AI Revolutionizes Online Shopping: One in Two French Consumers Already Won Over

AI, Already a Reflex for One in Two Buyers

The use of conversational artificial intelligence is intensifying in the realm of online shopping. A recent study published by Botnation on July 6, 2026, highlights this trend. Botnation, a French company specializing in conversational agents, conducted this survey with 4,204 participants. The results show that both consumers and sellers are inclined to integrate AI more into their decision-making processes, but under specific conditions.

The Methodology of the Botnation Study

The study was conducted online through a CAWI survey between June 19 and June 26, 2026. It involved 4,204 respondents, comprising 3,103 consumers and 1,101 sellers, including merchants, e-commerce professionals, shop managers, customer relationship professionals, and leaders of small and medium-sized enterprises. The sample was constructed using a quota method and adjusted based on sociodemographic data from INSEE. Participants were recruited through the BuzzPress panel, using email invitations and advertisements on Facebook and LinkedIn.

AI, Already a Reflex for One in Two Buyers

Nearly 49% of the French respondents reported using artificial intelligences such as ChatGPT, Claude, or Gemini before making a purchase. This usage takes various forms: 16% use it only for significant or expensive purchases, 13% verify the information obtained on other sites, 11% use it systematically to save time, and 9% primarily use it to summarize customer reviews and product tests.

In contrast, 29% of respondents still prefer to conduct their research via Google, merchant sites, or forums, and 18% do not trust the responses provided by an AI when choosing a product. The use of assistants integrated into websites follows a similar trend: 13% use them often, 36% sometimes, while 29% rarely use them and 18% never do.

This adoption of AI is accompanied by an openness to automated recommendations. Indeed, 58% of consumers would be willing to accept an AI agent recommending products instead of a human seller, with 19% doing so unreservedly if it saves them time.

Being Advised Without Letting Go of the Final Decision

Consumer acceptance of AI is conditioned by a clear division of roles. In the product search process, respondents primarily see AI as a decision-support tool. Thus, 29% would accept it comparing prices and product features, 22% would allow it to recommend the best product as long as they validate the purchase themselves, and 16% would limit it to suggesting product ideas.

However, the idea of fully autonomous purchasing remains difficult for many to accept: only 10% of participants would let an AI manage certain purchases on their behalf, while 18% refuse any AI intervention in their purchases. Caution is also warranted throughout the entire purchasing journey, with 59% of people open to automated purchasing, but mainly if a human can intervene in case of problems (31%) or for simple, low-stakes purchases (17%).

Certain categories of products are more suited to this type of assisted shopping. High-tech products and home appliances lead the way (51%), followed by clothing and accessories (38%), gifts (37%), and travel (36%), all areas where comparison and sorting can be time-consuming.

On the Seller Side, Broad Adoption but Under Control

Commerce professionals are not lagging behind. According to the study, 74% of sellers surveyed would be willing to let an AI agent recommend products to their customers, but rarely without safeguards. Indeed, 25% agree to this conditionally, wanting to maintain control over the recommendations, and one-third prefer using a supplementary AI, activated before human intervention or outside of business hours. Total trust in AI remains a minority view, with only 11% of sellers willing to rely entirely on it, believing it responds faster and sells better.

Sellers' hesitations are mainly related to customer relationships: 16% consider the human relationship essential in sales, and 9% fear that AI might make poor recommendations. When considering the use of AI, it is often delineated, between guiding the customer before human intervention (21%) and taking over outside of business hours (17%).

Trust Depends on Price and Responsibility

Two aspects raise significant reservations. Price, first: 46% of respondents believe that a product recommended or sold by an AI should cost less, either due to savings in time and personnel (29%) or because they consider AI advice less valuable than human advice (17%). Conversely, 34% think that the price should remain tied to the effectiveness of the service or the quality of the product, regardless of AI.

Responsibility, next: in the event of poor advice or an unsuitable purchase, 90% of French people place the responsibility on the professionals. The brand selling the product is cited first (26%), followed by the company using the AI and shared responsibility (18% each). Only 4% believe that the consumer should bear this responsibility, even if they chose to use AI.

Emmanuel Françoise, founder of Botnation, emphasizes that this survey reveals a willingness among consumers to integrate AI into their purchasing journeys, provided it is perceived as useful, transparent, and controlled. "AI should not replace trust; it should enhance it," he asserts.