Netomi’s lessons for scaling agentic systems into the enterprise

Source: OpenAI Date: 2026-01-08 URL: https://openai.com/index/netomi

Summary

Case study from January 2026 covering Netomi, an AI customer service platform, and their experience scaling agentic AI systems into enterprise customer service deployments. Netomi specializes in autonomous customer service — AI agents that resolve customer issues without human escalation for the majority of contacts. The “lessons” framing suggests this is less a product announcement and more a practitioner reflection on what works and what fails when deploying agentic systems at enterprise scale.

Implications

Agentic customer service as a matured use case. By January 2026, autonomous AI customer service wasn’t an experiment — Netomi was operating at enterprise scale with measurable resolution rates. The “lessons” post is practitioner wisdom from a company that had been building this for years. The insights on failure modes, escalation design, and quality control are more valuable than any launch announcement.

Resolution rate vs. customer satisfaction tension. Agentic customer service optimizes for autonomous resolution; customers care about satisfaction. These can diverge: an AI that “resolves” an issue by closing the ticket without actually solving the problem is technically a win for resolution rate but a loss for the customer. Netomi’s lessons likely address this tension, which is the central quality control challenge in autonomous customer service.

Thread: enterprise agentic deployment. Sits alongside the EliseAI case study, the SafetyKit deployment, and the ChatGPT Agent system card as markers of where agentic AI is actually deployed at scale vs. where it’s being piloted. Netomi’s at-scale deployment is more valuable as a signal than any pilot announcement.

Watch: Whether Netomi’s autonomous resolution rates continue improving as underlying models improve, and whether customer satisfaction scores track with or diverge from resolution rate gains.