DXC Technology partnered with LangChain on October 13, 2026 to scale governed agentic AI for mission-critical IT operations. Its OASIS orchestration platform is already in production and scaling toward 125 customer environments across regulated industries and multiple geographies. The combination matters for business because it links faster incident resolution with controls required for critical systems.

DXC and LangChain to Scale Governed AI Agents for Critical IT

OASIS platform, LangGraph use and LangSmith license

DXC was already using LangGraph inside OASIS, described as its intelligent orchestration platform and operating model for mission-critical IT operations. Under the partnership, DXC is also licensing LangSmith to support development, evaluation and observability of OASIS agents. LangGraph handles delegation, state management and human-in-the-loop workflows, while LangSmith provides agent development, evaluation and observability tooling. LangChain is known for the open-source LangGraph framework and the LangSmith engineering platform.

OASIS works as an open agentic orchestration layer across existing tools rather than replacing them, with ServiceNow and Dynatrace named as connected systems. Agents respond to operational events, investigate issues in parallel and pull in specialized capabilities across technology domains. People enter the process when judgment or approval is required, leaving experts to focus on higher-value decisions. DXC first introduced OASIS on April 28, 2026 as a managed-services operating model moving delivery from reactive support to real-time operations.

The production figures cited by DXC show large time reductions in specific IT cases. Filesystem incidents that took 53-90 minutes are resolved in 3-6 minutes, database performance incidents moved from 2-3 hours to 5-12 minutes. Complex problem management that required 2-6 days now takes 8-10 minutes. Across more than 500 measured use cases, OASIS reduced resolution time by 95% on average, with diagnostic accuracy of 92% or better when validated against senior engineer judgment, according to Dan Gray, President of Global Infrastructure Services at DXC.

What governed expansion means for enterprise IT

For companies running complex infrastructure, the setup offers capacity and speed without adding separate tooling stacks. OASIS is model agnostic and interoperable with third-party and customer-owned systems, so different models can be selected by quality, cost, latency and privacy. Deployments can span public cloud, private cloud, dedicated and sovereign environments without redesign, while customers keep control of data and workflows. Small firms gain access to parallel investigation, large enterprises gain a way to standardize operations across geographies.

Governance remains the condition for wider use, and several limits need checking before adoption. OASIS assesses actions by reversibility, impact, verifiability and agent track record, pauses sensitive actions for human approval and traces decisions, tool calls and outcomes. Documentation describes traceable actions, role-aware access controls and progressive autonomy that can be expanded or constrained. Buyers should verify which actions require approval, how enterprise context is maintained and what the white paper "DXC OASIS: The Harness for Mission-Critical IT Operations" defines for accountability.

The marker to watch is progress toward 125 customer environments in regulated industries and geographies. If DXC reports that count alongside sustained resolution and accuracy levels, governed agents will have moved from pilots to repeatable operations. LangChain COO Julia Schottenstein points to the same test: ownership of the agent stack, proprietary context and continuous improvement loops that keep agents reliable in production.