Send the answer that feels wrong

I start with the brand's own public record: official pages, branch pages, product wording, press pages and the English/French versions if both exist. Then I compare that record with repeated AI answers and name the exact ambiguity causing trouble. The output is a written audit with wording repairs and prompt tests.

I take on ownership confusion, parent-brand and product-line hierarchy, multi-location structure, geography drift, rebrand and heritage wording, product provenance and reputation-balance issues. I am most useful when the brand already has public evidence but the seams are weak.

I usually reply within a few working days through this site's form. If the case looks like a fit, I ask for the key public surfaces and a few sample AI answers before suggesting a format.

A consultation is usually a focused call plus a short written note. For larger work, I build a full entity audit, map the evidence seams, propose quotable wording and give you a repeatable test set for several answer styles.

Most serious engagements sit between USD 6,000 and 28,000, depending on the number of brands, locations, languages and evidence surfaces. I give a rough scope before any paid work begins.

I do not write fake reviews, bury bad history, impersonate customers, mass-produce directory text or promise that every AI system will answer the same way. I also avoid cases where the company wants a cleaner answer without correcting the public record.

I compare French and English separately when both matter. The two records often tell slightly different stories, and that difference is where AI systems start to wander.