Retail & commerce20253–4 months
Client: An omnichannel retailer
One of the project’s screens, in a phone frame. It is shown again with its caption under “On screen”.
At a glance
Search and AI visibility measurement
Structured product data and feeds
Category guides written as direct answers
From goal to result
The goal in full, our approach, what we built and what it set out to change.
Shoppers were asking AI assistants which product to buy. The answers cited review sites and forums, rarely the retailer’s own guides, even where those were the better source. Search rankings alone no longer showed how visible the retailer was.
The brief: get cited in AI answers as well as ranked in search results.
We fixed the questions first: a panel of priority shopper questions, checked on a schedule across the main AI assistants. Visibility is measured rather than guessed.
Then we made the content easy to cite: complete structured data, clean feeds and category guides written as direct answers with their evidence.
Priority questions by category, each with the answer we would want cited.
The product data model, feeds and schema markup, completed.
Category guides rewritten as direct answers, each with its sources.
The panel re-run on a schedule; citations reported by category and by assistant.
Structured product data, schema markup and guides written as direct answers, tracked against a fixed panel of priority questions in AI assistants.
The work sets out to make the retailer’s product data complete and its guides easy to cite. The tracker shows, question by question, which assistants cite the retailer’s guides and which cite someone else.
Share of AI answers is read from the fixed panel. The client approves any figure before it is published.
Picture story
These pictures are examples: licensed stock photos and mock-ups we built in code. They show the setting, not the client’s own work. Credits are at the foot of the page.
On screen
Example screens we built in code, like the ones the project made. They are not the client’s own. Open any full screen.
Before and after
An example before and after, built in code, of the kind of change the work made. These are not the client’s screens. Drag across to compare.
Before · not cited
After · cited
Outcomes
What each measure is compared with, when we read it and who reports it. Results show once the client approves them.
Share of AI answers citing the brand
Organic revenue from priority categories
Product data completeness
Figures pending client approval
We publish results only with the client’s written approval. Until then, we judge the project on the plan shown here. Ask for a private walk-through.
Credits
People are named only with their agreement.
Licensed stock photos and mock-ups we built in code, to show the setting. None is the client’s own work.
Keep going
Next case · 05 · Retail & commerce
Cut a season’s campaign production from six weeks and three agencies to days.
Read the next case
You will talk to the lead who would run your work.
01About 2 minutes
You get A reply from a named lead
02About 8 minutesRecommended
You get Options and prices after one call
03About 15 minutes
You get Confirmation and a named lead
We always agree a written scope and price before work begins. How each plan is priced
