Found in the answer, not just the results page

A specialist retailer invisible in AI answers despite ranking well in search. We restructured the catalogue and content for citation.

Read the case study

Services

  • GEO
  • Content

Details

Client
Withheld
Industry
Retail
Delivered
April 2026
  • A glossy violet plaque mounted clear of a terracotta tiled wall, hard shutter light raking across the tiles behind it.
  • A row of blank cardboard swing tags hanging from a rail against a terracotta wall, all facing away, hard shutter bars falling across them.
  • One blank swing tag turned to face the lens on its cotton loop, the deep coral wall behind it falling out of focus.
  • Three blank swing tags on a sunlit limestone sill, two standing squared against a rough block and one lying flat in front.

The problem

The site ranked respectably for its core terms and was still absent from the answers people actually read. Assistants summarising the category cited competitors, several of whom ranked below it. The content was written to be skimmed by a human already on the page, not quoted by a model deciding who to mention.

What we did

Rewrote the category and product pages so each opens with a short, self-contained answer to the question the page exists for, before any elaboration. Product attributes moved into structured markup so specifications are machine-readable rather than implied by a photograph.

Entity signals were tightened: one consistent business name, linked profiles, and specifications stated as text rather than living only inside images.

What changed

The pages now say what they are in a form that survives being extracted. That is the whole mechanism — an engine cannot cite a claim it has to infer from a product shot.

Result

40–60
Word answer blocks per page
100%
Product data in structured markup

Want the same for your business?

Send the site and the service you want more of. We reply with what we would change first.

Start a project