Leading Engineering in the Age of AI Agents: Braze's CTO on Transformation

Jon Hyman, co-founder and CTO of Braze, has steered the company's engineering organization through nearly 15 years of rapid growth. In this Q&A, we explore how he guided the team’s evolution into an AI-first powerhouse in just a few months, reshaping the engineering culture to thrive in the agentic era—where autonomous AI agents increasingly handle complex tasks.

How did Braze’s engineering team evolve over 15 years?

Braze started as a small startup with a handful of engineers focused on building a customer engagement platform. Over the years, Jon Hyman oversaw a deliberate scaling process: from a tight-knit team to a global engineering organization of hundreds. The key was maintaining a strong engineering culture while adapting to new challenges—moving from monolithic codebases to microservices, adopting cloud-native infrastructure, and integrating real-time data pipelines. Hyman emphasized hiring for adaptability and fostering cross-functional collaboration. As the company grew, the engineering team shifted from simply shipping features to optimizing for reliability, performance, and developer velocity. This long-term evolution set the stage for a more radical transformation: embracing AI as a core part of their engineering DNA.

Leading Engineering in the Age of AI Agents: Braze's CTO on Transformation — Source: stackoverflow.blog

What prompted the shift to an AI-first team?

The trigger was a combination of market signals and internal vision. Hyman observed that customer expectations were changing—they wanted smarter, more personalized engagement that could adapt in real time. Meanwhile, advances in machine learning and large language models made it feasible to embed intelligence directly into engineering workflows. Braze's leadership saw that waiting would mean falling behind. The decision to go AI-first came from the top: Hyman and the executive team committed to rethinking every part of the engineering stack. They wanted to move beyond using AI as an add-on feature; instead, AI would become the foundation for how engineers write code, test software, debug issues, and even manage infrastructure. The urgency was driven by the rise of agentic AI—systems that can autonomously plan and execute tasks—which promised to revolutionize software development.

What were the key steps in transforming to AI-first?

The transformation happened in a few intense months. First, Hyman formed a small AI task force to experiment with tools like ChatGPT, GitHub Copilot, and internal models. They ran hackathons to find high-impact use cases: automated code generation, intelligent code review, and predictive testing. Next, they established an AI infrastructure team to build MLOps pipelines, secure data access, and create guardrails for responsible AI use. Hyman also launched a company-wide training program, requiring every engineer to complete an AI literacy course. The most impactful step was integrating AI assistants into the daily workflow—every pull request now included AI-generated suggestions, and debugging sessions involved AI copilots. Within 90 days, most teams had adopted AI tools, and Braze began measuring productivity gains in real metrics like deployment frequency and bug resolution time.

How did the CTO lead the cultural change?

Hyman led by example. He personally used AI tools in his own code reviews and design discussions, showing that it wasn't just for junior engineers. He held town halls to address fears about job displacement, framing AI as a superpower that lets engineers focus on creative problem-solving. He also changed performance reviews to reward learning and experimentation, not just output. One pivotal move was creating a monthly “AI Showcase” where teams demonstrated their AI-driven innovations. This fostered a culture of curiosity rather than resistance. Hyman also established an AI ethics board with rotating engineering members to ensure transparency and fairness. He emphasized that transformation is a journey, not a destination—encouraging continuous upskilling and cross-team knowledge sharing. Within six months, the engineering team’s mindset shifted from “AI is interesting” to “AI is essential.”

What challenges did they face?

The biggest challenges were lack of expertise and tool fragmentation. Many senior engineers were skeptical about AI reliability, especially for production-critical code. There were also data privacy concerns: how to use customer data to train models without compromising security. Hyman’s team invested heavily in sandboxed environments and data anonymization to build trust. Another challenge was preventing “AI debt”—over-reliance on generated code that might not be maintainable. They solved this by enforcing strict code review policies for AI contributions. Finally, integrating diverse AI tools into a coherent workflow required significant engineering effort. The team built internal abstractions that unified multiple AI services (LLMs, code analysis, testing) into one platform. Through iterative improvements and transparent communication, they turned these obstacles into learning opportunities.

What advice does the CTO have for other engineering leaders?

Hyman advises leaders to start small but think big. Instead of a massive overhaul, choose one pain point in the engineering lifecycle—like code review or documentation—and pilot an AI solution. Measure the impact quantitatively and qualitatively. Second, invest in people: create a safe space for experimentation and pair experienced engineers with AI-curious newcomers. Third, don’t ignore infrastructure: a solid MLOps pipeline and data governance are prerequisites for scaling AI. Fourth, communicate the vision relentlessly. Explain how AI changes the role of engineers—from writing repetitive code to designing intelligent systems. Finally, be patient with the learning curve. The transformation won’t happen overnight, but the compounding benefits of an AI-first culture are immense. Hyman believes the agentic era is not about replacing engineers, but about empowering them to build more ambitious products.

How does this transformation align with the 'agentic era'?

The agentic era is defined by autonomous AI agents that can plan, execute, and learn from tasks. Braze’s AI-first engineering transformation directly prepares the team for this future. By integrating AI agents into their development lifecycle—such as agents that automatically triage bugs, suggest code refactors, or monitor system performance—engineers can focus on higher-level architecture and innovation. Hyman envisions a near future where each developer has a personal agent that handles routine coding, testing, and deployment, while the human oversees strategic decisions. This shift requires a new engineering mindset: building systems that are resilient to agentic failures, designing APIs that agents can call, and creating feedback loops for continuous improvement. Braze’s early adoption positions them to thrive in a landscape where software development becomes a partnership between humans and autonomous agents.

What are the results of the AI-first approach?

The results have been dramatic. Developer productivity—measured by lines of code shipped per day—increased by roughly 40%. Bug resolution time dropped by 30% because AI copilots quickly identified root causes. Code review cycles shortened from days to hours. More importantly, engineers reported higher job satisfaction: they spent less time on mundane tasks and more on creative problem-solving. Braze also saw faster time-to-market for new features, enabling them to respond more nimbly to customer requests. Perhaps the most significant outcome was a cultural shift: the engineering team became proactive in experimenting with new AI tools. Hyman notes that the transformation is still in its early days, but the momentum has made Braze an attractive destination for top engineering talent who want to work on the frontier of AI-driven development.

Tags: