Agents Don't Do Standups: Building the Post-Engineer Engineering Org — Mike Spitz, PFF

Integrating AI agents into engineering workflows significantly accelerates development cycles and eliminates traditional process bottlenecks. At PFF, shifting to an agent-first model resulted in a 25x increase in deployment frequency and a 10x improvement in output. By automating repetitive tasks like ticket creation, code reviews, and design documentation, the team replaced rigid Scrum ceremonies with lightweight, autonomous workflows. This transition requires a phased, cautious implementation starting with non-critical systems to ensure quality and maintain consistent product feel. Success depends on empowering curious engineers to leverage AI for deterministic tasks while offloading opinionated, low-value code reviews to agents. Ultimately, this approach transforms the engineering lifecycle into a composable, high-velocity factory, allowing teams to remain competitive by focusing human effort on high-level architecture and customer satisfaction rather than administrative overhead.