new AI features written by Codex
new AI features written by Codex: 95 percentEvidence strength: reported
Enterprise AI Evidence · Technology & Software
Cisco
Cisco deployed Codex broadly across enterprise engineering and AI-defense workflows.
Reported outcomes
defect resolution throughput increase
defect resolution throughput increase: 10 xEvidence strength: reported
engineering hours saved monthly
engineering hours saved monthly: 1,500 hoursEvidence strength: reported
Mapped use case
Repository coding agent
Applied AI can augment engineering work by reducing cycle time, increasing coverage and improving decision consistency while retaining explicit controls for high-consequence decisions.
Related evidence in Technology & Software
NVIDIA
NVIDIA engineers and researchers use Codex for complex engineering tasks and end-to-end machine-learning experiments.
Examine evidenceHume AI
Hume AI uses Claude in emotionally intelligent voice experiences across healthcare, customer service and consumer applications.
Examine evidenceRakuten
Rakuten uses Claude Code to accelerate feature development.
Examine evidenceDovetail
Dovetail uses Amazon Bedrock to build generative AI features for analysis of customer research and unstructured feedback.
Examine evidenceASAPP
ASAPP uses Amazon Bedrock and multiple models in an AI-native customer-service platform for complex voice and chat interactions.
Examine evidence
From intelligence to institutional decision
Test whether this pattern survives your institutional reality.
Use the evidence as a starting point for a controlled assessment, not as a substitute for your own baseline, risk, economics and operational validation.