How Braze’s CTO Engineered an AI-First Culture in Record Time

From Startup to Enterprise: A 15-Year Engineering Journey

For nearly 15 years, Jon Hyman, co-founder and CTO of Braze, has steered the company's engineering organization through explosive growth. What began as a small startup team has evolved into a powerhouse responsible for a customer engagement platform that processes billions of data points daily. Hyman's leadership philosophy has always centered on scalable architecture and developer productivity, but the biggest shift came recently: a complete transformation into an AI-first team—accomplished in just a few months.

Building for Scale

Under Hyman's guidance, Braze's engineering team prioritized infrastructure that could handle massive surges in messaging volume without compromising reliability. They adopted microservices, invested in observability, and cultivated a culture of continuous deployment. These foundational decisions made the later AI pivot possible by ensuring the platform could rapidly integrate new machine learning models and data pipelines.

The Pivot to AI-First Engineering

As the industry entered what Hyman calls the "agentic era"—where autonomous AI agents take on complex tasks—Braze recognized the need to reorient its engineering team. Instead of treating AI as a separate feature, Hyman embedded it into the core of how the team operates and builds.

Rapid Transformation in Months

The overhaul involved three key changes:

• Retraining engineers in prompt engineering, model fine-tuning, and AI safety—not as specialists, but as part of their daily workflow.
• Redesigning the platform to expose AI capabilities as first-class APIs, allowing customers to leverage predictive personalization and autonomous campaign optimization.
• Shifting team structures to create cross-functional pods that pair machine learning experts with product engineers, ensuring AI solutions directly address user needs.

Hyman notes that the transformation succeeded because of leadership commitment and a willingness to fail fast. The team launched experimental AI features early, gathered feedback, and iterated rapidly—a practice Hyman calls "agentic iteration."

Lessons for Engineering Leaders

Hyman's experience offers a playbook for other CTOs looking to embrace the agentic era. He emphasizes four takeaways:

1. Start with infrastructure: Ensure your platform can handle compute-heavy workloads and real-time data streams before layering on AI.
2. Invest in education: Every engineer should understand AI basics, not just the ML team. Hyman ran internal bootcamps and hands-on hackathons.
3. Measure differently: Traditional uptime and latency metrics remain important, but Hyman tracks AI success rate—the percentage of autonomous actions that meet customer goals without human intervention.
4. Keep humans in the loop: Even in an agentic system, Hyman believes engineers must retain oversight. His team builds guardrails and "escape hatches" to let humans override AI decisions when needed.

Jump to Lessons

Looking Ahead

Braze's engineering transformation is ongoing. Hyman sees a future where agents not only execute tasks but also contribute to system design and code generation. He predicts that within five years, Braze's platform will handle 80% of customer engagement decisions autonomously, freeing engineers to focus on new creative problems. For now, the lessons from Braze's rapid shift stand as a testament to what a focused, AI-first culture can achieve.