The marketer’s guide to GEO: how to get your brand discovered in AI search

Thought Leadership
Cristina Forlani We Are Social

Generative Engine Optimisation (GEO) is about improving how a brand is understood, referenced and recommended by generative AI. As people increasingly use tools like ChatGPT and Gemini to research brands and products, GEO is becoming an important part of brand discovery. Here’s what marketers need to know, and what they can actually do about it.

Your next customer might know quite a lot about your brand before they ever visit your website.

They might have asked ChatGPT to compare you with three competitors, asked Gemini which product best suits their needs, followed up with a question about the downsides and eventually asked which one it would recommend. By the time they reach you, a meaningful part of the research journey may already have happened somewhere you can’t see.

This is moving into the mainstream quickly. More than one billion people globally now use standalone generative AI tools every month, according to our Digital 2026 research with Meltwater, and Adobe found 25% of customers now cite an AI platform as a primary research tool.

At the same time, social has become a major discovery engine in its own right. 59.6% of Australians use social for brand discovery, almost level with the 60.3% using traditional search.

People are discovering brands across search, social and AI, and those systems are increasingly feeding each other.

Our Sydney team recently brought together Jeremy Cabral, Founder of Ark Advisory and former COO of Finder, Sarah Fuller, Senior Agency Lead at LinkedIn, our APAC CEO Suzie Shaw and Chief Strategy Officer Richard Parker to unpack what that means for marketers.

GEO: what marketers need to know

GEO goes beyond optimising your website. It’s about the wider body of information that helps AI understand who you are, what you know and whether you deserve to be part of an answer.

AI is also changing the research journey. Comparison and shortlisting can increasingly happen before somebody reaches your website, which makes the information available about your brand elsewhere more important.

That includes social content, creators, employees, experts, publishers, reviews and communities. And while the temptation might be to respond by producing industrial quantities of “AI-friendly” content, the better opportunity is to create useful information and credible evidence that genuinely adds something.

Which brings us to the acronym everyone is talking about.

What is GEO, or Generative Engine Optimisation?

Generative Engine Optimisation (GEO) is the practice of improving a brand’s visibility, authority and representation in generative AI answers.

In practical terms, it means making it easier for AI systems to understand what your brand is, what it knows and does and when it should be referenced, cited or recommended.

The important bit is where that understanding comes from. Your website is part of it, but AI can also draw on publishers, social content, creators, experts, employees, reviews, forums, communities and other third-party sources. So while GEO has grown out of the search conversation, it very quickly becomes a much broader brand and marketing question.

What’s the difference between GEO, AEO and SEO?

Here’s the simplest way we think about them.

SEO (Search Engine Optimisation) helps websites and content rank in traditional search results, primarily to generate visibility and clicks.

AEO (Answer Engine Optimisation) helps content provide a clear, extractable answer to a specific question, including through answer boxes, AI Overviews and other answer experiences.

GEO (Generative Engine Optimisation) builds the wider visibility and authority that helps generative AI understand, reference, cite and recommend a brand when synthesising information from multiple sources.

Why does GEO matter for brands?

Here’s the slightly uncomfortable bit for marketers: a customer can now do a lot of research about your brand without you knowing they’re there.

They can ask ChatGPT for the best options in a category, compare you with a competitor, interrogate the differences, ask what people complain about and get a recommendation, all before Google Analytics has registered so much as a page view.

For years, we’ve built marketing journeys around a series of relatively observable steps: search, click, website, compare, convert. AI can squash a surprising amount of that into one conversation.

The click is changing too. Pew Research found that users clicked a traditional Google result in around 15% of searches without an AI Overview, falling to 8% when one appeared. At the same time, Adobe found that people arriving from AI during its holiday analysis converted 31% better than traffic from other sources.

Put those things together and you get a rather different customer journey. Fewer people may arrive on your website, but those who do could arrive knowing far more and being much closer to a decision. Others may have already ruled you in or out before they get anywhere near it.

That changes what visibility means. Ranking and earning the click still matter, but marketers also need to think about whether AI knows enough about their brand to include it when customers are researching, comparing and narrowing down their options. That broader challenge is where GEO comes in.

Sarah Fuller from LinkedIn made a related point during the panel: if a brand isn’t showing up in AI engines, it risks not making the shortlist at all. In B2B especially, where much of the buying journey happens before someone speaks to sales, the visibility and credibility of individual leaders and experts can become commercially significant.

What sources influence AI answers?

This slide summed this up rather neatly:

Obviously brands have never had complete control over their reputation, but AI makes the wider information around a brand more important than ever, because it can actively pull those signals together when forming an answer.

A publisher might establish authority. A creator can demonstrate experience. An employee or executive can show expertise. Reviews and communities can provide the messy, useful reality of what people actually think.

There is already some interesting evidence here. Meltwater analysed 9.5 million AI citations across 16 B2B categories and found 47.5% came from LinkedIn, Reddit and YouTube, compared with 18.7% from company websites.

Within its LinkedIn research, 75% of citations came from individual profiles rather than company pages and 51% came from people with fewer than 10,000 followers.

Those numbers aren’t a formula for winning GEO. Citation behaviour varies significantly by model, category and prompt, and will keep changing. What they do tell us is that your brand’s AI footprint doesn’t stop at your website. Reputation, advocacy, creator strategy, thought leadership, PR and community are all potentially part of the picture.

How does social media influence GEO?

We spend a lot of time thinking about what social content does in the feed. Did someone stop? Watch? Share? Comment? Click? Buy? None of that goes away.

But one of our favourite examples from the panel suggests some social content may keep working long after the engagement graph has flattened out. Jeremy shared a story from Finder about an Instagram Reel built around original survey data. Internally, it wasn’t considered an especially important piece of content. Later, it started appearing as a source around commercially valuable AI prompts, because the Reel contained original information that was useful when AI was trying to answer a question.

For marketers, that adds another lens to the role of social. Alongside asking whether something will earn attention, we can ask whether it contains useful knowledge, genuine expertise or something distinctive enough to be worth referencing later.

That doesn’t mean turning every Reel into a research paper. [Please don’t!] It means recognising that social can contribute to a much bigger information footprint around the brand, and that a good post may have a useful life beyond the week it was published.

Sarah framed this as a question of lifetime value. Social content has traditionally been judged over a fairly short campaign window, but GEO gives some of it a much longer tail. A useful post, expert point of view or piece of original data may continue contributing to discovery well after the engagement curve has flattened.

Authority is built with evidence

One of Jeremy’s most useful ideas during the discussion was “manufacturing evidence”: making sure enough credible information exists for an AI system to understand what an organisation knows, does or can legitimately claim.

We think there are 3 parts to that:

Put those together and you’re giving both people and machines a much richer body of information to work with.

The goal isn’t to cover every conceivable query with another piece of content. It’s to work out where the important information gaps are and fill the ones your brand has a credible right to fill.

And useful information doesn’t always live in the obvious places. Jeremy pointed to financial services guides buried in website footers: hardly anyone’s hero campaign asset, but their clear, consistent structure can make them surprisingly useful to AI trying to establish what a company actually offers. Turns out the least glamorous pages on your website may finally be having their moment.

Owned, earned and social need to work together

One month LinkedIn is climbing the citation charts. Another study finds Reddit everywhere. Different models favour different sources. Platform access changes regularly. By the time someone publishes the definitive GEO playbook, there’s a decent chance part of it will already be wrong.

This is why we don’t think chasing individual citation patterns is much of a strategy. Research showed material differences between ChatGPT, Claude, Gemini, Perplexity, Grok, Copilot and Google’s AI products, and source behaviour also changes by category and prompt.

A more durable approach is to think about how useful information can travel. Original research might live on your website, fuel social content, give executives something interesting to talk about, provide creators or experts with useful material and generate earned coverage. Video can become a transcript. A good panel discussion can become a guide. [See what we did there?]

None of those assets needs to exist purely for GEO. They should do useful marketing jobs in their own right. Together, though, they create a richer and more consistent information footprint around the brand.

How can brands improve their visibility in AI search?

The technology is moving quickly enough that waiting for the definitive playbook is tempting. Unfortunately, we could be waiting a while. A sensible place to start is with the questions customers actually ask.

1. Understand the questions that matter

Look beyond the obvious category question and map the research, comparison and refinement journey around it.

2. Audit your current AI visibility

Run those questions across the platforms your customers are likely to use. Look at whether you appear, how you’re described, which competitors appear and which sources underpin the answers.

3. Find the information gaps

What information is missing, inaccessible, inconsistent or simply absent? Where do competitors have more credible evidence than you?

4. Create evidence worth referencing

Invest in original research, proprietary insight, expertise, useful product information and material that genuinely adds something new.

5. Build authority beyond your own channels

Think about the creators, experts, employees, executives, publishers and communities that can credibly contribute to the areas you want to be known for.

6. Make social work harder

Keep creating work people choose to watch, share and talk about. But look for opportunities to build useful information, genuine expertise and original evidence into that work too.

7. Keep testing

Models change. Sources change. Customer behaviour changes. Treat GEO as an ongoing learning programme rather than a one-off optimisation project.

How should marketers measure GEO?

Referral traffic is useful, but it can’t tell the whole story. If someone spends 15 minutes asking an AI tool about a category, repeatedly encounters your brand and eventually visits you directly, conventional attribution is unlikely to capture much of what happened.

We’d look at a combination of visibility across commercially relevant prompts, representation and whether the information about the brand is accurate, competitive presence, the sources influencing answers and the information gaps where credible evidence is missing.

Then connect those signals back to the commercial measures you already care about, including branded search, direct traffic, qualified demand, leads and sales.

The objective is to understand whether your brand is becoming easier for AI to find, understand and confidently include when customers are making decisions.

Don’t optimise the humanity out of marketing

GEO naturally pulls the conversation towards utility: clear information, structured answers, comparisons, evidence and lots of lovely bottom-of-funnel intent.

Marketing has spent enough decades learning that people aren’t quite so tidy and rational. We choose brands because they’re familiar, because somebody we trust recommended them, because they say something about us, because everyone seems to have one or because we simply fancy the thing.

AI getting better at comparing functional attributes doesn’t make any of that disappear. It may actually make brand preference more important as price, specifications and availability become easier for machines to compare.

So GEO shouldn’t become an excuse to optimise all the humanity out of marketing. Brands need enough useful information and credible evidence to be understood by machines, while continuing to build the relevance, reputation and desire that make people want to choose them.

How visible is your brand in AI?

AI is already researching, comparing and recommending brands in your category.

We Are Social can help you understand how your brand is showing up, where the gaps are, and what to do next, from GEO strategy and visibility assessment through to social, creator and content activation, testing and measurement.

Talk to us about building your GEO programme.

GEO FAQs

What is GEO in marketing?

GEO, or Generative Engine Optimisation, is the practice of improving how a brand is understood, referenced, cited and recommended by generative AI systems. It considers the wider information ecosystem around a brand, including owned websites, social content, creators, experts, publishers, communities, reviews and earned media.

What is the difference between GEO and SEO?

SEO primarily helps webpages rank in traditional search results and generate clicks. GEO considers how generative AI understands and recommends a brand using information from many sources, including websites, social platforms, publishers, creators, experts and communities.

What is the difference between GEO and AEO?

AEO, or Answer Engine Optimisation, focuses on making information easy for an answer engine to extract when answering a specific question. GEO is broader, focusing on the overall authority and information ecosystem that helps generative AI understand, reference and recommend a brand.

Does social media affect GEO?

Yes, social content can contribute to the information ecosystem used by AI systems. Research reviewed by We Are Social has found platforms including YouTube, Reddit, LinkedIn and Facebook appearing among AI citation sources, although their importance varies significantly by model, category and prompt.

Can LinkedIn, Reddit and YouTube influence AI search results?

They can. Different AI systems cite social platforms to different degrees and citation patterns change over time. The important question isn’t simply which platform receives the most citations overall, but which credible sources influence the prompts and decisions that matter in your category.

How do you optimise a brand for ChatGPT and other AI tools?

Start by identifying the prompts customers use when researching your category. Audit how your brand and competitors appear, identify the sources influencing those answers and find information gaps. Then strengthen the available evidence through useful owned content, original research, social, creators, experts, earned media and other credible third-party sources.

How can I check whether my brand appears in AI search?

Create a representative set of customer questions and refinement prompts, then test them across the AI platforms relevant to your audience. Track whether the brand appears, how it is described, which competitors are recommended and which sources are cited. Repeat the exercise over time because results can change.

What makes content more likely to be useful to AI?

There is no guaranteed citation formula. However, original information, clear structure, accessible factual information, demonstrated expertise and credible corroboration give AI systems stronger evidence to work with. The aim should be to create genuinely useful information rather than content designed purely to manipulate an AI result.

How do you measure GEO?

Useful GEO measures include visibility across relevant prompts, competitive presence, accuracy of brand representation, citation and source patterns, information gaps and downstream commercial signals. Referral traffic alone is unlikely to capture AI’s full influence on a customer’s decision.

Who should own GEO inside a business?

GEO is likely to require collaboration across brand, search, social, content, communications, PR, technology and sometimes legal. AI systems can build their understanding from information across an entire organisation, so GEO is difficult for any single specialist team to solve alone.