In China, as of this week, a shopper can point an iPhone camera at a pair of trainers and land on the product page within seconds, with the price, its price history and an AI-generated summary of reviews. The takeaway for management is clear: the next product search may not start with words, but with a photo. And whoever has the best-prepared catalogue will win it.
What happened
On 4 November, Apple opened the public beta of its artificial intelligence assistant on iPhones sold in China. It came with so-called “visual intelligence”: the user points the camera at a real object and the phone suggests third-party services to search for it.
One of the first connected apps is Dewu, China’s leading platform for trainers and fashion. It works like this:
- The shopper points the camera at a product and picks Dewu as the search service.
- Matching products on sale appear, with their prices.
- Tapping one opens its product page directly in the app: materials and features, an AI-generated reading of how the product performs and what buyers say, a price trend chart and real reviews.
- The purchase happens right there, without going through any search engine.
There are no usage figures yet, and the feature is only available in China, in beta, and to users who have the app installed. But the journey already exists: from seeing a product in the street to having it in the basket in three taps.
The bigger picture: the camera as a search engine
For twenty years, selling online has meant showing up when someone types what they are looking for. This move points to a different entry point: the customer doesn’t know the model name, but has it right in front of them. They see trainers on the underground, a lamp in a restaurant or a part in a competitor’s warehouse, and they ask the camera.
What matters is not the recognition technology, which already existed, but who captures the sale. Here, the phone doesn’t send the customer to a list of links, but to a single platform offering price, context and instant purchase. And an AI that reads product pages and reviews on the customer’s behalf helps them decide.
China tends to be one or two years ahead of Europe in mobile shopping habits. That is why it is worth looking at now, not when it arrives.
Why it matters to your business
Sales. If search starts with an image, the winning product is the one the AI recognises and can describe with confidence. A catalogue with poor photos or incomplete product pages won’t compete, even at a better price.
Margin. When the customer sees price history and a review summary on the same screen, comparison is instant. Brands that compete only on price will find it harder; those that explain why they are worth what they cost, easier.
Dependency. In the Chinese case, the sale goes through an intermediary. If the journey looks similar in Europe, the question is whether the phone will take the customer to your website or to a marketplace. That depends largely on what you do beforehand.
What to review now
Nothing needs to change tomorrow, but the management team should be asking a few questions:
- Would an AI understand our catalogue? Clean product photos from several angles with the product clearly identified; complete, consistent names, materials, sizes and uses across every channel.
- Are reviews visible and readable? If an AI is going to summarise what customers say, there need to be reviews, they need to be genuine, and they need to be in a format it can read.
- What is the single source of truth for product data? If information is spread across the ERP, spreadsheets and the website, each channel will tell a different story. And AI rewards consistency.
- Where do we want the search to end? In our own store, on a marketplace, or both. It is a channel and margin decision, not just a marketing one.
An illustrative example: a footwear brand with 600 SKUs keeps its photos in a shared folder, compositions in the ERP and descriptions only on the website. The day an AI assistant searches for “grey running shoe with white sole”, that brand depends on someone having joined those three pieces together.
How we approach it at TRILOGI
For us, this story isn’t about Apple or China: it’s about the quality of product data. When a client asks how to prepare for AI-driven search, we start with the catalogue, not with advertising.
First, we map where each piece of product data lives and what state it is in. Then we centralise photos, attributes and copy in a single place, usually a PIM connected to the ERP and the store, so that the website, marketplaces and any AI assistant receive the same complete, up-to-date information. Only then do we work on the visible layer: images designed to be recognised, product pages that answer customers’ real questions and well-structured reviews.
If you’d like to see how we do it, we explain it in our PIM integration service.
Every Tuesday, the 5 stories that matter: subscribe to the newsletter.
