Ever wondered why ChatGPT suggests specific brands when you ask for a product recommendation?
Until recently, searching for answers followed a familiar routine: open Google, scan a list of blue links, and pick the most credible source. Today, millions skip that process entirely and ask AI directly. A question such as “What’s the best cooking cream for pasta?” returns a curated explanation – often including product recommendations, comparisons, and even usage ideas. The interesting part is not the answer itself, but the invisible process behind it: why these products, why these sources, and why this tone of certainty?
The more people rely on AI for guidance, the more brands are forced to ask a new strategic question: how does a product become part of an AI-generated answer at the exact moment a user is making a decision?
This evolution is giving rise to Generative Engine Optimisation (GEO) – an emerging discipline where visibility is earned inside conversations rather than traditional search rankings.
Search has evolved. So must marketing
AI-powered search is no longer experimental. Global surveys indicate nearly half of internet users now engage with AI chatbots for information or purchase advice. In the MENA region, adoption is significantly higher; around 74 per cent of users in the UAE and 68 per cent in Saudi Arabia report weekly usage – clearly outpacing global behaviour.
This matters because generative engines do not generate lists; they generate conclusions. Traditional SEO attempts to influence what Google ranks. GEO seeks to influence what models such as ChatGPT, Gemini and Google’s AI Overviews consider accurate, recommendable and trustworthy enough to include. In fact, research suggests that a brand’s corporate site may contribute only around 5–10 per cent of the data generative engines draw from. The rest comes from the broader digital ecosystem: publications, reviews, social communities, Wikipedia, video commentary, retail platforms and user-generated discussion spaces like Reddit.
In other words, authority is becoming decentralised. Search has stopped being a page and is becoming a conversation.
SEO: Still the starting point
GEO does not replace SEO; it builds on it. Technical foundations – speed, structure, mobile readiness, semantic tagging and high-quality copy – remain the mechanism through which generative engines interpret a brand with confidence. Without that foundation, AI simply lacks reliable material to evaluate.
Beyond that, GEO demands something beyond technical strength: external validation. A brand must exist credibly in the wider internet, not just its own properties. Generative models weigh what users say on review platforms, what journalists mention in credible publications, and how communities speak about products in forums, Reddit threads, and social channels. PR visibility, Wikipedia accuracy, structured data, and verified facts are increasingly shaping how AI understands a brand’s identity.
Retail platforms have become a critical arena for GEO. Generative engines frequently pull data from product listings, star ratings, ingredient breakdowns and user reviews. A rich e-commerce presence therefore becomes a strategic input, not just a conversion tactic.
Traditional performance measurement also changes. Instead of only tracking rankings on Google, brands will need to track how often they are mentioned inside AI-generated answers. This is called “share of AI voice” – essentially, how frequently a model chooses your brand when responding to users.
Together, these shifts turn GEO from a visibility tactic into a reputation engine – strengthening a brand’s authority across every digital touchpoint, far beyond traditional search.
How Arla Foods approaches the shift
Within Arla Foods, this transformation has prompted a forward-thinking question: how do our brands appear when generative engines make recommendations? Rather than assuming AI will eventually recognise trusted FMCG brands, Arla Foods Consumer Experiences team began assessing how products such as Puck and Lurpak surface across retail platforms, chatbot answers and community discussions. The goal is not to chase a trend, but to understand where generative engines currently gather information and how they represent our categories today.
Partnering with CARAT, we have started establishing early measurement frameworks to examine how frequently our brands are cited within generative responses, whether AI references the correct products, and which external signals contribute most to those mentions. This approach is deliberately exploratory: it is about understanding the mechanics behind AI recommendations long before those patterns become fully established.
Importantly, this phase is not a celebration of results. It is groundwork – defining baselines, learning how AI perceives our categories, and preparing strategic principles for a future where answer engines impact how families choose everyday products. In a region where AI usage already exceeds global norms, this preparation becomes more than a competitive advantage – it becomes a prerequisite for relevance.
Generative search will not fully replace traditional search, but it will change the starting point of discovery. The brands that appear confidently within AI answers tomorrow will be those building credibility, accuracy and visibility across the ecosystem today.
In a world where answers increasingly replace links, visibility becomes less about being clicked – and more about being chosen.
By Talal Alrifai, Search Lead at Arla Foods
