$ cat jobs/product-director-ai-agentic-safelite-cd337038d363.json
Product Director, AI & Agentic
Does this position interest you? You should apply –even if you don’t match every single requirement! We're known as an auto glass company. That's the focus of what we do. But beyond the glass, we're so much more. We'll help you build a fulfilling career and encourage you to have a life. Let us be the best place you'll ever work. Role Summary We are building the next generation of how customers, clients, and colleagues interact with Safelite . Agentic AI is at the center of it. The Product Director, Agentic Experiences is the single-threaded owner of our agentic experiences across customer, client, and colleague audiences. You will turn vision and strategy into an executable roadmap, then run the build, the experimentation, and the performance that make these experiences real. This is the senior operator who makes the agents ship and improve. You run discovery and delivery day to day, you are deep in the tools and the data, and you lead a team of product managers who each own a space. You bring structure to ambiguous work and keep a portfolio of agentic experiences moving across audiences and channels. What You'll Do Own the agentic roadmap across customer, client, and colleague Translate vision and strategy into an executable roadmap and clear priorities across all three audiences. Run discovery and delivery day to day, from problem definition through launch and continuous improvement. Make tradeoff decisions based on customer impact, business value, effort, risk, and learning potential. Balance near-term performance, foundational platform work, and the next generation of agentic capability. Build and improve agents as living systems Scope, build, and ship agents that handle thousands of customer, client, and colleague conversations, and improve them in production rather than shipping and leaving them. Run the experimentation engine: deploy tests, simulate responses and outcomes, predict behavior, and find opportunities for agents to self-improve. Own how agent perf