Decathlon: Integrating AI into Business Strategy and Process

Move People Through the Wonder of Sport Our Story

● Michel LECLERCQ surrounds himself with 6 sportspeople to create a new concept: A store entirely dedicated to all sports under one roof, in Englos, near Lille ● Opening of our first store outside of France, in Dortmund, in Germany ● Opening first store in the Netherlands, Amsterdam Arena Store ● Launching of our first iconic product « 2 seconds tent » A product that symbolizes the innovation process throughout the company 1976 1986 2000 2005 Our Story

103138 passionate teammates 82 nationalities 1709 stores in 61 countries 142M customers & 400M users of our products Move People Through the Wonder of Sport Our Story

1709 stores in 61 countries Today 21 1.500+ Our Story

Decathlon NL Data Team ● 3 Data Scientists ● 3 Data Engineers ● 4 Analysts Decathlon United (HQ) ● 145+ Data Team Decathlon Network ● 61 Data Teams Our Structure

Our Goals for AI Integration AI Driven Efficiency, Innovation and Customer Experience

➔ Bring data into daily work life ➔ Increase efficiency and accelerate decision making ➔ Make the complex simple ➔ Understand our Customers and Consumers ➔ Anticipate and reduce risks ➔ Gain more out of the given insights ➔ Provide predictive, proactive solutions ➔ Support and drive business innovation Our Goals for AI Integration

Barriers to Integration Alignment with Current Strategy Dispersed Skill Sets Executive Buy In Adoption Legal Clarity

Our Projects

Challenge 1 … An employee spend on average 2 hours* per day just searching and gathering for information ● 50K hours per 100 employee in a year ● At least €1M within a year, money spent without much value being added The solution … ● AI Powered Knowledge Base ● Use an external partner for the implementation *McKinsey Our Projects

AI Powered Knowledge Base ● Implement an internal knowledge base ● Reduce the time spent on searching & gathering information by at least a factor of 2 ● Savings of around €0.5M per 100 employee per year

Challenge 2 … A repair employee spend on average 30 minutes on assessing and pricing a buy back (second hand) bike ● 250 hours equivalent to 33 FTE to assess 500 bikes a day ● At least €1M cost per year ● Not scalable & Non-standardized pricing The solution … ● Machine Vision For Automated Damage Assessment ● Implement an in-house solution Our Projects

Machine Vision For Automated Damage Assessment ● Machine vision model for bike assessment ● Reduce the assessment time by a factor of 6 ● Enable the circular business to scale up ● Savings of at least €1M per 100K bikes assessed per year ● Improved Customer Experience

Our Projects Challenge 3 … There are nearly 4 fraud cases in online retail per 100 inhabitants per year in the Netherlands* ● ~20% come from return, ~15% through gift cards … etc. ** ● Tens of thousands Euros in losses The solution … ● AI Based Fraud Detection System ● Implement an in-house solution *Statista **Digital Commerce 360

AI Based Fraud Detection System ● Implement pattern recognition model to detect anomalous behaviours in transactions ● Setup alerts when a fraud pattern is detected ● Set action plans based on fraud categorization

Barriers to Integration Alignment with Current Strategy Dispersed Skill Sets Executive Buy In Adoption Legal Clarity Projects Support existing initiatives Foster Collaboration Demonstrate Value Embedded in Process Approved

THANK YOU THANK YOU ANY QUESTIONS?