Traffic from generative AI to the brand has increased tenfold since 2025. Boulanger is restructuring its local data to maximize its mentions. The brand relies on trust and authority to influence algorithms.
According to the first edition of GEO barometer by Valiuztraffic from generative AI to brand sites like Boulanger was multiplied by 18. Published in early 2026, this study covered the period from December 2024 to December 2025. “More than 90% of this traffic comes from ChatGPT, with a peak observed in August 2025, corresponding to the arrival of ChatGPT-5 on the French market,” says Guillaume BotSEO manager at Boulanger.
But, currently, direct traffic from LLMs like ChatGPT to the site boulanger.com represents less than 1% of the overall web audience, he assures. However, the opportunities offered by visibility in AI models are numerous.
Harmonize information from 223 stores across France
To appear in these answers, Baker thoroughly restructures its data. The goal is not to consider AI models solely from the angle of traffic acquisition, but as powerful levers of notoriety and indirect business contribution. “The main thing for us is brand visibility,” says Guillaume Bot. “Unlike traditional search engines, these Generative AI are designed as “closed” ecosystems. They favor resolving the query within the conversation itself, without necessarily encouraging the user to click to an external link.”
Their strategy is to maximize the brand’s presence in the AI’s recommendations. “By being cited and repeated organically, Boulanger naturally fits into the consumer decision-making process.” To do this, Boulanger first focuses on visibility and conversion. “Our priority is to guarantee that our brand appears relevantly on each query related to our offer,” says Guillaume Bot. “The objective is to provide an immediate response which naturally directs the Internet user towards our stores or our e-commerce site.”
In this sense, Boulanger has partnered with Yext to centralize, harmonize and automate the management of local information for its 223 stores across France (hours, services, photos, reviews, brand corners). This is made possible by multicasting data to Google Maps, Waze and Google Business Profile via the Yext Knowledge Graph. “We provide in-store teams with intuitive tools, including a mobile application, to allow them to easily enrich their local content.”
This hybrid model makes it possible to rely on a centralized database, while promoting the uniqueness of each point of sale via local news, highlighting employees and specific offers. All of this information is then instantly synchronized on local mini-sites and on the thousands of contact points in our partner network. This strategy, which consists of making web-to-store an infrastructure for activating the physical network, has borne fruit in SEO. The brand recorded +30% of Google clicks outside the local brand at the start of 2025.
Second lever: reputation and trust. “Beyond Google reviews, we monitor satisfaction via the Net Promoter Score, or NPS, on all third-party sites and social networks,” enthuses Guillaume Bot. “These trust signals are important for AI algorithms.” Remember that NPS measures the probability of brand recommendation by customers. “Rather than asking the customer if they are satisfied, we are developing the concept of ambassador,” smiles Mikael Stephan. “AI interprets this as trust signals and can put us in front.”
Worked and designed with Yext in November 2023, their web application also allows them to be taken up more easily by AIs, according to Mikael Stephan. “We have a good vision of Google reviews to which the store can respond in real time in the field with the right elements,” he says. “The American giant can rely on our customer experience.”
Finally, Boulanger is based on content authority. “We want to make our brand a reference source of information,” points out Guillaume Bot. “By constantly enriching our site with diverse content, quality videos, content focused on usage and repairability, we ensure that conversational AIs have reliable and qualitative data to recommend Boulanger.” According to Semrush Enterprise AIO data, Boulanger generates over 30,000 monthly mentions in AI responses.
Watch out for the competition
However, certain difficulties remain. “LLMs sometimes favor manufacturer sites for prompts associated with product characteristics,” says Guillaume Bot. “We counter this bias by in-depth work on the uniqueness of our product sheets. On the other hand, certain dynamic data such as reviews and delivery times remain difficult to read for the models. They depend on our APIs. Optimizing the structure of our pages for these AI is a delicate trade-off. It must not degrade our Web Perf which impacts the user experience on our website.”
