Engineering for the Agentic Era: How Braze’s CTO Led a Rapid AI-First Transformation

From Startup to AI-First Powerhouse: A 15-Year Engineering Journey

For nearly fifteen years, Jon Hyman has served as the co-founder and CTO of Braze, the customer engagement platform. Over that period, he has steered the company’s engineering organization through explosive growth, from a small startup to a publicly traded enterprise with hundreds of engineers. But perhaps the most remarkable pivot came recently: a complete reorientation toward an AI-first mindset—achieved in just a matter of months. In this article, we explore the principles, strategies, and cultural shifts that defined this transformation and what it means for engineering teams entering the so-called “agentic era.”

The Core Philosophy: Engineering as a Strategic Advantage

Hyman’s approach to engineering leadership has always prioritized long-term scalability. Early on, he invested heavily in infrastructure and platform teams to prevent technical debt from slowing innovation. As Braze grew, he maintained a flat organizational structure with clear ownership boundaries, allowing autonomous teams to move quickly while staying aligned with company goals.

Key Structural Choices

• Platform teams that own shared services (messaging, data pipelines, APIs) to reduce duplication.
• Product-aligned squads that focus on customer-facing features while leveraging the platform.
• Strong DevOps culture with continuous delivery and observability from day one.

These decisions created a resilient foundation that could absorb rapid change—like the AI pivot—without breaking existing workflows.

The AI-First Pivot: From Months to a New Mindset

When generative AI and agentic architectures began to reshape the technology landscape, Hyman recognized that incremental updates would not be enough. He challenged his team to become AI-first—not just adding AI features, but rethinking how engineering itself operates.

What “AI-First” Means at Braze

For Hyman, an AI-first engineering organization is one where machine learning models and agentic systems are treated as first-class citizens—integrated into the development pipeline, deployment processes, and product architecture. This required:

1. Investing in ML infrastructure – Building feature stores, model registries, and A/B testing frameworks specifically for AI.
2. Upskilling every engineer – Not just data scientists, but backend and frontend engineers learned prompt engineering, model evaluation, and agent orchestration.
3. Shifting from deterministic to probabilistic thinking – Acceptance that AI-powered features have confidence scores and may need fallback logic.

The Speed of Transformation

What seemed like a multi-year journey was compressed into a few months. Hyman credits this to the preexisting culture of experimentation and the platform team’s ability to rapidly prototype new services. Internal hackathons and “AI sprints” allowed teams to ship early prototypes and iterate based on real customer feedback.

Preparing for the Agentic Era

Braze’s engineering organization is now looking ahead to what Hyman calls the “agentic era”—where autonomous agents execute complex tasks on behalf of users. Instead of rigid workflows, agents can plan, reason, and take actions across multiple systems.

Architectural Implications

To support agentic behavior, Braze has begun redesigning its APIs and event streams to be agent-friendly. Key changes include:

• Event-driven actions – Agents subscribe to customer events and trigger personalized campaigns in real time.
• Context-rich prompts – APIs now pass conversation history, user preferences, and business rules so agents can make informed decisions.
• Human-in-the-loop guardrails – For high-stakes actions, agents can request approval before executing.

Cultural Shifts for an Agentic Workforce

Hyman emphasizes that engineers themselves must evolve. He encourages teams to think of their code as “agent training data”—clean, well-documented, and predictable. Additionally, engineers now regularly collaborate with product managers to define the boundaries of agent autonomy, ensuring that Braze’s platform remains trustworthy and compliant.

Lessons for Other Engineering Leaders

Hyman believes that any engineering organization can replicate Braze’s transformation, but it requires deliberate effort. His advice:

1. Start with a strong platform – You cannot move fast on shaky foundations. Invest in infrastructure before adding AI.
2. Democratize AI knowledge – Avoid creating an elite “AI team.” Instead, train everyone so that AI thinking permeates all engineering decisions.
3. Embrace failures as learning – Not every agentic prototype will succeed. Create psychological safety for experimentation.
4. Think about ethics from day one – Agentic systems can amplify biases; build fairness and transparency into the design process.

Conclusion: The Future of Engineering at Braze

Jon Hyman’s journey from startup CTO to leader of an AI-first engineering organization offers a blueprint for teams navigating the agentic era. By maintaining a strong technical foundation, rapidly upskilling engineers, and embracing probabilistic architectures, Braze has positioned itself to deliver next-generation customer experiences. The transformation was not easy—but it proves that with the right culture and leadership, even a fifteen-year-old company can reinvent itself in months.

For more insights on engineering leadership and AI transformation, explore our other articles on platform engineering and agentic architecture.