When the world’s largest banks, airlines, and insurers consider adopting artificial intelligence, they face a dilemma. These organizations operate under strict security and compliance requirements, and a single AI hallucination could disrupt critical operations or trigger regulatory action. Yet the pressure to modernize grows daily. The announcement of a multi-year global alliance between DXC Technology and Anthropic represents a significant attempt to bridge this gap.
DXC Technology, an IT services giant with approximately 115,000 employees across 70 countries, has operated the legacy systems of major financial institutions, airlines, and government agencies for decades. The company has now committed to training tens of thousands of Claude-certified forward-deployed engineers (FDEs), who will embed directly into customer organizations. These engineers will bring Anthropic’s Claude AI into the same transaction-processing, claims-handling, and operational systems that DXC has managed for years.
The partnership begins with DXC already having tested Claude inside its own operations. In April 2026, DXC launched OASIS, its AI-native orchestration platform for managed services. According to the company, more than 95% of OASIS’s code was generated by Claude, with human engineers reviewing the output. DXC estimates this accelerated software development by a factor of ten. The platform now serves over 50 customers, and the firm plans a global rollout.
When an AI model writes 95% of the code for a mission-critical platform, the nature of software engineering shifts from writing to reviewing. This transformation requires a new kind of workforce. DXC will recruit engineers from its existing development teams and certify them through Anthropic Academy, supplemented by DXC’s own curriculum designed specifically for the mission-critical environments its customers operate.
The alliance targets four initial application areas. In insurance, DXC will deploy Claude for agentic solutions that modernize core systems while respecting each customer’s business context. Under the Modernization as a Service (MaaS) initiative, Claude will help analyze, refactor, and modernize legacy codebases faster and with greater accuracy than traditional approaches. For cybersecurity, DXC is developing an always-on security engineer subagent built on Claude Security, deployed across security operations centers. Finally, in application services, OASIS agents will embed Claude directly into the maintenance and management environments DXC operates for enterprise customers.
The real test of enterprise AI is not whether it can generate a marketing email, but whether it can handle a claim adjustment in a regulated market without introducing error or bias. This distinction explains why the DXC-Anthropic partnership focuses on industries where mistakes carry high costs. A hallucination in a bank’s transaction system could have immediate financial and regulatory consequences. An airline reservation system failure could strand thousands of passengers. The certification process for DXC’s engineers aims to prevent such scenarios.
Critics might point out that major cloud providers including Microsoft, Google, and Amazon have their own AI offerings for regulated industries. Microsoft’s Azure OpenAI Service, for example, already powers banking and healthcare applications. However, the DXC-Anthropic partnership differs in two respects. First, DXC operates the underlying legacy systems, meaning Claude can be integrated directly into existing workflows rather than requiring a separate integration layer. Second, DXC’s engineers will be certified specifically to work with Claude in these environments, creating a focused expertise that general-purpose integrations may lack.
In regulated industries, trust is the highest form of technical debt that cannot be paid down overnight. DXC’s 50-year history of running critical systems provides a foundation that no start-up can replicate. The company’s CEO Raul Fernandez described the alliance as combining "trust and experience with the most advanced AI technology available."
For enterprise leaders in banking, insurance, aviation, and government, the partnership signals that AI adoption in regulated environments is no longer a theoretical question. The question has shifted to implementation: how to deploy AI without violating compliance requirements, how to train engineers who understand both the technology and the regulatory landscape, and how to measure success when failure carries existential risk.
The coming years will reveal whether this model works at scale. If successful, it could establish a blueprint for AI adoption in every industry where trust, security, and accuracy are non-negotiable. If it fails, the reasons will likely involve not the technology itself, but the challenges of integrating AI into systems designed long before machine learning existed.
For now, the alliance represents a bet that the most valuable application of advanced AI is not replacing human judgment in critical systems, but augmenting it with engineers who understand both the power and the limitations of the technology.