What Is Generative Engine Optimisation (GEO)? A Guide for SMEs

Generative engine optimisation (GEO) is the work of making your website easy for AI assistants such as ChatGPT, Gemini, Claude, Perplexity and Google’s AI Overviews to find, understand and cite when they answer a question. It builds on search engine optimisation (SEO) rather than replacing it: the page still has to be crawlable and indexable. GEO adds three things on top. The AI companies’ own crawlers must be allowed in, each page must answer one question plainly, and the site must describe your company in the same way everywhere, so an assistant can quote it with confidence.

For a clinic in Johor Bahru, a logistics firm in Singapore or a distributor in Surabaya, this matters because a growing share of customers ask an assistant before, or instead of, searching. The assistant reads a handful of sources and writes one answer. Your site is either one of those sources or it is not mentioned.

Where the term comes from

The term comes from a research paper. “GEO: Generative Engine Optimization”, by researchers from Princeton and partners, was first published on arXiv in November 2023 and accepted at KDD 2024 (Aggarwal et al., arXiv 2311.09735). The authors built a benchmark of queries across domains and tested ways of rewriting source pages. They report that some methods, such as adding citations, quotations and statistics, raised a source’s visibility in generated answers by up to 40% in their tests. They also found that results varied by domain.

Two cautions apply. “Up to 40%” is the best case in a controlled benchmark, not a typical result on a live site. And the methods that helped are the ones a careful editor would apply anyway: say where a claim comes from, quote the exact figure, and be specific.

How GEO differs from SEO

SEO aims to rank a page in a list of links. GEO aims to have a page used and cited inside an answer. The foundations overlap almost completely, and the differences are in emphasis.

	SEO	GEO
What the user sees	A list of links	One written answer with a few cited sources
Who reads your page first	Googlebot and Bingbot	Those, plus OAI-SearchBot, Claude-SearchBot, PerplexityBot and others
What wins	Relevance, quality and links to the page	Relevance and quality, plus a passage that answers the question directly and can be quoted
Unit of competition	The page	Often a paragraph or a table within the page
How it is measured	Rankings, clicks, impressions	Citations and mentions in answers, referral visits from assistants, impressions in Search Console

Google is clear that its own AI features need nothing exotic. Its Search Central guidance says there are “no additional requirements to appear in AI Overviews or AI Mode, nor other special optimizations necessary”. A page needs to be indexed and eligible to show a snippet, and no special files or schema are required (Google Search Central: AI features and your website, retrieved 2026-10-11). For Google, good GEO is good SEO.

Other assistants run their own crawlers and their own search indexes. That is where GEO adds work SEO alone does not cover.

What makes a page easy for an AI assistant to cite

Assistants that search the web retrieve passages, then write an answer from them. Our post on document Q&A that cites its sources describes the same pattern inside a company. A passage is more likely to be used when it can stand on its own and answer the question without the paragraphs around it.

In practice, citable pages share five traits.

1. The answer comes first

Open each page, and each major section, with a direct answer in two or three sentences. Then give the detail. A service page that opens with “We are passionate about excellence” gives an assistant nothing to quote. One that opens with “We build offline-first inspection apps for oil and gas plants in Malaysia, delivered in a 4-week sprint” does.

2. One page per question

A page that tries to answer ten questions answers none of them well. Give each service, each product and each common customer question its own page with a clear heading. Link the pages to each other so both crawlers and readers can move between them.

3. Specifics instead of adjectives

Names, numbers, places, dates and steps can be quoted. Words like “leading”, “innovative” and “seamless” cannot. The GEO paper’s strongest methods were exactly this: cite sources, quote precisely and include figures. Only publish figures you can back up.

4. The same facts everywhere

An assistant has to decide who you are before it can recommend you. Use the same company name, the same description of what you do, the same service names and the same markets on your site, your LinkedIn page, your Google Business Profile and any directories you appear in. If your site says “IT solutions” and your LinkedIn page says “software house”, the assistant has two weaker signals instead of one strong one.

5. Structured data that matches the page

Structured data (JSON-LD) describes your organisation, services, articles and FAQs in a format machines read reliably. Google does not require it for AI features, and it notes that structured data should match the visible text on the page. It still helps every system that reads it to connect your company name to your services and locations.

A GEO checklist for your site

Work through these in order. The first three are technical and are usually small developer tasks. The rest are content work that continues.

Let the right crawlers in

Each AI company separates crawling for model training from crawling for search answers, and each has its own robots.txt token.

• OpenAI: GPTBot collects training data. OAI-SearchBot surfaces sites in ChatGPT search, and OpenAI states that sites opted out of it “will not be shown in ChatGPT search answers”, though they may still appear as navigational links. ChatGPT-User fetches pages when a person asks ChatGPT to, and is not used to decide what appears in search (OpenAI: Overview of OpenAI crawlers, retrieved 2026-10-11).
• Anthropic: ClaudeBot collects training data, Claude-SearchBot indexes content for search responses, and Claude-User fetches pages when a person asks Claude a question (Anthropic: Does Anthropic crawl data from the web?, retrieved 2026-10-11).
• Perplexity: PerplexityBot surfaces and links websites in Perplexity’s results and is not used to train foundation models. Perplexity-User fetches pages for a user’s question and “generally ignores robots.txt rules” (Perplexity: Perplexity crawlers, retrieved 2026-10-11).
• Google: Googlebot crawls for Search, including AI Overviews and AI Mode. The separate Google-Extended token controls use in Gemini training and grounding, not inclusion in Search (Google: Google’s common crawlers).

If you want to stay out of training but appear in answers, block the training crawlers and allow the search ones:

User-agent: GPTBot
Disallow: /

User-agent: ClaudeBot
Disallow: /

User-agent: OAI-SearchBot
Allow: /

User-agent: Claude-SearchBot
Allow: /

User-agent: PerplexityBot
Allow: /

Check your firewall and CDN too. Bot-protection settings sometimes block AI crawlers before robots.txt is ever read, and the site owner never finds out.

Make sure pages render without JavaScript

Not every AI crawler runs JavaScript the way a browser does. If your content only appears after scripts load, they may see an empty page. Open your site with JavaScript disabled, or view the page source. If the text is not there, a crawler may not see it either. Static or server-rendered pages avoid the problem.

Add structured data

At minimum, describe your organisation (name, logo, description, contact, areas served and links to your official profiles), a service entry for each service page, and BreadcrumbList for navigation. Add Article markup to blog posts. Keep every value identical to the visible text.

Publish an llms.txt (optional)

llms.txt is a plain Markdown file at the root of your site that summarises who you are and links to your most useful pages. Jeremy Howard proposed it in September 2024, and it remains an informal proposal rather than a standard. Google has said it does not use llms.txt (Search Engine Land). It is cheap to add, so we add one, as on our own site, but do not expect it to move results on its own.

Rewrite key pages answer-first

Start with the pages that earn you money: services, products, pricing and locations. For each one, write the question it answers as the H1 or the opening line, answer it in the first paragraph, then add detail, examples and an FAQ built from questions customers really ask your sales team.

Cover the questions customers ask before they buy

Make a list of the questions your sales team hears every week, such as “How much does it cost?”, “How long does it take?”, “Do you work in Malaysia?” and “What do we need to prepare?”. Give each a clear answer on your site. These are often the exact questions people now put to an assistant.

How to measure GEO

Measure before you change anything, so you can see what moves.

• Google Search Console. Google reports clicks and impressions from AI Overviews and AI Mode within the overall “Web” search type in the Performance report. They are not split out, so watch impressions for the pages you rewrote.
• Analytics referrals. Visits from assistants that link out show up as referrals from domains such as chatgpt.com, perplexity.ai and gemini.google.com. Track them as their own channel.
• Server logs. Your logs show whether OAI-SearchBot, Claude-SearchBot and PerplexityBot are reaching your pages, and how often.
• Ask the assistants. Once a month, ask ChatGPT, Gemini, Claude and Perplexity the ten questions your customers ask, with search turned on, and record whether you are mentioned and which page is cited. Answers vary between runs, so look for patterns over months rather than reacting to one result.
• Enquiries. Add “How did you hear about us?” to your contact form, with an option for “AI assistant”. It is the measure that matters most.

What GEO cannot do

GEO cannot make an assistant recommend you. Each assistant chooses its own sources, those choices change, and the same question can produce a different answer tomorrow. Be wary of anyone who promises a citation or a ranking. What GEO can do is remove the reasons you are being left out, such as a blocked crawler, a page that never answers the question or an inconsistent company description, and make your best pages easy to quote.

Where to start

If you do three things this month, make them these:

1. Check robots.txt and your firewall, and make sure the search crawlers listed above can reach your site.
2. Rewrite the opening paragraph of your five most important pages so each one answers a single question plainly.
3. Set up Search Console and an AI referral channel in analytics, so next month you can see what changed.

These are the first steps of our SEO and GEO work: an audit of what search engines and AI crawlers can reach, fixes made in your code, then pages structured around the questions your customers ask. If you want to know where your site stands, send us the address and what you want to be found for. Our technical lead will tell you what stands out first.