A brand can be strong within its category and still lose the sale if an AI shopping tool decides another category better answers what the shopper wants.
A request for a high-protein breakfast, dinner for five in 20 minutes, or something to take into the office does not specify a product. It gives the agent an occasion and leaves it to work out which products could fit.
Toby Hilliam, co-founder and CEO of Appetise, said that could change how FMCG brands think about where they need to build relevance.
“The agent often reads the occasion first, then works out which products fit it,” said Hilliam.
“That’s actually how people have always shopped, and AI just makes the mechanism visible: nobody buys ‘pasta’; they buy dinner, and pasta is one ingredient in a decision that looks different depending on whether it's a date night, or a need to feed a family of five in 20 minutes.”
That can put products into a wider competitive set. If the requirement is an easy family dinner, pasta may be competing with tacos, a frozen meal, or something from another category altogether.
Appetise's data also suggests shoppers are not equally willing to hand over every product decision.
Shoppers have been more comfortable using AI to help decide what to cook than allowing it to decide what food to buy. Hilliam said the willingness to delegate also changed depending on the product.
“The direction of travel looks like informed delegation, not full automation: think more sous chef than head chef,” said Hilliam.
“Shoppers want AI to handle the routine, repeatable calls and want to stay more in control of the personal ones like coffee or chocolate.”
Existing brand preference appears to provide some protection.
Read more about AI & the trolley in the latest issue here
