DXC and LangChain partner to scale governed agentic AI across enterprise IT operations

DXC Technology and LangChain have announced a partnership to advance the use of governed agentic AI across mission-critical enterprise IT environments.

The collaboration builds on DXC's use of LangGraph, LangChain's open-source agent orchestration framework, within DXC OASIS, its intelligent orchestration platform for IT operations. DXC is also licensing LangSmith to support agent development, evaluation and observability.

DXC OASIS is already in production and is scaling towards 125 customer environments across regulated industries and multiple geographies. The partnership aims to help enterprises deploy AI agents while maintaining the governance, human oversight and operational controls required for critical systems.

Faster resolution across IT operations

DXC OASIS uses AI agents to respond to operational events, investigate issues in parallel and coordinate specialised capabilities across technology domains. Human experts remain involved when judgement, approval or intervention is required.

According to DXC, the platform has reduced resolution time by an average of 95% across more than 500 measured use cases, with diagnostic accuracy of 92% or better when validated against senior engineer judgement.

The company has also reported improvements across specific operational scenarios. Filesystem incidents that previously took 53 to 90 minutes can now be resolved in three to six minutes. Database performance incidents have moved from two to three hours to five to 12 minutes, while complex problem management cases that previously required two to six days have been reduced to eight to 10 minutes.

Dan Gray, president of Global Infrastructure Services at DXC Technology, said agentic AI could transform how enterprises operate critical technology, but added that models alone are not enough. He highlighted the importance of combining agent technology with operational expertise and governance to support enterprise-scale deployment.

Building governance into AI agents

A central focus of the partnership is enabling AI agents to operate within defined boundaries while preserving accountability and human oversight.

Through LangChain's technology, DXC OASIS assesses actions according to factors including reversibility, potential impact, verifiability and an agent's track record. Sensitive actions can be paused for human approval before execution.

The platform also tracks agent decisions, tool calls and outcomes, providing visibility into how agents operate and supporting oversight in environments where reliability and compliance are essential.

DXC OASIS grounds agents in customer-specific and system-specific information, operational expertise and vendor documentation. This enables agents to work with the context of individual enterprise environments rather than relying solely on general model knowledge.

The approach is designed to help organisations expand the use of AI agents in production without losing control over critical processes.

Combining LangGraph and LangSmith

The partnership brings together two components of LangChain's agent engineering ecosystem.

LangGraph provides the orchestration framework for coordinating agents, managing state and supporting workflows that include human intervention. LangSmith supports the development, evaluation and observability of those agents, helping teams assess performance and improve reliability over time.

DXC is integrating these capabilities into OASIS to support the development and continuous improvement of agents operating across complex IT estates.

Julia Schottenstein, chief operating officer at LangChain, said enterprises adopting agentic AI need to adapt the agent infrastructure to their business, ground agents in proprietary context, and establish governance and continuous improvement processes to make them reliable in production.

An open and flexible approach to enterprise AI

DXC OASIS is designed to support different AI models and interoperate with third-party and customer-owned systems. Organisations can select models for individual workloads according to requirements such as quality, cost, latency and privacy.

Deployments can span public cloud, private cloud, dedicated infrastructure and sovereign environments without requiring a redesign of the underlying architecture.

This model gives customers flexibility over their data, workflows, infrastructure and AI choices while allowing them to build on existing technology investments.

DXC and LangChain have also published a technical white paper, DXC OASIS: The Harness for Mission-Critical IT Operations, outlining their approach to deploying AI agents with governance, accountability and human oversight.

Moving agentic AI into critical enterprise systems

The partnership comes as organisations explore how AI agents can move beyond experimentation into operational environments where reliability, security and traceability are essential.

For enterprise IT teams, the challenge is not simply enabling agents to perform tasks. It is ensuring they can coordinate complex workflows, act within defined authority and provide measurable improvements without compromising operational control.

By combining DXC's experience operating mission-critical systems with LangChain's agent development and orchestration technology, the companies aim to advance a model of enterprise AI that brings automation and human expertise together.

About DXC Technology

DXC Technology is an enterprise technology and innovation partner delivering software, services and solutions to global enterprises and public sector organisations. Its capabilities span managed infrastructure services, application modernisation and industry-specific software solutions.

---

About LangChain

LangChain develops tools and frameworks for building, evaluating and operating AI agents. Its ecosystem includes LangGraph, an open-source framework for agent orchestration, and LangSmith, a platform supporting agent development, evaluation and observability.

---

Source & Credits

DXC Technology's official announcement  / Hotwire for DXC