DesignOps in Practice: Building a customer insights pipeline

To bridge the structural gap between customer-facing operations and product design, I architected a DesignOps-led AI insights pipeline that transforms unstructured call transcripts into a prioritized, evidence-based roadmap. Moving away from high-friction manual forms that saw zero adoption, I instrumented existing workflows to capture raw customer signals without requiring any behavior change from internal teams. Using Claude Code, I built a custom synthesis engine that ingests thousands of recordings and deliberately separates "customer pain points" from "feature requests," empowering the design team to solve root problems rather than just reacting to the loudest voice in the room. This infrastructure has shifted our organizational culture from reactive shipping to defensible, informed design, ensuring that every new feature is grounded in verified customer reality.