Decathlon: Turning Retail Data into Intelligent Store Operations

01 / 05

The challenge

Large retail operations generate massive amounts of data across inventory, store operations, product availability, merchandising and customer interactions. Despite this volume, information often remains fragmented across systems and teams. Store managers need actionable insights, not additional dashboards. Operational teams need to detect stock discrepancies, understand their causes and take timely action. The opportunity was to combine traditional analytics with computer vision and generative AI to make store operations more intelligent.

02 / 05

My contribution

I led and coordinated AI product initiatives within a €4M AI transformation programme, connecting business needs, engineering capabilities and delivery teams, with a 13-person team.

03 / 05

The approach

• Full Monitored StoreDevelop computer vision use cases to monitor store operations and identify relevant events or anomalies through visual data.
• Manager CompanionIntroduce AI-assisted experiences that help store managers access operational information and make informed decisions.
• Stock Discrepancy IntelligenceUse data analysis to identify inventory discrepancies, improve visibility and support operational investigations.
• Knowledge Graph & LLM ComparatorInvestigate structured knowledge relationships and language-model comparison approaches to improve information access and AI solution choices.

04 / 05

Impact

The initiatives contributed to a broader approach to intelligent retail operations, combining several complementary capabilities: greater visibility into store operations and inventory discrepancies, improved access to contextual information for operational teams, a foundation for AI-assisted decision-making at store level and stronger alignment between business workflows, analytics and AI products.

05 / 05

What I learned

The value of retail AI comes from connecting intelligence directly to operational decisions, not from building isolated models.