DXC Technology partners with LangChain to deploy AI agents in mission‑critical IT operations
October 24, 2026 · 2 min read
Also published in 日本語, Nederlands, हिन्दी, 中文, العربية, Indonesia, 한국어, Deutsch, Français, ไทย, Italiano, Türkçe, Русский
DXC Technology has partnered with LangChain to integrate AI agents into mission‑critical IT operations. The collaboration leverages LangGraph and LangSmith to enhance the OASIS platform, which orchestrates critical systems. DXC will acquire a LangSmith license to support development, evaluation and observability of OASIS agents. The expanded solution will operate across 125 customer environments in regulated sectors worldwide. In production use cases, file‑system incidents that previously required 53 to 90 minutes can now be resolved in 3 to 6 minutes, and database performance issues that once took 2 to 3 hours can now be fixed in 5 to 12 minutes. Complex problem handling time has been cut from 2 to 6 days to 8 to 10 minutes.
Dan Gray, president of Global Infrastructure Services at DXC, said AI‑based agents can fundamentally transform how enterprises manage critical technology, but models alone are insufficient. OASIS has reduced resolution time by 95% across more than 500 use cases, achieving diagnostic accuracy of 92% or higher when compared with expert engineer criteria. The partnership enables agents to operate within defined authority limits, with controls for reversibility, impact, verifiability and history. Sensitive actions can be paused until human approval is obtained, and all agent decisions, tool calls and results are logged for visibility. OASIS provides agents with contextual information about each business environment, allowing them to reason based on system‑specific and customer‑specific data, DXC operational expertise and vendor documentation.
This context, combined with governance policies and human oversight, allows safe scaling of agents in production without losing responsibility or control. Organizations can select different models for each workload according to quality, cost, latency and privacy requirements. The open approach offered by the collaboration gives enterprises flexibility and control as AI technologies evolve. OASIS is model‑agnostic, supports interoperability with third‑party systems and can continuously improve by learning from agent execution data.