How Mcdonald’s and JPMorgan chase safely scale experimentation

Описание: AI has more teams experimenting than ever, but 95% of AI pilots fall flat in the face of data silos and no safe way to fail fast. Most teams use feature flags to turn things on and off. The best teams use them to transform how they build, release, and learn.

In this interactive panel, Daniel MacDade, director of product management at McDonald’s, and Sorin Cioban, head of front-end engineering, at JPMorgan Chase & Co. join Harness’s Ryan Vila and tech journalist Jennifer Riggins to share how they scaled experimentation from a single toggle to a full-fledged delivery discipline. Together, they’ll unpack the real practices, both technical and business, technical and business metrics, and cultural shifts that helped them ship features up to 50x faster, release confidently, and make every deployment a data-driven decision.

You’ll walk away lessons learned from real-world enterprise case studies on turning feature management into a strategic advantage — not just a safety net.

In this session, you’ll learn how to:

Expanding who can experiment: In times of rapid change, how to enable developers to safely experiment, even without being a site reliability engineer and experimentation expert.

Accelerate Delivery: Automate rollouts, eliminate downtime, and safely push code to production faster than ever.

De-risk Deployments: Use progressive delivery, with canary rollouts, user segmentation, and blue-green deployments.

Prove Impact with Data: Measure feature performance with A/B and multivariate experiments that validate real business value before full release.

Govern with Confidence: Apply guardrails, auditability, and consistent flag management across services, teams, and environments.