By Johann Wrede, Chief Marketing Officer, UserTesting
The last two years have seen the playbook for marketing change drastically. Global end consumer usage of AI apps like ChatGPT and Gemini has surged by 62% over that period, to the point that AI is fundamentally reshaping how brands must engage their customers. One IBM report states that 41% of end consumers now use AI assistants to research products, 33% to look for reviews, and 31% to search for deals, and those numbers are expected to grow.
This means the first time a customer encounters a brand may now be through an AI-generated summary description. In short this summary, more often than not, is now your representative – how it describes a brand could be make-or-break for customer engagement. Thus the brands who stand out amongst their competitors today are the ones that know how to engage with LLMs directly – not just on the brand website but consistently across all communication channels from social to earned media.
But as we’ve seen with the rise of ‘AI slop’, content that swings too far toward AI-optimisation risks diluting brand presence and disengaging customers. Brands need to find the sweet spot between efficiency and customer experience through iterative research and audience response testing.
Sitting alongside traditional SEO, brands are quickly realising that GEO (Generative Engine Optimisation) is a strong route to making them and their products more likely to appear in AI-generated answers. But targeting LLMs doesn’t require a new marketing strategy, rather a refinement of current SEO thinking.
The brands that get cited reliably can often share certain qualities: consistent messaging, keyword-led website copy, accurate product descriptions, comprehensive FAQs that pre-empt questions before customers think to ask them, strong review volume and clean schema markup. If any of that is inconsistent or hard to parse, the model will surface someone else.
Unlike SEO, where page rankings shift but tend to hold their position over time, GEO results are far more transient. LLMs constantly update and refresh what they cite, and a brand surfaced in AI summaries today may not be there in a couple of months. To this end, GEO demands regular auditing of what models are actually saying about the brand and testing whether content still earns the citation.
Historically the marketing playbook has demanded content to be set up for human consumption only – led by design, storytelling, a unique brand voice and conversion strategies. While all of those items are still crucial to stand out amongst competitors, they don’t support machine interpretation. What’s more, GEO relies heavily on third-party intel i.e. your brand is only as strong as the industry describes it to be.
All of this means there’s a fundamental gap to GEO-readiness that needs to be addressed by the marketing and data functions. Marketers must balance shaping a distinctive brand voice through consistent messaging and storytelling, while working with web and social data/analytics experts (“data teams”) to enhance LLM visibility.
Striking the balance
The risk for marketers is leaning too far towards an AI-optimised digital presence, diluting brand quality, differentiation and disengaging customers.
However, AI optimisation and brand identity aren’t necessarily in conflict. Done with the right inputs and rigorous testing, brands can build a strong AI-optimised presence without diluting their voice. The problem is that most brands are treating them as separate workstreams, with the data team owning GEO and the brand team owning everything else. The gap between the two is where quality degrades.
If we consider the traditional marketing funnel, GEO sits at the awareness stage. It gets you cited. A citation that leads somewhere thin is a dead end. Brands still have to earn the conversion, and that requires something structured data alone cannot deliver. Brand differentiation is what will engage customers and move them down the funnel.
Building for the real world
End consumers are increasingly demanding authenticity in the AI world. That’s because there is an emotional side to customer interactions. This should be every marketer’s bread and butter. Content and products need to feel meaningful, while also making the customer feel valued.
Fundamental to this is understanding how real people respond to brand experiences. In the GEO era, it feeds directly into content credibility. A brand that regularly tests content with actual customers tends to produce material AI engines are more likely to cite, because it has already proven its value with real people before it went out.
Building something worth surfacing to AI engines and to customers alike is the actual challenge and it starts quite simply with knowing what your audience thinks.
