Cisco and OpenAI redefine enterprise engineering with Codex

Cisco has integrated OpenAI’s Codex into its production engineering workflows, turning the model into an AI‑native development teammate that operates at enterprise scale. Since deployment, Codex has written over 95 % of new AI features, automated defect remediation with its CLI, and boosted defect resolution throughput by 10‑15×. The company reports that Codex‑driven cross‑repo build optimization cut build times by roughly 20 % and saved more than 1,500 engineering hours per month across its global environments. Codex also accelerated framework migrations, compressing weeks of React 18→19 work into days and enabling autonomous compile‑test‑fix loops in large C/C++ codebases.

Cisco’s AI Defense product, built almost entirely with Codex, demonstrates the model’s ability to handle safety and security risks, while the open‑source Defense Squad was moved from ideation to community release in under a week. By providing continuous feedback to OpenAI, Cisco helped shape Codex’s compliance, long‑running task management, and pipeline integration features for large‑scale enterprises. The collaboration has become a repeatable model for adopting next‑generation AI, with Codex now used across multiple Cisco business units to improve productivity, code quality, and time‑to‑resolution. Engineers now treat Codex as a core part of their workflow, generating plan documents and code that reviewers can easily understand and validate.