On April 1, during the AWS Summit in Paris, Decathlon presented a review of AI integration within a content strategy. For the retailer active in 79 countries (2025 turnover: €16.8 billion, up by 5.6% at constant exchange rates, including 20.2% online), the project involved a digital asset management system (DAM), which manages and stores 19 million of multimedia assets, each with their own life cycle. These images, videos, audios, 3D or 360° formats, previously lacking meta descriptions, are used to present products on the e-commerce site (technical photos), marketing actions (social media campaigns, local activations…) and brand identity (logos, institutional assets…). The goal of the project was to make the use of this content easier for internal staff and increase compliance with the European directive on accessibility.
A DAM associated with high costs and improvable quality
According to Lévi Bernadine, Machine Learning Engineer at Decathlon Digital, “until now, we had €7 million in annual fixed costs, linked to the need to correct and translate our content into French and English, with a certain number of anomalies that persisted afterwards, such as abnormal classification, non-compliance of product sheets in terms of CSR, etc.”
