AI chatbots are rapidly becoming part of everyday shopping behaviour. Increasingly, shoppers are asking tools such as ChatGPT and Google’s AI assistants for product advice before making purchase decisions.
Recent estimates suggest that around 22.5 million people in the UK are already using large language models, with Generation Z leading adoption.
For FMCG brands, this raises an important question:
What answers are shoppers actually seeing when they ask AI for product advice?
To explore this, we analysed real prompts and responses from our AI Consumer Panel. The findings highlight how AI is already influencing product discovery and why fresh information, particularly fresh reviews, is becoming increasingly important.
Before diving into the details, here are the key insights from our research:
One of the most striking findings from our research was the variety of questions shoppers ask within a single category.
Instead of simple queries such as “best yoghurt”, shoppers ask much more detailed questions:
In one study alone, we collected more than 780 real prompts and responses about yoghurt from shoppers using AI tools.
This highlights an important behavioural shift. AI queries are more conversational and exploratory than traditional search. Instead of entering keywords, shoppers ask complete questions about ingredients, health benefits, taste and dietary suitability.
For brands, this means visibility across many different questions now matters far more than ranking for a single search term.
Another major difference between AI and traditional search is the number of brands mentioned in answers.
This contrasts sharply with a traditional Google results page, which might display dozens of brands and links.
In AI-generated responses, if your brand is not included in the answer, you are effectively invisible. Understanding how those answers are generated is therefore critical.
Looking more closely at yoghurt responses, Fage appeared most frequently in the AI answers we analysed, even though it is not the overall market leader in the category.
The explanation is category simple: Many shoppers were asking specifically about Greek yoghurt or thicker yoghurt styles, categories where Fage is strongly associated.
This reveals an important fact: Topics chosed by consumers lead to brands chose by LLMS.
If the conversation focuses on Greek yoghurt, certain brands naturally appear. If the question focuses on healthy yoghurt options, the AI may recommend a type of yoghurt, such as Skyr, rather than a specific brand. In other words, AI recommendations tend to follow category logic rather than market share.
Another interesting finding from the yoghurt research was the role of retailers in AI answers. Tesco appeared frequently in responses, particularly when shoppers asked about product availability or value. AI models often rely on structured and accessible information from retailer websites when generating responses.
In many cases, retailer product pages effectively become sources of truth for how AI systems describe products and categories.
We undertook a similar analysis in another category: face cream and moisturiser.
Once again, shoppers asked a wide variety of questions, but most focused on specific skincare needs, including:
This highlights another important shift, brands are not simply competing to be the “best moisturiser.” They are competing to be the best answer to a specific need state.
In the face cream analysis, CeraVe dominated the responses, appearing in more than 40 percent of answers:
Several factors likely contribute to this visibility:
AI systems combine structured information, such as ingredients and product claims, with signals including consumer reviews and retailer presence.
Interestingly, traditional influencer marketing appeared to have very little influence on the answers generated.
Another notable outcome from the skincare analysis was the prominence of Superdrug in AI responses.
Superdrug’s website provides clear product categorisation, ingredient information and structured descriptions, making it easier for AI systems to interpret.
By contrast, some retailer websites are harder for AI systems to navigate and extract structured information from.
As a result, Superdrug appears more frequently in AI-generated explanations and recommendations.
One of the clearest conclusions from the research is that fresh information matters.
AI answers frequently include statements such as: “Highly rated by users” and “Popular among shoppers”
Recent reviews strengthen the credibility of these recommendations.
Our broader analysis of supermarket reviews shows that reviews older than six months rapidly lose their impact, while fresh reviews significantly improve shopper trust and conversion.
In an AI-driven discovery environment, fresh signals help ensure brands remain visible in recommendations.
AI-driven product discovery is still evolving, but several practical actions are already clear.
Brands should focus on:
These actions increase the likelihood that a brand will appear when shoppers ask AI for product advice.
This analysis is just one example of what we are now seeing through our AI Consumer Panel and category research programmes.
We can run the same research for your category to show:
In a world where AI responses often mention only a handful of brands, understanding this landscape early could create a significant competitive advantage.
To get AI Consumer Panel research for your category, get in touch.
Email sales@checkoutsmart.com to request a category analysis.