Cisco engineering director explains how AI agents reshaped software development workflow

In a recent interview, Cisco senior director of engineering Sergio Freitas described a rapid transformation in his team's workflow following the launch of frontier AI models such as Anthropic's Claude Opus 4.5. The company moved from using AI for simple autocomplete to deploying autonomous agents that can write code, manage GitHub pull requests, and respond to Jira tickets. Freitas now oversees between ten and twenty agents daily, with a primary orchestrator spawning sub-agents for specific tasks.

Although the agents can act independently, they operate under strict guardrails and all code must be reviewed by at least two engineers before reaching production, keeping ultimate liability with humans. The shift has accelerated project speed, introduced higher multitasking demands, and required engineers to develop new mental habits, but it has not increased overall work hours. Security scans and legal reviews remain integral parts of the process.