AI shopping assistants are creating very different competitive environments for brands, with the same product search producing vastly different recommendations, retailers and price ranges depending on the platform, according to new research from Jellyfish.
AI shopping assistants study
The digital marketing agency, part of The Brandtech Group, analysed shopping recommendations across eight product categories in the US, UK, Australia and Singapore, examining how ChatGPT, Google AI Mode and Amazon’s Alexa for Shopping select and recommend products.
The findings reveal significant differences in the brands and retailers surfaced by each assistant, raising questions about how marketers should approach visibility as consumers increasingly turn to AI for shopping advice.
Jellyfish has also launched Shopping Optimization, a new capability within its Share of Model platform, designed to help brands understand why AI assistants recommend particular products and how to improve their visibility.
Different assistants, different shopping experiences
One of the clearest findings is that there is no single AI shopping environment in which brands can compete.
When asked to recommend men’s suits in the UK, ChatGPT drew on 30 retailers, while Google’s AI Mode considered just four.
The differences extended across other categories. In US athletic wear, ChatGPT drew on 171 retailers compared with 126 for Google’s AI Mode.
Amazon’s shopping assistant presented a different picture. When asked for toy recommendations, it featured products from 177 brands but sourced them almost entirely from Amazon itself. ChatGPT, by comparison, drew on 24 retailers and Google’s AI Mode on 37.
Across the eight categories studied, approximately 80% of the recommendations measured came from ChatGPT and 20% from Google’s AI Mode, although the split varied considerably by category. In home appliances, the difference reached 97% versus 3%.
Established brands face new competition
The research also suggests that AI shopping could make it harder for established brands to dominate product recommendations.
A single shopping query surfaced as many as 290 competing brands. In US fashion, recommendations featured 120 brands, but the most frequently recommended brand accounted for just 7% of the results.
Traditional price positioning was similarly blurred.
In one UK search for men’s suits, AI recommendations ranged from £27 to £3,295. For gaming chairs in the US, prices stretched from $79 to $3,479.
The findings suggest that premium and budget brands can appear alongside one another in the same AI-generated shortlist, potentially changing how consumers discover and compare products.
Jellyfish launches Shopping Optimization
To help marketers navigate these differences, Jellyfish has introduced Shopping Optimization within its Share of Model platform.
Unlike conventional digital-shelf tools, which track product rankings on individual retail websites, the new capability examines recommendations at individual product, or SKU, level.
It allows marketers to compare which products different AI assistants recommend, their ranking, pricing and ratings, and the retailers from which they can be purchased.
Brands can also investigate why particular products are recommended or overlooked, identify opportunities to improve their visibility and track changes in recommendations over time.
“Until now, marketers have had little visibility into why AI shopping systems favour one product over another,” said Natasha Wallace, Chief Solutions Officer at Jellyfish.
“Share of Model’s Shopping Optimization gives brands a clear understanding of the factors driving AI recommendations, so they can focus their Optimization efforts and measure the business impact.”
John Dawson, Vice President, Strategy at Jellyfish, said the emergence of AI shopping assistants represents a fundamental change in how brands compete for consumers.
“Brands have spent decades optimising products for search engines, marketplaces and retailer shelves. Agentic commerce changes that equation.
“AI shopping assistants are becoming active participants in purchase decisions, creating a new decision-maker brands must optimise for.
“Our own data shows the same product question can return a 30-retailer shortlist on one assistant and a closed, single-store answer on another – so ‘winning AI’ isn’t one race, it’s many, and most brands can’t yet see the starting line.”
Shopping Optimization extends Jellyfish’s Share of Model platform beyond tracking brand mentions in AI-generated responses to analysing the individual products consumers are shown.
The new functionality is available globally and forms part of the agency’s wider Generative Engine Marketing offering, alongside its Creator Intelligence and AI Ads Optimization capabilities.