AI-based search results are challenging marketers’ status quo, disrupting organic search and changing the way in which consumers discover information.

Essity conducted some analysis of its global data across several brand categories in the 150 countries it operates in, finding that its organic search traffic had declined by 35%.

Keiko Hikichi, global brand strategy and performance manager, Essity, said: “After running some pilots, we concluded that mental availability is now shaped within AI answers and that we had to take into account another metric.

“AI is now the goalkeeper for mental availability. Citation is the new metric. If you are not cited, basically you are invisible. And it doesn’t matter how long you spent building brand, and it doesn’t matter how strong you think your brand is out there. If your brand is not machine-readable, you are invisible. In the moment of truth, when people are searching for solutions, that’s when the bottle neck starts – if you are not part of that conversation, you are not there in the solutions.”

To look into the issue further, Essity looked at how products within its Tempo toilet tissue range were performing in AI search results across various AI models in Germany.

“According to us, as brand managers, we believed that eco-friendliness and sensitive to skin were two of the most relevant product attributes we wanted to communicate. It turned out that LLMs thought otherwise.”

The analysis found that there was a gap in performance due to AI results focusing on ‘scent’. “We didn’t use the word ‘scent’, we used ‘fragrance’ – and because of this, we were losing visibility,” said Hikichi.

She cautioned that marketers now need to consider how LLMs work, beyond their traditional insight work. “We tend to believe that what we’re doing is correct, based on human understanding. But the other side of the coin is to have a good understanding of how LLMs have been trained.”

Essity has optimised its AI search across four pillars: omnipresence (being visible), context-rich content (whether content has high authority within online communities, e.g. Reddit); being machine-readable; and trust (signal amplification).

Speaking in the same session, Jodie Gillary, head of brand activation at Kantar, said: “Search isn’t dead, it has just been diverted to AI tools, as people ask longer, more complex questions about their needs. What we’re seeing is that over 80% of Google searches in travel results in an AI overview. Even when people are not actively engaging with gen AI, they are forced to look at it as the result of a question.”

This shift is creating new “jobs to be done”, said Gillary. “This is not just for SEO or content teams, it’s a marketing job, because there is a new consumer in town – the machine. Before you can even get to a human, you’ve got to be chosen by the LLM and you’ve got to get them to summarise how you want your brand to be portrayed.”

Gillary cautioned that brand size or fame doesn’t translate to performing well in AI search, with Kantar analysis on airlines finding that Qatar and Emirates – small brands, relative to competitors – were over-indexing on AI results.

Machines and humans also interpret brand relevance differently – what will differentiate a brand in a person’s mind will differ from how an LLM interprets difference. “Humans and machines do not see ‘meaningful difference’ in the same way in terms of brands,” said Gillary. “If you are a brand that has a more emotive proposition, you might struggle, because the machines like tangible, functional difference – they don’t understand emotion or empathy.”