Fevad and KPMG share vision of AI agents in e-commerce at NRF Europe

Generative artificial intelligence tools are reshaping e-commerce as nearly a third of online shoppers now use the technology during their buying journeys, according to an extensive study presented by the Federation of Distance Selling (Fevad) and KPMG at the NRF Retail’s Big Show Europe conference in Paris.

AI Agents Handle Shopping Tasks with Varying Autonomy

The transformation introduces AI agents capable of understanding user intent, comparing offers, and preparing orders, and in some cases, acting on behalf of the consumer to the extent of settling the purchase, as announced by the Federation of Distance Selling (Fevad). However, these tools operate across varying degrees of autonomy. François-Xavier Leroux of KPMG noted that while conversational assistants help formulate needs and recommend products, preparers can build a shopping cart for client validation, and execution under mandate allows the AI to complete purchases within defined parameters.

Despite these capabilities, fully autonomous purchasing remains marginal. According to Odoxa data cited by Fevad, 31% of internet shoppers use generative artificial intelligence during their purchasing paths, with that figure climbing to 73% among regular users. Consumers, however, remain reluctant to cede complete control over financial transactions, preferring to maintain final authority at the moment of payment. As the action scope of the agent widens, concerns regarding personal data, payment security, and liability grow increasingly critical, making it essential for merchants to offer tangible benefits while allowing customers to validate, correct, and resume manual control.

Kingfisher Adopts Google Technology to Improve Product Discovery

While traffic originating from major AI models currently accounts for less than 1% of total e-commerce traffic, conversion rates on those visits multiply by two, three, or four times. This indicates that consumers arriving via AI tools possess highly refined purchase intentions. Because a significant portion of research occurs outside merchant websites within assistant mobile applications, retailers face challenges in attributing traffic accurately.

Merchants must now optimize product information, pricing details, and inventory timelines so that automated agents can interpret and recommend their catalogues. Beyond front-end discovery, this evolution impacts internal merchant operations such as logistics and customer relations.