Cisco used OpenAI's Codex to write the majority of AI Defense, its end-to-end enterprise AI security product, and nearly every new feature currently in development. The result: feature delivery dropped from several quarters to weeks, according to DJ Sampath.
That compression matters. Enterprise security software is not a fast-moving category by design. Cutting delivery time by that magnitude while maintaining production quality is the operational claim worth interrogating in the full piece.
The linked article goes deeper on how Cisco is structuring AI-native development as a permanent part of its engineering process, not a one-off experiment. If you work in enterprise software or security infrastructure, the process details are what you want.
[WATCH ON YOUTUBE →]