Application Refactoring · IT Operations Automation · Security and Compliance
Cisco needed an AI engineering teammate that could operate across large multi-repository codebases, enterprise governance controls, and mission-critical software programs without sacrificing security or review discipline.
Reported impact
Cisco used Codex to accelerate AI Defense delivery, save more than 1,500 engineering hours a month, reduce build times and defect-resolution effort, and institutionalize AI-assisted development inside a global enterprise engineering organization.
Cisco embedded Codex directly into enterprise engineering workflows instead of treating it as a standalone assistant. The company used Codex to write most of AI Defense and nearly every new feature in the product, turning work that once took quarters into weeks. Beyond product delivery, Codex analyzes build logs and dependency graphs across more than 15 interconnected repositories, automates defect remediation on large C/C++ codebases through iterative compile-test-fix loops, and speeds framework migrations such as large React upgrades. Cisco also helped shape Codex for enterprise deployment by providing feedback on workflow orchestration, security controls, and long-running task support, and it participates in the Daybreak initiative with governed access to GPT-5.5-Cyber for defensive use cases. The result is an AI-native development model where Codex acts inside existing review, security, and governance frameworks while teams increasingly size work by how long a Codex run will take.
Code, build logs, dependency graphs, and defect reports flow into Codex and Codex-CLI, which generate plans, propose code, run tests, and iterate fixes inside Cisco's existing review and governance systems. The same enterprise feedback loop informs AI Defense, Defense Squad, and cyber workflows using GPT-5.5-Cyber.
"Codex was used by Cisco's team to write the majority of AI Defense"
"Features that would have taken several quarters to get into customers' hands dropped to weeks."
"What once took weeks of manual effort now completes in hours, delivering a 10-15× increase in defect resolution throughput"
"more than 1,500 engineering hours saved per month"
Review how this reference was scored and trace each normalized claim back to its documented source.
Scoring methodology
Customer Reference Value measures usefulness as a customer reference. Evidence Confidence measures the quality and traceability of available evidence, not independent verification.
Traceability
Field coverage shows which facts each source supports. Dates are shown only when documented by the source.
Related
OpenAI Sample Sep 8, 2026 1Password 1Password needed to compress the time between planning and production, reduce investigation and review overhead, and expand builder capacity without weakening the company's zero-knowledge security posture. $0.8M estimated annual engineering capacity value from Codex AI site reliability engineering (SRE) agent · AppSec harness · Codex OpenAI Sample Mar 11, 2026 Rakuten Rakuten needed to ship faster, recover from incidents faster, and increase autonomous execution across a large engineering organization without relaxing security or reliability standards. 3-4x faster potential build time for projects, from quarters to weeks AI agent service · CI/CD workflow · Codex 
AWS Sample Sep 18, 2023 CyberGRX The legacy ML modeling solution for CyberGRX took 8 days to complete predictive modeling for all the companies in its database — 150,000 at the time. Four engineers were required to monitor and manage it, which added… CyberGRX reduced the runtime of its solution from 8 days to 1 hour for its quarterly modeling, improved staff… AWS Lambda · AWS Step Functions · AWS Step Functions Distributed Map 