THG Ingenuity Launches Gemini-Powered AI Stylist

The reported figures compare users who chose to generate an image with a broader site-average activewear cohort. They therefore describe an observed performance difference, rather than a controlled causal measurement of the tool's incremental effect. Users willing to engage with a virtual try-on feature may differ from typical visitors in shopping intent or time available to browse.

Jo Drake, CTO Platform at THG Ingenuity, described the product as a scalable virtual try-on experience and said the collaboration with Google Cloud could improve consumer confidence and reduce operational costs. Maureen Costello, Google Cloud's vice president for the UK, Ireland and Sub-Saharan Africa, said making the tool available on Marketplace could help retailers address ecommerce engagement challenges.

Virtual try-on systems combine image generation or image transformation with an ecommerce interface, making the quality of identity preservation, garment rendering, latency, moderation, and measurement central operational concerns. The deployment's engagement metrics indicate that image generation can be incorporated directly into a shopping flow rather than kept as a separate discovery feature.

For teams evaluating comparable tools, the reported Myprotein comparison underscores the importance of experiment design. Randomized holdouts, consistent attribution windows, and return-rate measurement by product category are needed to distinguish uplift from self-selection. Retail Times reports that THG Ingenuity presents reduced returns as an intended outcome, but neither source provides return-rate figures from the initial deployment.