AI at Scale: How NTT DATA Cut a 3-Day Job to 30 Minutes with Codex

📅 July 25, 2026

🏷️ 🚀 Enterprise AI Case Study

⏱️ 60-Second Summary

Japan's IT services giant NTT DATA Group expanded OpenAI's coding agent, Codex, to roughly 9,000 employees across technical and non-technical teams. The headline result: Codex completed an incident analysis in 30 minutes that had previously taken five engineers three full days — a 99.3% time reduction. Driven by NTT DATA's "Client Zero" philosophy of internal AI adoption before client rollout, non-technical staff now use Codex for expense report extraction, self-service data reporting, and task automation.

🤔 Why This Matters

While many corporate AI efforts remain confined to basic chat interfaces or narrow developer pilots, NTT DATA presents a blueprint for operationalizing agentic AI across a large, non-tech-native enterprise. By establishing a central Center of Excellence (CoE) to handle governance, security boundaries, and training first, NTT DATA scaled Codex safely to 9,000 staff, demonstrating how AI transitions from a novelty tool to enterprise-wide core infrastructure.

👨‍💻 Engineering Impact

• 99.3% Time Reduction: Complex root-cause incident analyses reduced from 15 engineer-days (5 engineers × 3 days) down to 30 minutes.
• Developer Workload Relief: Domain experts can build their own reports and data automation scripts, freeing developers to focus on core product architecture.
• Governance First: Clear guidelines on system connectivity, data touchpoints, and human-in-the-loop review criteria ensure safe autonomous execution.

🌍 Non-Technical Team Use Cases

• Expense Automation: Pulling transit expenses from credit card statements directly into travel reimbursement forms.
• No-Code Reporting: Transforming raw data into insights without needing business intelligence tools or data engineering queues.
• Workflow Scripting: Automating file management, document summarization, and routine administrative tasks.

🔥 Key Takeaways for Any Org

1. Everyday Chat Prepares Teams for Agents: ChatGPT habituation makes Codex adoption seamless.
2. Broad Rollout Drives Organic Adoption: Peer word-of-mouth creates momentum across departments.
3. Governance Must Scale Early: Establish security boundaries before expanding access.
4. Treat Launch Day as Day One: Continuously refine workflows using telemetry and employee feedback.
5. Centralize Best Practices: Use a central CoE to curate playbooks for team-wide replication.

🔗 Official Source

• OpenAI Customer Stories — NTT DATA Group