One of the biggest changes taking place in the advertising ecosystem as a result of AI is undoubtedly the consumer shift towards using AI tools over search browsers to discover products. In this guest byline, Sean Betts, Chief AI Officer at Omnicom Group UK marks the shift from search engine optimisation (SEO) to generative engine optimisation (GEO), and explains why the two aren’t the same.
For the last twenty years, digital marketing has been built around the click. Brands created content, optimised pages, bought search ads, improved rankings and measured whether people arrived on their websites.
This ecosystem still matters. Search is not disappearing. Websites are not becoming irrelevant. But a new layer has emerged between brands and consumers. People are increasingly asking AI platforms to summarise choices, compare products, explain trade-offs and recommend what to do next.
This changes the marketing challenge for brands from “Can people find us?” to “Can AI platforms understand us and recommend us?”
That is where GEO, or generative engine optimisation, comes in. But the phrase needs careful handling. GEO is not SEO for AI. It is not a new way to stuff keywords into pages or game AI platforms. It is the practice of improving how a brand, product or service is understood, cited and recommended by GenAI platforms.
This is important because AI platforms do not behave like traditional search engines. Search engines rank links. AI platforms interpret intent, retrieve information, synthesise sources and generate answers. They do not just point people towards the web. Increasingly, they become the interface through which people experience it.
This means brands need to think differently about how visibility is created.
The first difference is that AI platforms search differently. A traditional search query is often short and transactional. A prompt is usually longer, more conversational and more context-rich. Someone may not search “best running shoes”. They might ask, “What are the best running shoes for a beginner with knee pain who mostly runs on pavements and does not want to spend more than £120?”
This is a very different kind of intent. It means brands are not only competing for broad category visibility but are now competing to be relevant in much more specific contexts.
The second difference is that AI platforms see websites differently. Humans are very visual and prefer good design, imagery, video, animation and interactivity. AI platforms mostly parse text and structure. A website can look beautiful to a person and still be very difficult for an AI platform to understand. If important content is hidden behind JavaScript, blocked from crawlers, poorly labelled or expressed mainly through imagery, it may be effectively invisible to the AI platforms now mediating consumer discovery.
The third difference is that authority is broader than the brand’s own website. AI platforms draw from a wider ecosystem: brand sites, publishers, reviews, comparison pages, Reddit, YouTube, blogs, ecommerce platforms and expert sources. The exact mix varies by category and platform, but the principle is consistent. Brands are judged not only by what they say about themselves, but by how the wider web explains, compares and discusses them.
What Should Marketers Do Now?
First, make your own site machine-readable. This is the most practical place to start because it is within your control. Audit robots.txt and crawler access. Check whether key content is visible before JavaScript loads. Use clear, semantic URLs that describe what pages are about. Add structured data through schema.org where appropriate. Create answer-first content that maps to the questions people ask in your category. Build FAQs that explain products, services, claims, limitations and comparisons clearly.
This is about making your brand intelligible to machines. Accessibility matters here too. Strong accessibility foundations, clear headings, descriptive labels and well-structured content help people, but they also help AI platforms understand what a page contains.
Second, build authority beyond your own website. Marketers need to understand which sources AI platforms rely on in their category. That means auditing AI answers, identifying the pages and platforms that are being cited, and mapping the sources that shape how the category is described.
For some categories, official brand sites and product pages will matter most. For others, review sites, expert blogs, publishers, YouTube videos or Reddit discussions may carry more influence. This requires a broader view of digital reputation. Keeping comparison content fresh, supporting credible third-party coverage, answering real customer questions and understanding community conversations all become part of GEO.
This also changes how brands should think about content. Scale is not always the winning factor. Relevance often matters more. A niche YouTube video or detailed blog post that directly answers a specific question may be more useful to an AI platform than a polished but generic brand campaign page.
Third, start preparing for AI-mediated commerce. AI platforms are moving closer to product discovery, recommendation and transaction. That makes structured product data increasingly important. Brands and retailers should make sure product information is complete, accurate and current: pricing, availability, specifications, ratings, reviews, imagery, support information, store locations and FAQs.
The point of this is to help AI platforms understand when a product is relevant, who it is suitable for, how it compares with alternatives and why it should be recommended.
This approach is simple, but it needs a wide variety of specialists. Understand how people prompt in your category. Audit how your brand currently appears. Identify which sources shape those answers. Improve the onsite, offsite and commerce signals AI platforms can read.
Then measure, test and repeat.
In this environment, visibility will not belong only to the brands that rank highest. It will belong to the brands that are easiest for AI platforms to understand, verify, contextualise and recommend.
