Insights/AI Search and GEO

AI Search and GEO: How to Get Your Business Cited by ChatGPT, Gemini and Google AI

Written by Casey MountfordPublished 7 August 2026
Share

Not all that long ago, getting found online meant one thing to most businesses: getting onto page one of Google.

That still matters (a lot), but it is no longer the whole picture. Search behaviour is changing, and so are the places people turn when they want an answer. Someone researching a service, comparing suppliers or simply trying to understand a problem might now ask ChatGPT or Gemini, use Google AI Mode, read an AI Overview or turn to Microsoft Copilot before they ever click a traditional search result.

For businesses, that creates another question alongside the usual “where do we rank?”:

When somebody asks an AI platform a question related to what we do, are we actually part of the answer?

That is where AI search visibility and Generative Engine Optimisation (GEO) come in.

There is plenty of noise around both already. Some of it is useful; some of it makes GEO sound like an entirely new marketing discipline that requires businesses to forget everything they know about SEO and start again.

Thankfully, the reality is a little more sensible.

What is Generative Engine Optimisation (GEO)?

Generative Engine Optimisation, usually shortened to GEO, is the process of improving the chances of a business, brand or piece of content being surfaced within answers created by generative search engines and AI platforms.

Depending on the platform and the query, that visibility can look quite different. Your website might be cited as a source, your company could be mentioned directly, a relevant page may appear alongside an AI-generated response or your brand might be included when somebody asks for a recommendation or comparison.

The term has academic roots rather than being something the marketing industry invented last week. The original 2023 research paper, GEO: Generative Engine Optimization, explored whether publishers could influence their visibility within generative responses and reported visibility improvements of up to 40% for some approaches within its experimental environment.

That figure needs some context, though. A July 2026 review of 45 GEO studies found that the evidence remains highly dependent on the engine, query, content and stage of the retrieval process. The review also found that generic optimisation rules don't transfer particularly reliably between platforms, while citation-focused rewrites can sometimes make retrieval worse rather than better.

So, GEO is real. The universal checklist that guarantees you a ChatGPT citation isn't.

How is AI search visibility different from traditional SEO?

Traditional search has generally presented people with a selection of results and left them to investigate. They click a website, return to Google, try another result, change the search and gradually piece together the information they need.

Generative search can do much more of that research before somebody reaches a website.

Ask an AI platform something like, “What should I look for when choosing an SEO agency for a small UK business?” and it can immediately explain the services to expect, likely warning signs, questions worth asking and differences between various approaches. Depending on the platform, the answer may also cite a comparatively small number of external sources.

The person searching hasn't disappeared, and neither has the underlying need. What has changed is the layer sitting between that person and the businesses hoping to reach them.

That means visibility now involves asking more than whether a page ranks third, fifth or tenth in Google. Businesses increasingly need to understand:

  • whether their brand is appearing for commercially relevant AI queries;
  • which pages are being used or cited as sources;
  • where competitors are appearing instead;
  • which topics AI platforms associate with the business;
  • whether information being surfaced about the company is accurate; and
  • whether that visibility ultimately contributes to traffic, branded search, enquiries or sales.

Some of those questions are new. Much of the work required to improve the answers isn't.

Does GEO replace SEO?

No. In fact, this is probably the most important thing to understand before anybody starts trying to sell you an entirely separate “AI optimisation” strategy.

Google has been fairly unequivocal about its own AI search features. Its current guidance says there are “no additional requirements” for appearing in AI Overviews or AI Mode and that established SEO best practices remain relevant.

In May 2026, Google went a step further and published specific guidance on optimising for generative AI Search. The resource, authored by Google's John Mueller, again pointed businesses back towards valuable, original content and strong SEO fundamentals rather than special GEO files or technical hacks.

OpenAI ‘s own publisher guidance says any public website can potentially appear in ChatGPT Search, but website owners should make sure OAI-SearchBot isn't blocked if they want content to be discovered, surfaced, cited and linked.

So, if your site has serious crawl problems, weak content or barely any authority, GEO isn't going to bypass those issues.

Strong organic visibility still depends on many of the same foundations:

  • technically accessible pages;
  • clear search intent;
  • genuinely useful information;
  • sensible site structure and internal linking;
  • authority beyond your own website;
  • accurate business information; and
  • content that offers something more useful than the hundred pages already saying roughly the same thing.

That's why our approach to SEO and GEO services connects technical SEO, content and authority rather than treating AI visibility as an isolated project.

And if you're still deciding where organic visibility belongs alongside faster acquisition channels, our SEO vs Paid Ads guide looks at how the two fit together commercially.

What actually improves AI search visibility?

There isn't one GEO checklist that guarantees your business will suddenly start appearing in ChatGPT, Gemini or Google AI.

What we do have is a growing amount of first-party guidance from the platforms themselves, alongside independent research into the kinds of information generative systems retrieve, trust and reference.

In practice, most of the work falls into several familiar areas.

Start with technical accessibility

Before rewriting a single paragraph for AI, make sure the platforms you're targeting can actually access the website.

For Google, important pages need to be crawlable, indexable and eligible to appear within Search. For ChatGPT, OpenAI specifically tells publishers not to block OAI-SearchBot if they want their site content included in summaries and snippets.

A sensible technical review should therefore include:

  • robots.txt rules;
  • accidental noindex directives;
  • canonicalisation;
  • internal linking;
  • broken links and redirect chains;
  • site architecture;
  • page speed and usability; and
  • whether important information is actually available in accessible page content.

It's not the exciting side of GEO (if there is one), but there isn't much value in carefully restructuring an article for AI search when the relevant crawler can't get to it.

This is also why search visibility and website build and development need to work together. Sometimes what looks like a content problem is actually a technical problem underneath.

Build content around the questions people actually ask

Keywords still matter. They help us understand demand, language, competition and commercial search intent, and they aren't suddenly obsolete because somebody can type a longer question into ChatGPT.

What conversational search gives us is more context.

Someone might type “accountant Manchester” into Google.

The same person could ask an AI assistant, We're a Manchester-based business turning over around £750,000 and we've outgrown our current accountant. What should we look for when changing firms?

Those searches are clearly related, but the second tells us far more about the person's situation, concerns and decision.

Strong AI search content should account for the questions surrounding the core keyword: what somebody wants to compare, what they are unsure about, which objections they might have, what information they need before taking the next step and what a genuinely useful answer looks like.

That doesn't mean turning every page into 30 FAQs (please don't). It means covering a topic well enough to help somebody understand it.

It's also why our approach to content marketing starts with what an audience is searching for and what they need to understand during a buying decision, rather than simply filling a monthly content calendar.

Give AI systems something worth citing

This is where a lot of business content falls down.

There is an enormous amount of technically relevant copy online that contains almost no useful information.

You know the sort:

“Our experienced team provides tailored, industry-leading solutions designed around your unique requirements.”

It could belong to an accountant, solicitor, IT provider, HR consultancy, marketing agency or office furniture business. There is nothing particularly wrong with it, but there is also very little substance for a human reader (or a retrieval system) to take away.

Useful content gives people something specific to work with. That might include:

  • original research or data;
  • genuine pricing information or realistic ranges;
  • quotes from named experts;
  • detailed case studies and measurable outcomes;
  • examples from real client work;
  • first-hand experience;
  • properly sourced statistics;
  • meaningful comparisons; or
  • situations where a product or service genuinely isn't the right option.

Google's latest guidance specifically encourages publishers to create valuable, unique and non-commodity content for generative Search.

That fits neatly with what good content should have been doing long before GEO became a search term.

If twenty businesses have published essentially the same article and yours contains first-hand expertise, named contributors, useful examples and credible evidence, you've given the reader more reason to trust it. You've also given search and retrieval systems something more substantial to work with.

(Which is why uploading 100 generic AI-generated blogs isn't quite the GEO masterstroke some LinkedIn posts would have you believe.)

Make expertise obvious

If you're publishing specialist advice, make it clear where that expertise comes from.

Who wrote it? What relevant experience do they have? Are important claims supported? Can readers follow the source and check the evidence themselves? Does the page make it obvious which organisation is responsible for the information?

That doesn't mean forcing a giant author biography into every article. It does mean avoiding anonymous, generic content that could have been written by absolutely anybody.

Expertise needs to be real before we worry about making it machine-readable.

Build authority away from your own website

Your website is only one part of the information available about your business.

Industry publications, directories, reviews, professional memberships, partner websites, news coverage and credible third-party mentions all contribute to the wider picture of who you are and what you're known for.

This becomes particularly important when somebody asks an AI platform for a recommendation rather than a factual explanation. Your own website saying you're excellent is one source. Independent organisations, publications and customers saying positive things about you provide a different level of corroboration.

Microsoft's Jordi Ribas, Corporate Vice President for Search & AI, described grounding in February 2026 as becoming “a foundational layer of the AI ecosystem” because it connects generative models with external information.

That is one reason Digital PR and link building remain highly relevant. The value isn't simply another backlink on a spreadsheet. Genuine coverage gives search systems, AI platforms and potential customers independent evidence connecting a business with its expertise, services and reputation.

For local businesses, the same logic applies to accurate listings, reviews and Google Business Profile information. Consistency isn't exciting, but being accurately represented across the web makes your business considerably easier to verify.

Don't forget the person reading it

It is surprisingly easy to spend so much time thinking about AI readability that you make a page worse for actual people.

We wouldn't recommend that.

The purpose of search visibility is still to connect your business with a person who has a question, problem or need. If they eventually arrive on a page that is awkward, repetitive, stuffed with keywords or written like an encyclopaedia entry purely because somebody told you an LLM would prefer it, you've won the citation and lost the customer.

Clear navigation, useful headings and sensible formatting help machines, but they also help people find what they need. Strong website copy still needs to explain, reassure and persuade.

If you're getting visibility but people aren't taking the next step, our guide to why websites don't convert covers some of the other issues worth looking at.

What about structured data and schema?

Structured data is another area where some GEO advice gets ahead of the evidence.

Schema is useful. Google's structured data guidance explains that it gives Search explicit clues about the meaning and classification of information on a page.

But there isn't a special “GEO schema” that unlocks AI citations.

Google specifically states that no special schema.org structured data is required for its generative AI Search features. Existing structured data should be valid, accurate and match the content visible on the page, just as it should for traditional Search.

So yes, implement appropriate schema where it genuinely helps search engines understand the page.

Just don't expect adding FAQ markup on Friday to make Gemini fall in love with you by Monday.

What does good GEO content look like in practice?

Imagine an HR consultancy wants visibility around the question:

“How much does outsourced HR cost for a small business?”

A weak answer might say:

Costs vary depending on your requirements. Every organisation is different, so contact our team for a bespoke quotation.

Technically, it answers the question.

Barely.

A genuinely useful page would explain the different pricing models available, provide realistic ranges where possible, show what affects those costs and perhaps work through an example for a typical 20-person company. It could explain what tends to be included, what usually costs extra and when outsourcing might actually be less appropriate than employing somebody internally.

Only once the reader understands the subject does the consultancy need to explain how its own service works.

The second version is more useful because it helps somebody make a decision. It's also far richer for a retrieval system because there are actual facts, relationships and explanations it can draw from.

That's the mindset we prefer when creating SEO-led content.

Not “how do we make this paragraph appeal to ChatGPT?”

How do we make this one of the most useful pages available on the subject?

Can you measure AI search visibility?

Increasingly, yes.

Measurement was one of the biggest weaknesses in the early GEO conversation. Traditional SEO has familiar metrics such as impressions, rankings, clicks, organic sessions and conversions, whereas AI-generated answers were much harder to monitor consistently.

That picture is already changing.

Microsoft's AI Performance report in Bing Webmaster Tools now shows website owners which pages are being cited across supported AI experiences, how citation activity changes over time and which grounding-query themes are associated with the content. Microsoft is also careful to point out that citations do not represent rankings, authority or clicks (an important distinction).

Google began rolling out dedicated Generative AI performance reports in Search Console in June 2026. The reports show impressions within AI Overviews and AI Mode, along with the pages, countries, devices and dates associated with that visibility, although the rollout initially covers only a subset of sites.

OpenAI also confirms that websites allowing OAI-SearchBot can track referral traffic from ChatGPT through analytics platforms such as Google Analytics.

Alongside those first-party signals, there is still value in manually checking the sorts of questions that actually matter to your business and monitoring whether competitors are repeatedly appearing where you aren't.

A sensible AI visibility review might therefore look at:

  • the commercially important topics you want to own;
  • whether your brand appears around those topics;
  • which pages earn citations or visibility;
  • which competitors appear consistently;
  • referral traffic from AI platforms;
  • changes in branded search demand;
  • whether the information shown about your business is accurate; and
  • whether that visibility contributes to meaningful commercial activity.

The word commercially matters there.

Being cited hundreds of times for an informational topic that never creates a customer isn't automatically better than appearing a handful of times when somebody is actively comparing providers.

Visibility should still earn its keep.

What we wouldn't recommend doing for GEO

As with almost every fast-growing area of marketing, GEO has already produced plenty of shortcuts, hacks and supposed secret formulas.

We'd be wary of strategies built around:

  • hundreds of thin pages targeting individual AI prompts;
  • aggressively shortening good content because “LLMs prefer chunks”;
  • publishing large volumes of generic AI-generated articles;
  • stuffing ChatGPT, AI or GEO terminology onto unrelated pages;
  • treating schema as a standalone AI optimisation strategy;
  • manufacturing recommendations or fake conversations on third-party communities; or
  • promising guaranteed citations.

Google's own 2026 guidance actively challenges several of the myths developing around GEO, including the idea that websites need special AI files or machine-specific content formats. Its advice remains remarkably familiar: create useful, distinctive content and make sure the technical fundamentals work.

The latest research points towards the same need for caution. The 2026 GEO review found no single technique with proven, stable, long-term effects on organic AI discoverability across platforms.

By all means adapt your strategy for AI search. Just don't destroy a useful website in the process.

Frequently asked questions

What does GEO mean in marketing?

GEO stands for Generative Engine Optimisation. It describes the work involved in improving the visibility of a website, business or brand within AI-generated answers and search experiences, including ChatGPT, Google AI, Gemini and Microsoft Copilot.

What is the difference between SEO and GEO?

SEO traditionally focuses on improving visibility within organic search results. GEO broadens that focus to include citations, references and brand visibility inside generative search experiences. There is significant overlap between the two, particularly around technical accessibility, content quality, search intent and authority.

Can you optimise a website for ChatGPT?

There is no guaranteed formula for appearing in ChatGPT Search. OpenAI does recommend allowing OAI-SearchBot to access public website content if publishers want that information to be discovered, surfaced and cited. Creating useful, relevant and well-supported content also gives retrieval systems better information to work with.

Do I need separate SEO and GEO strategies?

For most businesses, not completely. GEO is better treated as an extension of wider organic search, content and brand activity because many of the underlying requirements overlap. There may be new tools, measurements and research methods involved, but that doesn't mean abandoning established SEO principles.

Can I track AI search visibility?

Yes, although measurement remains fragmented between platforms. Bing Webmaster Tools now provides dedicated AI citation data, Google has started rolling out AI visibility reporting within Search Console, and referral visits from ChatGPT can be monitored through website analytics.

Can GEO guarantee that my business appears in AI answers?

No. Generative answers vary by platform, query, model, context and user. GEO can improve the accessibility, relevance and authority of information about a business, but nobody can credibly guarantee that a particular platform will cite or recommend it.

Author

Casey Mountford

Marketing Assistant & Junior Copywriter · Casey works across content creation, email campaigns, social media, Meta ads and on-site SEO, bringing a sharp eye and proper get-stuck-in energy to every project.

Want this thinking applied to your business?

Reading is useful. Acting on it is better.

If something in our Insights section has made you think about your own setup, that's a good sign. Book a free 30-minute strategy call and let's talk about what the right approach looks like specifically for your business.