‘Mental availability is now shaped within AI answers,’ according to Keiko Hikichi, global brand strategy and performance manager at Essity, and marketers must look at metrics and define success in ‘a very different way’.
Speaking at the Market Research Society’s Annual Conference on 10 March, Hikichi described how Essity had worked to improve the AI visibility of one of its brands in Germany. Using analysis from Kantar, Essity identified that the toilet paper was not being rated highly for its scent because its advertising generally used the term fragrance (or the German equivalent) to describe the smell of the paper. Tweaking the language boosted the brand’s AI visibility significantly, she said: ‘Just because we didn’t use the word AI was trained to search for, we were losing business.’
Jodie Gillary, head of brand activation at Kantar, explained that with searches for brands declining by ‘as much as 80% in some cases’, marketers need to adapt to a discovery environment where consumers ‘increasingly ask longer, more complex questions’.
‘53% of users fully or mostly trust the answers they get from gen AI,’ she added. ‘And over a third of people use AI to discover new brands.’
Gillary described AI as ‘the new consumer’ and advised brands to adjust their marketing to appeal to both humans and machines.
‘Before you even get to a human, you have to be chosen by a machine,’ she said. Gillary added that, according to Kantar’s findings, many large brands, like AirBnB or EasyJet, have shockingly low rates of AI mentions compared to smaller and more frequently selected brands, like Emirates or Hilton. This is not just true of travel brands — she gave the example of asking an AI chatbot for the best brand of tea and coffee. PG Tips and Tetley, two of the UK’s biggest brands, did not even make the top ten, and nor did Nescafé in coffees. Instead, the AI primarily selected luxury or newcomer products, like Fortnum and Mason, Grind, and Pact. The only massive everyday brand which was selected by the AI was Yorkshire Tea.
‘If you want to show up in AI search results you have to have a strong brand,’ she explained, clarifying that what makes a brand strong is ‘different for machines than humans’: ‘If you are a brand that has a more emotional offering, machines do not understand emotion, nuance, metaphor. They like tangible difference.’
Asked how marketers can identify what they are doing that AI isn’t picking up, Hikichi said that, while Essity uses an external research company, the process can begin in house.
‘Look at your website traffic and look at impressions,’ she said. ‘Every good marketer already knows their baseline. If you are losing “vanilla coffee”, that’s your hint that people are searching for that elsewhere.’
Once these terms have been identified, she explained, marketers can work to introduce them into more advertising copy.
Gillary added that marketing favoured by humans, such as emotion-driven or visually impactful campaigns, should be prioritised on social platforms where AI less frequently scrapes content.
‘YouTube probably does the best with LLMs,’ she said, attributing this in part to its content, which is often educational, but also to the site’s auto-generated captions which allow text-based LLMs to read the content of videos they would not otherwise be able to access.
This two-pronged approach increases the AI-appeal of a brand, putting it on people’s radar when they ask a chatbot for product recommendations, while still generating positive consumer sentiment on social media.
‘As brands, you need to consider that everything you are doing needs to be machine readable,’ Hikichi concluded.
