Cisco and OpenAI Redefine Enterprise Engineering with AI Agents

Cisco is using OpenAI's Codex not merely as a development tool, but as an AI engineering team member operating in complex production environments. Through this, Cisco has realized enterprise-grade AI capabilities even in large-scale multi-repository systems and C/C++-centric codebases.

Codex delivers value beyond existing code completion features through the following Agentic AI capabilities:

• Understanding and reasoning across large, interconnected repositories
• Proficient handling of complex programming languages
• Autonomous CLI-based compile-test-fix loop execution
• Operating within existing review, security, and governance frameworks

Applying this to real workflows has confirmed concrete results. Build optimization reduced build times by approximately 20% and saved over 1,500 engineering hours per month, while the defect resolution (CodeWatch) process saw throughput improve 10-15x. Additionally, the Splunk team's React 18 to 19 migration work was shortened from weeks to days.

The collaboration between Cisco and OpenAI goes beyond simple technology adoption, presenting a repeatable model for adopting next-generation AI through close technical partnership with real workloads.