Case Studies
Measured movement, raw answers.
Every figure below is measured the same way, a fixed battery of real buyer questions (your brand name never in the question), re-asked across the AI engines weeks apart, with every raw answer retained. Not a dashboard estimate, the actual answer the buyer sees.
The numbers
Three brands, moved.
Across specialty chemicals, natural food colourants and handmade art, on the same before-and-after method.
Every result above is measured on a fixed buyer-question battery, re-asked across the AI engines weeks apart, with the raw answers retained. Last updated September 2026.
Featured · Handmade art
Tangled Monk: 8 to 16 in a month
A Singapore handmade-art brand, measured 10 Jul to 24 Aug 2026.
The starting point
When buyers asked the AI engines for handmade and ceramic art, Tangled Monk was named in a handful of answers and rank 4 among the brands the AI recommended, a 7.1-point share-of-voice gap behind the leader. All 59 pieces and 36 products were missing from the sitemap, so many pages the AI might cite were not even discoverable, and the off-site footprint was thin: 11 of 27 target listings, with 0 independent editorial pieces.
What we did
We worked the on-site levers first. Every piece and product went into a complete sitemap with the AI crawlers unblocked, so the catalogue became retrievable. Pages were rewritten answer-first and held to a citability bar before shipping, one page's citability rose from 1.3 to 7.3. Identity and Product schema were made consistent so every source the AI checks agrees on the same facts.
The result
In one month, direct AI recommendations went from 8 to 16 and share of voice from 7.9% to 18.8%, with nothing regressed. Crucially, 15 of the 16 recommendations were live Gemini web citations, the retrieval behind Google AI Overviews, not training-memory recall, so this is visibility a buyer actually encounters in a live search.
What we had not done yet (the honest line)
The doubling came from on-site work alone. The off-site lever, footprint 11 of 27 with 0 independent editorial pieces, had not been started, which is exactly why the AI still preferred rivals on the closest near-win questions, and where the next gains come from. We report the upside we have not touched, not just the wins.
Also
Exporters, made citable.
Specialty chemicals
From 0 to 21.4% share of voice on Gemini. Identity consistency rose from 4.0 to 14.0 out of 20, and the AI now recites the company's own facts back to the buyer: ISO 9001:2015 certified, 99% purity, Certificate of Analysis.
Natural food colourants
A brand the AI could not find at all. After its identity was rebuilt from the ground up, one consistent entity across Organization schema, Google Business Profile and Wikidata, share of voice went from 4.1% to 14.3% in six weeks.
Want the same read on your brand?
A free GEO audit shows exactly who the AI names instead of you today, traced to the raw answers, and whether your gap is measurement or execution.
Get your free audit →