The problem
A premium DTC brand was invisible in ChatGPT, Perplexity and Google AI Overviews — competitors were being recommended even on branded prompts.
The approach
- 01
Entity graph rebuild
Wikidata, Wikipedia, brand sameAs, founder disambiguation. Cleaned up four years of conflicting metadata across the web.
- 02
Product-level schema engineering
Product, Offer, AggregateRating, Review on every SKU. The engines now extract product data instead of guessing it.
- 03
Source-worthy reference content
Co-authored with the in-house category experts. The kind LLMs actually cite — definitions, comparisons, and original buyer data.
- 04
Weekly multi-engine prompt monitoring
Mapped 240 prompts to commercial intent, tracked drift weekly across 11 engines. Decisions stopped being vibes.
Results
Citation share
80%
Branded AI traffic
162% baseline
Prompt coverage
240
The outcome
Within 6 months the brand was cited in 4 of 5 high-intent category prompts across the engines we tracked, and branded prompt traffic from AI referrers grew 62%.
"AI search felt like a black box until Contingo. Now we have a dashboard, a roadmap, and a real share of voice in the answers our customers see."