Cisco and OpenAI redefine enterprise engineering with Codex

For decades, Cisco has been at the forefront of building and operating complex, mission-critical software systems. As generative AI evolved, Cisco leveraged its expertise in scaling advanced technology within demanding real-world environments to integrate OpenAI's Codex into its core enterprise engineering. This move made AI-native development central to how enterprise software is constructed.

Cisco’s approach transformed Codex from a standalone developer tool into an AI engineering teammate capable of operating at enterprise scale. This was achieved by integrating Codex directly into production engineering workflows, exposing it to massive multi-repository systems, C/C++-heavy codebases, and stringent security and compliance requirements. This collaboration has significantly shaped how Cisco builds new products, exemplified by AI Defense, where Codex compressed critical engineering work from several quarters to mere weeks.

Applications of Codex have yielded remarkable results across Cisco's operations. For instance, in cross-repository build optimization, Codex analyzed complex dependency graphs across more than 15 interconnected repositories, leading to a 20% reduction in build times and saving over 1,500 engineering hours monthly. Another significant achievement is defect remediation at scale; using Codex-CLI, automated defect repair now takes hours instead of weeks, boosting throughput by 10-15 times. Furthermore, framework migrations that typically took weeks are now completed in days, allowing engineers to focus on higher-level decision-making.

Cisco’s continuous feedback from real-world production use has been instrumental in accelerating Codex’s readiness for large enterprises, particularly in adhering to compliance, managing long-running tasks, and integrating with existing development pipelines. This partnership also extends to advancing AI security, with Cisco actively participating in OpenAI’s Daybreak initiative, enhancing cyber defense and securing software continuously through models like GPT-5.5-Cyber.

This deep technical partnership, focused on real workloads and leadership alignment, has established a repeatable model for adopting next-generation AI within Cisco. Today, Codex is widely used across multiple business units, improving productivity and code quality. The effectiveness of this AI integration is now a key metric, with teams increasingly evaluating tasks based on "How long will that Codex run take?" rather than traditional effort estimations.