B2B services firm: from 0 of 80 prompts cited to 23 of 80 in 90 days
Situation
The firm offered specialty B2B services to a narrow buyer set. Their site ranked acceptably in Google for branded queries, but AI tools either could not find them or described their services incorrectly. None of the 80 baseline prompts we ran through ChatGPT, Perplexity, and Claude cited their site.
Approach
We rebuilt the service-page surface around definition blocks, FAQ sections, and comparison tables. We shipped a fuller JSON-LD graph: Organization, Person (the principal), Service, FAQPage, and BreadcrumbList. We hand-authored an llms.txt at the root, fixed the entity-language inconsistencies that had been confusing AI tools, and configured robots.txt to explicitly allow the major AI crawlers.
Outcome
After 90 days, 23 of the 80 baseline prompts cited the firm by name. Roughly 85% of those citations described their services accurately. Search Console showed AI Overview impressions for the first time. The firm now has a citation-baseline tracking document they re-run every quarter.
Representative engagement. We have withheld the client identity for privacy. The numbers above are from this specific project. We do not promise the same outcome on a different one.