The overview
A confidential GEO engagement for a major European visitor attraction that was being overlooked in AI-generated recommendations. The work combined prompt-level measurement, technical SEO diagnosis, entity analysis and third-party source research to explain why—and to establish a credible path forward.
A brand can appear in AI answers without a functioning website, but it cannot control or compound that visibility without strong owned evidence.
01
The challenge
The attraction competed with museums, landmarks and family destinations for inclusion in AI-generated trip-planning answers, but it was not being recommended consistently. Because this was early in the development of GEO, there was no dependable off-the-shelf playbook: the measurement framework had to be designed alongside the investigation.
02
Building the baseline
I helped define representative prompts, comparison criteria and evidence categories so performance could be assessed systematically rather than through a handful of anecdotes. The baseline separated brand mentions, recommendations, citations, source types and the contexts in which competitors were preferred.
03
The technical blocker
The investigation uncovered a fundamental owned-site problem: an age-gate implementation prevented Googlebot from reliably crawling and indexing the attraction website. Thin generated pages, limited structured data and poor performance compounded the issue, restricting the material that conventional and AI-powered search systems could retrieve.
04
The counter-intuitive finding
Despite the inaccessible owned site, the attraction still had a degree of AI-search presence. Reddit discussions, travel-review platforms, Wikipedia and Wikidata supplied enough corroborating evidence for models to understand and sometimes recommend the entity. That demonstrated an important principle: AI visibility is shaped by the wider information ecosystem, not only by a brand's website.
05
The value delivered
The engagement turned an emerging, poorly understood problem into a measurable programme: repair crawlability, strengthen the entity and content layer, improve retrievable first-party evidence, and monitor both owned and third-party sources against a repeatable prompt set.
06
Turning a new discipline into an operating method
The deliverable was more than a snapshot. It created a repeatable way to test representative prompts, record recommendations and citations, compare competitors, classify the sources influencing answers and measure the same questions again after implementation. That turned an unfamiliar subject into work that technical, content and brand teams could own.
The strategic takeaway
What this work demonstrates
The project exposed both the technical foundations of GEO and the importance of the wider information ecosystem surrounding an entity.
