DXC partners LangChain to govern enterprise AI agents

DXC Technology has partnered with LangChain to expand the use of AI agents in mission-critical IT operations, with a focus on governance and human oversight in enterprise deployments.

Under the agreement, DXC is using LangChain's LangGraph and LangSmith tools within DXC OASIS, its orchestration platform for IT operations. OASIS is in production and scaling towards 125 customer environments across regulated industries and multiple geographies.

The partnership comes as companies face growing questions about how to control AI agents in live operational settings. Gartner expects more than 40% of agentic AI projects to be cancelled by the end of 2027 because of inadequate risk controls, cost, or unclear value.

DXC said the system is designed to let AI agents act within defined authority while keeping a record of decisions, tool calls, and outcomes. Sensitive actions can also be paused for human approval before execution.

This approach is aimed at enterprise IT operations, where outages and faults can create operational and regulatory risks. OASIS is intended to let agents respond to operational events, investigate issues in parallel, and involve human staff when approval or judgment is needed.

DXC said the platform is already being used by 67 customers and is scaling towards 125 customer environments, signalling broader deployment across its client base.

Performance figures released by DXC covered more than 500 measured use cases. According to the company, OASIS reduced resolution time by 95% on average, while diagnostic accuracy reached 92% or better when checked against senior engineer judgment.

DXC also provided examples from production use cases. Filesystem incidents that previously took 53 to 90 minutes were reduced to three to six minutes. Database performance incidents that took two to three hours were cut to five to 12 minutes, while problem management tasks that once required two to six days were shortened to eight to 10 minutes.

Governance focus

A central element of the partnership is the controls around what agents can do and when they require oversight. DXC said actions are assessed using factors including reversibility, impact, verifiability, and an agent's track record.

DXC also said OASIS grounds agents in customer-specific and system-specific information, as well as DXC operational knowledge and vendor documentation. That is intended to reduce reliance on general model knowledge and keep agent decisions tied to the context of each customer environment.

LangChain, known for its open-source LangGraph framework and LangSmith engineering platform, is providing the software layer for agent orchestration, development, evaluation, and observability. For DXC, the partnership formalises technology it was already using inside OASIS.

"Agentic AI has the potential to fundamentally change how enterprises operate their most critical technology, but models alone are not enough," said Dan Gray, President, Global Infrastructure Services, DXC Technology. "Across more than 500 measured use cases, OASIS has reduced resolution time by 95% on average, with diagnostic accuracy of 92% or better when validated against senior engineer judgment. Through our work with LangChain, we're bringing together advanced agent technology, operational expertise, and governance to put AI agents to work safely at enterprise scale."

Open model

DXC said OASIS is model-agnostic and can work with third-party and customer-owned systems. That allows customers to choose different AI models for different workloads based on factors such as quality, cost, latency, and privacy.

Deployments can run across public cloud, private cloud, dedicated, and sovereign environments without redesigning the underlying architecture, DXC said. Customers retain control of their data, workflows, and operational processes built on top of the platform.

The deal also reflects a wider trend among enterprise technology providers to pair AI automation with auditability and approval controls as customers move beyond pilot projects. In regulated sectors in particular, suppliers are under pressure to show not just that agents can complete tasks, but that their actions can be traced and constrained.

Julia Schottenstein, Chief Operating Officer at LangChain, said customers are increasingly focused on how agent systems are structured and controlled in production settings. "Agents are transforming the application layer for every business, and the enterprises pulling ahead are taking ownership of their agent stack," said Julia Schottenstein, Chief Operating Officer, LangChain. "That means adapting the harness to their business, grounding agents in proprietary context, and building the governance and continuous improvement loops that make agents reliable in production. We're proud to bring that model into mission-critical enterprise operations through our partnership with DXC."