Cisco Engineer Sees Shift in Software Development with Agentic AI

Sergio Freitas, a senior director of engineering at Cisco, manages between 10 to 20 agents daily who oversee their own agents. He notes that managing these AI-powered agents does not mean longer working hours but can feel more intense.

The introduction of agentic AI has significantly altered the pace and nature of work for software engineers. Freitas attributes this shift in part to the launch of frontier models such as Anthropic's Claude Opus 4.5 at the end of last year. This marked a turning point where AI began to be deeply integrated into workflows, allowing for delegation and collaboration with agents.

The role of software engineers has transformed from manual coding to focusing on architecture, design, and workflow orchestration. Engineers now work alongside agentic AI, which can spawn sub-agents to complete tasks based on the complexity of the project. These agents are constantly running and aware of various integrations when granted explicit permission.

Cisco's move into loop engineering has also seen a change in how agents interact with humans. Instead of asking questions, they now loop or continue working until receiving final responses. While this brings more value to collaboration, it also requires engineers to remain vigilant and review the code produced by these AI tools.