Competitive AI-visibility audit
Documented who ChatGPT, Gemini and Perplexity were naming for every priority niche, and the signals earning those citations.
> which Hong Kong law firms specialise in virtual assets, fintech and Web3?
The case study you can verify before you finish reading it.
Ask Perplexity that question today and the first firm it names is TITUS Solicitors — a Hong Kong law firm whose Generative Engine Optimization programme I ran end to end, from first audit to first-named answer.
Last verified: 2 July 2026 · AI answers vary slightly between runs — the live check is the point.
01 / The situation
Strong firm. Strong niche. Invisible to the engines buyers had started using.
TITUS Solicitors is a modern Hong Kong law firm with genuine depth in virtual assets, fintech and Web3 — exactly the niche where Hong Kong's regulatory momentum was sending founders, exchanges and issuers looking for counsel. But those buyers had changed how they look. They were asking ChatGPT and Perplexity who to call — and the assistants were naming other firms.
The firm wasn't losing pitches. It was missing from the shortlist before a pitch could happen.
The mandate, starting November 2024: make the AI engines recognise, trust and recommend TITUS for the niche it actually deserved to own — and prove it with numbers.
02 / The method — six parts
Nothing exotic. Everything connected.
GEO levers only work together — a schema fix without content gives the engines nothing to cite; content without entity clarity gives them no one to credit. The programme ran all six in sequence.
Schema & author markup
Structured data across the site so engines could read the firm, its practice areas and its authors as machine-legible facts.
Knowledge-graph entity
A Wikidata entity and consistent cross-web identity, so assistants resolve TITUS as one real, authoritative thing.
Directory & profile authority
The third-party corroboration layer — the sources AI engines actually cite when they recommend a firm.
Answer-first bilingual content
An English + Traditional Chinese content engine structured to be the quotable answer — with a legal-accuracy QA loop, because in law a wrong answer is worse than no answer.
Measurement
Everything benchmarked in GA4 and Search Console from day one, so the results below are data, not anecdote.
03 / The results, year on year
Before and after.
- +183% organic search impressions, year on year
- ~2× organic clicks — roughly doubled
- 23.5 → 8.2 average Google position — page three to page one
And the headline outcome no dashboard shows: ask Perplexity who handles virtual-asset and fintech law in Hong Kong — TITUS is the first name it gives.
Source: GA4 and Google Search Console, year-on-year comparison; Perplexity result checkable live via the button above. The firm also owns a repeatable playbook — the visibility doesn't leave when a retainer ends, because there is no retainer.
04 / What this means for you
> could this work for my business?
If your buyers ask AI before they call, the same method applies.
Law is one of the hardest categories to do this in — regulated, accuracy-critical, trust-sensitive. If it works there, it transfers: professional services, fintech and virtual assets, startups, and Hong Kong SMEs. The first step is seeing how AI describes you today.
Published 2 July 2026 · Ivor Ngo, Ivorization · Full narrative version on the blog