NTT DATA Cuts Incident Analysis from Days to Minutes Using Codex

NTT DATA Group has implemented Codex, OpenAI's code-fluent model, to automate the analysis of complex system incidents. The move represents a significant shift in how the enterprise manages technical failures, replacing manual engineering marathons with AI-driven diagnostics.

According to confirmed reports, the implementation has drastically reduced the time required for incident analysis and root cause identification. Previously, resolving these complex system issues typically required a team of five engineers working over a period of three days. By utilizing Codex via ChatGPT Enterprise, NTT DATA has compressed that timeline to approximately 30 minutes.

The Shift to AI-Driven Diagnostics

This transition comes as global IT service providers face increasing pressure to maintain uptime for massive, interconnected digital infrastructures. Traditionally, root cause analysis (RCA) in large-scale systems is a labor-intensive process involving the manual review of logs, code traces, and system dependencies. By integrating a model specifically trained on code and technical documentation, NTT DATA can now synthesize these data points at a speed impossible for human teams.

Industry Implications

The reduction of a three-day process to a half-hour window has profound implications for operational overhead and service level agreements (SLAs). For data-heavy enterprises, the ability to identify the source of a system failure in minutes rather than days minimizes downtime and prevents the cascading effects of prolonged outages. Furthermore, it frees high-level engineering talent from the repetitive task of log parsing, allowing them to focus on permanent architectural fixes rather than immediate firefighting.

Future Outlook

While the current application focuses on incident analysis, the success of the Codex implementation suggests a broader trend toward the automation of specialized technical workflows. Industry observers will be watching to see if NTT DATA expands this AI integration into predictive maintenance—identifying potential system failures before they occur—or if similar models are deployed across other areas of its global IT operations. It remains to be seen how this shift will impact the long-term staffing requirements for traditional site reliability engineering roles.