FranceRetail and e-commerce2023 – ongoing

Cdiscount

GitHub Copilot, product sheet recategorization and pre-sales chatbot

Cdiscount recategorized 23 million product files with AI and gave Copilot to about 500 tech staff

WorkingVendor's case studyVendor ↗

Problem

The French marketplace lists 70 million products, and many sheets sat in the wrong categories. Developers also spent time on routine code.

Approach

Its tech arm Peaksys gave GitHub Copilot to about 500 tech staff. The company combined in-house models with Azure OpenAI Service LLMs to recategorize product sheets, and added a pre-sales chatbot via partner iAdvize.

Results

23 million of 66 millionProduct files recategorized after ten monthsVendor's case study[1]

half a working day a weekAverage time saved per developer using GitHub CopilotVendor's case study[1]

70%Customer satisfaction rate for the pre-sales chatbotVendor's case study[1]

Takeaway

Ten months in, 23 million of 66 million files had been recategorized, and the recategorized files converted at 30%. Microsoft reported developers saving about half a working day a week.

Sources

Cdiscount customer story (Azure OpenAI Service)
Microsoft4 Jul 2024Vendor publication

Source summaries

[1]Microsoft4 Jul 2024

Microsoft reports three AI uses at Cdiscount: coding help, product recategorization and a chatbot. It lists time saved, files recategorized and satisfaction.

Full details are in the original sources.

Added 25 Oct 2026 · Updated 25 Oct 2026