Synopsis
This talk shares our experience at Qonto integrating AI into a large engineering team of 120 developers. It covers the strategy we took, the experiments we ran, what actually worked (and what didn’t), and how we measured real impact across the org without slowing delivery.

Abstract:
“Start using AI” sounds simple. But getting engineers to actually integrate it into their daily work, especially in a fast-paced environment with constant change, is a whole different story.

At Qonto, I lead part of a 120-person engineering org. Over the past year, we’ve worked on making AI a useful, repeatable part of how we ship software without disrupting product delivery or adding overhead.

This talk is a look behind the curtain:
- The strategy that helped us stay flexible instead of locking into one tool or vendor
- The hands-on experiments we tried and what we learned, both the wins and the dead ends
- How we kept a balance between improving developer experience and hitting delivery goals
- How we tracked adoption, measured time saved, and identified where AI actually helped
- What we’re testing next as the AI landscape continues to shift

It’s a practical, real-world perspective on bringing AI into engineering. Not from a vendor, not from a research lab, but from the ground floor of a product team under pressure.

Takeaways:
- A strategy for introducing AI into a large, fast-moving engineering org
- Examples of real use cases that drove value, and ones that didn’t
- Tips for getting buy-in from delivery-focused teams
- Lessons from failed experiments and how we adjusted
- How to think about measuring success beyond just tool usage