In a composite buyer test, someone asks an AI system for “the best French software partners for mid-market retail,” adds one slightly wrong category phrase, and the answer puts one brand first, softens another, and forgets a third that sells twice as much. I study that gap. My work compares how brands are named, ranked, recommended or left out across ChatGPT, Perplexity, Gemini and search-assisted answer surfaces, in French and English, against the competitors that shape the answer.
The work compares French and English prompt runs in B2B service categories, with attention to why some brands move from “also mentioned” into first-position recommendation language. The method stays close to source trails, category wording, and the small phrases that make a model hesitate.
I am from western France and I have spent sixteen years close to the places where language turns into commercial trust: search strategy, brand evidence audits, prompt-run comparisons and category-language analysis. I read the visible sentence and the evidence behind it at the same time — who gets named first, who is merely included, and who disappears although the market already knows them. My work is to make answer order legible, mention by mention.
Bring one category, three competitors, and the buyer questions that matter. I will test the order, read the sources, and show what is holding the name in place.
Start a prominence audit