Professional Services|Product Engineering|Increase Efficiency

CapgeminiCode generation

Privacy rules and poor accuracy capped AI use at 30%. A self-hosted model hit 90% precision, driving adoption to 100%.

Nov 3, 2025|3 months ago

Key results

Client Projects
50+

The company

Capgemini logo

Capgemini

capgemini.com

IT consulting and digital transformation services for global enterprises.

IndustryProfessional Services
LocationParis, France
Employees100K+
Founded1967

Result highlights

  • Productivity gains across 50+ client projects

The story

A global system integrator with 350,000 employees across 40+ countries serves clients in highly regulated sectors like aerospace, defense, and the public sector.

Strict data sovereignty rules prevented the use of cloud-based AI, while initial open-source attempts stalled at 30% adoption due to poor accuracy. Developers managing legacy applications from the 1980s found these tools insufficient for critical languages like Ada and C++.

Capgemini deployed a self-hosted engineering assistant powered by Mistral's Codestral model to satisfy privacy mandates. The team swapped the underlying model into existing IDE and CI/CD pipelines without disrupting established workflows or infrastructure. Extensive developer coaching accompanied the technical rollout to ensure effectiveness across both modern and legacy codebases.

Scope & timeline

  • Developer adoption increased from 30% to 100%

Quotes

Explore similar

Find AI opportunities for your
business context

Understand what's working with 2,383 recent AI case studies across industries. We structure things so you can find high-impact strategies for your exact context.

Graphic placeholder

Industries covered