The challenge
The brand ranked respectably on Google for its core category terms — top five, sometimes top three. But when the team started actually testing ChatGPT, Perplexity, and Google's AI Mode with the same questions their prospects were asking, the brand simply wasn't in the answer. A specific competitor, ranking lower on traditional search, was getting named consistently instead. Leadership had noticed a slow, hard-to-attribute softening in inbound interest and suspected — correctly — that this was part of why.
What made it harder
- The site's content was written in a traditional SEO style: long intros, keyword-led headings, answers buried several paragraphs deep
- No structured data beyond basic Organization schema existed — nothing that gave an AI system a clean, extractable fact base about the brand
- The competitor being cited instead had objectively weaker backlinks and lower rankings, which made the gap concretely diagnosable rather than mysterious
- The brand's off-site presence was thin — almost no Reddit, YouTube, or forum footprint, which meant AI systems building context from beyond the site had little to draw on
The approach
- Ran a full AI-readiness audit across 14 real AI crawlers to confirm access wasn't the issue — it wasn't; the content itself was the problem
- Identified the exact buyer prompts where the brand should have been the answer and wasn't, then reverse-engineered what the cited competitor's content structure had that this site's didn't
- Rewrote key pages answer-first — direct, extractable two-to-three-sentence answers immediately under each heading, with supporting depth after, not before
- Implemented a full JSON-LD entity graph (Organization, Person, Service, FAQPage) to give AI systems a clean, unambiguous fact base to cite from
- Built targeted presence on two industry forums and a YouTube explainer series specifically because AI systems were observed pulling context from exactly those source types for this category
- Re-ran the same prompt set monthly to track citation appearance and refine wording where the brand appeared but wasn't the primary cited source
The outcome
- Moved from zero appearances to consistent citation across the tracked prompt set within the engagement window
- Closed most of the 'share of model' gap against the competitor that had been getting cited instead
- Traditional rankings held steady or improved slightly as a side effect — the AEO restructuring didn't cost anything on the SEO side
- Left a repeatable monthly monitoring process so the team can catch drift and keep expanding coverage without starting from scratch each time
The takeaway
Ranking and citation are different games measuring different things, and a brand can win one while quietly losing the other for months before anyone notices — because there's no Search Console equivalent flagging the drop. The fix wasn't more content; it was restructuring what already existed to be extractable, plus giving AI systems a clean entity to point to instead of a wall of unstructured text.
Anonymised under NDA. Representative of a real engagement; specifics withheld per client agreement.