Across the retailer's 28 stores and its growing e-commerce arm, buying glasses online and buying them in person were two very different experiences. In store, a customer picks up a pair and looks in a mirror. Online, they chose from a flat product photo and a size guide that, in practice, went largely unread.

That gap showed up directly in returns. Online frame purchases came back at nearly twice the rate of in-store purchases, and the reason was consistent: shoppers could not judge how a frame actually sat on their own face until the physical pair arrived. Every one of those returns was margin the retailer had already spent to acquire and fulfil the order.

The disconnect extended into the stores themselves. When a customer who had browsed online came in for an appointment, the optician had no view of what they had looked at or shortlisted beforehand. Every in-store fitting started from a blank page, even for customers who had already done real research online.

We built an AR virtual try-on that puts frames on the shopper's own face using camera-based face tracking, so the moment a size guide used to try and fail to convey now happens visually, in seconds. It runs in the iOS app and directly in the mobile browser, so shoppers who never download an app still get the same experience.

A 3D asset pipeline converted the retailer's frame catalogue into try-on-ready models, so every frame a shopper could browse online was also one they could see rendered on their own face. This closed the biggest gap between the online and in-store experience without asking the retailer to rebuild its catalogue.

Saved favourites now sync to store appointments, so when a customer who has browsed and tried frames on online books a fitting, the optician opens the appointment already knowing their shortlist. The online and in-store journeys became one continuous experience rather than two disconnected ones.