DXC & Primary launch AI Zero Trust security service

DXC has formed a strategic partnership with Primary and will serve as the exclusive managed services provider for Primary's AI-native Zero Trust Platform.

The partnership centres on a joint offering for organisations deploying artificial intelligence systems in production, including enterprises and government agencies. The service is designed to govern how AI agents and enterprise AI applications access data, while applying Zero Trust security principles to those environments.

The deal reflects a broader shift in corporate technology use as companies move beyond AI trials and begin putting autonomous systems into day-to-day operations. That transition has created new security and governance challenges, particularly where AI systems interact with sensitive internal data, business processes and identity systems.

Traditional security models have largely been built around users, devices and networks. The companies argue those frameworks do not fully address the risks posed by AI agents that act across multiple systems and make decisions or make requests without direct human input.

Under the agreement, DXC's global cybersecurity organisation will deliver the managed service layer around Primary's platform, including consulting, implementation, integration, governance and ongoing security operations.

The service is intended to give customers greater oversight of how AI systems interact with enterprise data and applications. It is also designed to support identity controls, policy enforcement and compliance in regulated sectors, where data handling and sovereignty rules can be strict.

Security focus

The offering is aimed particularly at organisations in highly regulated environments. Sectors with stricter obligations around data access, auditability and control have been among the most cautious adopters of AI tools, as concerns remain about how automated systems obtain information and what permissions they hold once deployed.

Zero Trust is a security model based on continuous verification rather than assumed trust inside a network boundary. Applied to AI, it means checking whether an AI agent should be allowed to access a specific dataset, identity or workflow at the moment a request is made, rather than relying on broad standing permissions.

That approach has gained attention as businesses explore agent-based AI systems that can retrieve files, query databases, trigger actions and interact with internal software. For security teams, the central issue is not only whether those actions can be observed, but whether they can be governed in real time according to business rules and regulatory requirements.

DXC said the service would help customers secure AI agents and enterprise AI applications, govern AI access to enterprise data, strengthen identity and policy enforcement, reduce risk across enterprise environments, support compliance requirements and increase visibility across users, applications, data and AI agents.

Dawn-Marie Vaughan, Cybersecurity Global Offering Lead at DXC, said security and governance must keep pace as organisations expand AI in operations.

"As enterprises move from AI pilots to production, security and governance must evolve just as quickly," said Dawn-Marie Vaughan, Cybersecurity Global Offering Lead, DXC Technology. "Trusted AI requires more than securing infrastructure; it requires governing how AI systems access data, identities, and business processes. Together with Primary, we are helping customers bring Zero Trust to enterprise AI, with the visibility, control, and operational confidence needed to move from pilot to production."

Control layer

Primary's role in the partnership centres on its Unified Control Plane, which it describes as AI-native and built to enforce policy across the AI workflow. The aim is to create a single control layer that can interpret signals from different parts of an organisation's technology estate and use them to determine whether an AI-driven action should proceed.

For large organisations, that can include telemetry from identity systems, applications, infrastructure and data sources. Bringing those signals together has become a major technical and governance challenge as AI tools spread beyond standalone assistants into systems that can take action inside core enterprise environments.

Michael Marx, Co-Founder + President at Primary, said the key challenge is deciding when an AI action should be permitted and enforcing that decision quickly enough to be useful.

"The defining security challenge of enterprise AI is not simply seeing what an AI agent is doing-it is having enough context to determine whether that action should be allowed, as well as the corresponding infrastructure to automate enforcement of that real-time decision," said Michael Marx, Primary's Co-Founder + President. "Primary transforms cross-domain telemetry into continuous, context-aware policy enforcement across the full AI workflow. With DXC's global implementation and managed security expertise, customers can establish a unified control layer that protects sensitive data, reduces operational risk, and enables AI systems to act securely within clearly defined enterprise boundaries."

The partnership puts DXC in charge of managed services delivery for the platform, giving it a defined role in a market where enterprises are seeking outside support to integrate AI governance into existing security operations. It also highlights how service providers and software companies are pairing up to address one of the more difficult aspects of enterprise AI adoption: controlling autonomous access to data and systems.