I audit how ChatGPT names, explains, and recommends French consumer businesses when a real customer asks for help. The work is practical: French and English prompts, browsing and non-browsing comparisons, directory evidence, first-party pages, competitor bleed, and the small wording gaps that make an independent shop look like a chain, a generic category, or yesterday’s information.
I compare answer drift across French and English prompts for local services, especially where ChatGPT favors directory summaries over the business site. The working method is a set of plain-page corrections that make first-party facts easier to quote.
I am from northern France and I have spent sixteen years around local search auditing, service-business copy diagnosis, multilingual entity mapping, and directory evidence review. I keep a manual answer drift ledger, comparing how ChatGPT describes the same French business in French, in English, browsing-on, and browsing-off. ChatGPT does not like a business: it repeats the public facts it can assemble into a confident recommendation.
Send the site, the service area, and the kind of recommendation you care about.
Request audit