RaceX moved from 63/100 to 94/100 on the Visible2 readiness scan after a controlled entity and consistency cleanup. RaceX is a Dubai racehorse ownership company and one of Jennifer McShane Bary’s own ventures. We disclose that relationship because this is a product-validation case study, not an independent client testimonial.
The starting problem
The first scan found a technically readable site but weak entity-definition and corroboration scores. Manual validation then exposed an important product issue as well: the live site already contained a valid SportsOrganization graph that the early parser had failed to recognise correctly.
What changed on RaceX
- The LinkedIn company profile was updated to remove legacy membership, jockey and location language.
- The canonical site name was standardised to RaceX rather than an SEO tagline.
- The existing Organization graph was verified in live source rather than duplicated.
- A CEO relationship was made explicit in the entity graph.
- The YouTube channel identity, description and contact details were brought into line with the current operating model.
- Instagram, LinkedIn and YouTube were connected through the Organization sameAs array.
The result
The later scan reported 100% Entity Definition, 100% Name Consistency and 100% Corroboration Readiness, with an overall readiness score of 94/100.
What we deliberately did not do
We did not create duplicate Organization schema. We did not rewrite the entire website. We did not add pages simply to reach a perfect score. The remaining findings included a slow sample response, one thin sampled page and question-answer structure — all of which needed context before action.
What 94/100 did not prove
It did not prove RaceX would be recommended by ChatGPT, Gemini or Google AI Mode. The readiness scan establishes that the site is understandable and technically eligible. Actual recommendation visibility still requires repeated buyer-question testing.
What Visible2 learned from its own case study
The remediation improved RaceX, but it also improved Visible2. The parser was updated to recognise SportsOrganization correctly, scan freshness became a product requirement, and the report logic was changed so a PASS cannot be followed by advice to install a duplicate version of the thing that already passed.
That is the standard we want from AI visibility work: measure, inspect the evidence, change only what the evidence supports, then measure again.