages that answer the question and writes a response with links. Google calls its version of this “query fan-out” and says AI Overviews and AI Mode may issue multiple related searches across subtopics to find supporting pages, according to its AI features documentation.

Two things follow. Your page must be retrievable by that platform, and the right passage on your page must be easy to extract and safe to quote.

Which crawlers need access for each AI platform?

Many sites blocked AI bots in 2023 and 2024 to stop model training, then wondered why they never appear in AI answers. Most platforms now separate training crawlers from search crawlers, so you can make a deliberate choice.

Platform	Crawler that affects citations	Training crawler (separate)	What the docs say
ChatGPT search	OAI-SearchBot	GPTBot	Sites that block OAI-SearchBot will not appear in ChatGPT search answers. Changes take about 24 hours to apply.
Perplexity	PerplexityBot	Not used for training, per Perplexity	PerplexityBot surfaces and links sites in Perplexity results and respects robots.txt.
Google AI Overviews and AI Mode	Googlebot	Google-Extended (Gemini training)	Pages must be indexed and eligible for a snippet. Google-Extended does not affect inclusion in Google Search.
Microsoft Copilot and Bing AI answers	Bingbot	Not applicable	Citations are reported in Bing Webmaster Tools AI Performance.

Sources: OpenAI crawler overview, Perplexity crawler documentation and Google’s list of common crawlers. Check robots.txt, and also any CDN or firewall bot rules, because a block at the CDN level has the same effect.

What does the research say gets content cited?

The GEO paper tested several content changes on a benchmark of queries and measured how visible a source became inside generated answers. Its authors report that the best methods improved visibility by up to 40%, with results varying by subject area.

The methods that performed best were adding relevant statistics, adding quotations from credible sources and citing sources. Keyword stuffing, a classic SEO habit, performed worse than leaving the content alone. The practical reading is simple: AI systems favour passages that contain specific, attributable facts.

Microsoft’s Bing team gave similar guidance in an October 2025 post on optimising for AI search answers. It recommends clear H2 and H3 headings, Q&A formats, comparison tables, JSON-LD schema and claims anchored in measurable facts. It also warns that AI systems may not render content hidden in tabs or expandable menus, and that key details should not live only in images or PDFs.

How do you build pages AI assistants want to cite?

Lead with the answer

Put a 40 to 70 word direct answer near the top of each page that targets a question. Then expand. Assistants often lift a single passage, so that passage should make sense on its own without the paragraph above it.

Use question-style headings

Phrase H2s the way buyers ask: “How much does a Shopify store cost in India?” rather than “Pricing considerations”. This mirrors how assistants break a prompt into sub-questions.

Add facts that can be attributed

Use your own numbers where you have them: delivery times, warranty terms, sample sizes from your customer surveys, prices with the date. Cite outside data with a link to the original source. Vague claims such as “industry-leading” give an AI system nothing to quote.

Use tables for comparisons and specifications

Comparison and specification queries are common in AI search. A plain HTML table with clear column headers is easier to parse than a styled graphic.

Make authorship and freshness visible

Show the author, the publish date and the last updated date. Update pages that contain prices, regulations or product details, and say what changed.

Keep structured data honest

Organization, Product, Article and FAQ markup in JSON-LD helps machines read your pages. Google’s AI features documentation says your structured data should match the visible text on the page.

Why does off-site presence matter for GEO?

When someone asks for “the best D2C eyewear brands in India” or “a good Shopify agency in Mumbai”, the assistant rarely relies on one brand’s own website. It retrieves listicles, reviews, directories, forum threads and news articles, then summarises the consensus.

That makes digital PR and reputation part of GEO:

Keep your brand name, category, location and key facts identical across your website, Google Business Profile, LinkedIn, marketplaces and directories.
Earn coverage in publications and comparison articles your buyers actually read.
Answer genuine questions in communities where your category is discussed, under your real name.
Collect reviews on platforms relevant to your market, and respond to them.
Keep Merchant Center and Business Profile data current, which Google lists as a best practice for its AI features.

How do you measure GEO results?

Measurement is still partial, but it is much better than a year ago.

Bing Webmaster Tools AI Performance: launched in public preview on 10 February 2026, it shows total citations, cited pages, grounding queries and page-level citation counts across Microsoft Copilot, AI summaries in Bing and select partner integrations. See Bing’s announcement.
Google Search Console: traffic from AI Overviews and AI Mode is counted in the Performance report under the Web search type.
ChatGPT referrals: OpenAI’s publishers and developers FAQ says ChatGPT adds utm_source=chatgpt.com to referral links, so you can segment that traffic in GA4.
Prompt tracking: keep a fixed list of 30 to 50 buyer prompts and check monthly whether your brand is mentioned or cited in each assistant. Record the answer, not just a yes or no.

A 90-day GEO checklist

Audit robots.txt, CDN and firewall rules for Googlebot, Bingbot, OAI-SearchBot and PerplexityBot. Decide separately on training crawlers such as GPTBot and Google-Extended.
List the 30 to 50 questions buyers ask before choosing you. Map each to one page.
Add a quick answer block, question-style headings and a FAQ section to your top 10 commercial and informational pages.
Add verifiable numbers, dates and cited sources to those pages. Remove vague superlatives.
Move key facts out of tabs, images and PDFs into plain HTML text.
Validate Organization, Product and Article schema and make sure it matches visible content.
Align brand facts across your site, Google Business Profile, LinkedIn, marketplaces and directories.
Plan two to four digital PR placements a quarter in publications your category relies on.
Verify your site in Bing Webmaster Tools and Google Search Console. Create a GA4 segment for AI referrers.
Run your prompt list monthly and log changes in mentions and citations.

Be wary of guarantees. No agency or tool can guarantee a citation in ChatGPT or Google AI Overviews. Platforms change models and retrieval methods often. Focus on the inputs you control: access, content quality, structure and reputation.

Common GEO mistakes to avoid

Blocking all AI bots with one robots.txt rule, which also blocks the search crawlers that drive citations.
Publishing large volumes of thin AI-written pages. Google’s spam policies cover content created mainly to manipulate rankings.
Writing for one assistant only. ChatGPT, Gemini, Perplexity and Copilot use different indexes and crawlers.
Treating GEO as separate from SEO. Pages that do not rank or load well are rarely retrieved.

How Squarebase can help

Squarebase runs SEO and GEO as one programme: crawler and technical audits, answer-first content, schema and citation tracking across Google, Bing and ChatGPT. See our search engine optimisation services or talk to our team about a GEO audit for your brand.

Frequently asked questions

What is Generative Engine Optimisation (GEO)?

Generative Engine Optimisation is the practice of making your content and brand easy for AI assistants such as ChatGPT, Gemini, Perplexity and Google AI Overviews to retrieve, trust and cite in their answers. The term comes from a research paper published at KDD 2024.

Is GEO different from SEO?

GEO builds on SEO rather than replacing it. Most AI assistants retrieve live web pages before answering, so crawlability, indexing and rankings still matter. GEO adds focus on extractable answers, attributable facts, consistent brand information across the web and tracking citations instead of only rankings.

Do I need special files or markup to appear in Google AI Overviews?

No. Google states that there are no additional requirements or special optimisations needed to appear in AI Overviews or AI Mode. Your page must be indexed and eligible to show a snippet, and standard SEO best practices apply, including structured data that matches visible content.

Should I block GPTBot to protect my content?

That is a business decision. OpenAI uses GPTBot for model training and OAI-SearchBot for ChatGPT search. You can block GPTBot and still allow OAI-SearchBot. If you block OAI-SearchBot, OpenAI says your site will not appear in ChatGPT search answers.

How can I track traffic from ChatGPT?

OpenAI says ChatGPT adds the parameter utm_source=chatgpt.com to referral links from its search results. In GA4 you can filter or build a segment on that source. For Microsoft Copilot and Bing AI answers, use the AI Performance report in Bing Webmaster Tools.

Sources

Written by Squarebase, Squarebase

Squarebase is a 360° marketing agency, founded in 2017, working on performance marketing, Shopify, websites and SEO, AEO and GEO for brands in India, the USA, the UK, Germany and Australia.