EY has introduced a new Agentic Cyber Defence Centre in collaboration with Microsoft, combining AI agents, security technologies, and governance controls to help organisations modernise cybersecurity operations.
The initiative builds on the EY–Microsoft partnership, bringing together Microsoft’s security platform and threat intelligence capabilities with EY’s cybersecurity expertise, AI frameworks, and implementation services. The centre is designed to help businesses coordinate security operations across fragmented environments while maintaining oversight of AI-driven decisions.
As organisations increasingly deploy AI agents to investigate threats and respond to incidents, managing their authority and accountability has become a central challenge. EY’s platform aims to address this by introducing policy-based controls, human oversight, and tamper-evident records documenting agent activity.
Managing AI agents in cybersecurity operations
The centre focuses on enabling AI agents to perform more cybersecurity tasks while ensuring organisations retain control over automated actions. EY says its approach incorporates controls that define what agents can do, how actions are authorised, and how security teams can review their decisions.
Tamper-evident records are intended to provide visibility into agent activity and help organisations investigate decisions, track changes, and coordinate incident responses. These records can support accountability by preserving evidence of what agents did, although they do not guarantee that automated decisions were correct.
Richard Watson, EY Global Consulting Cybersecurity Leader, said organisations need to rethink how cybersecurity operations are structured rather than simply making existing security operations centres faster.
The platform combines EY’s cybersecurity intellectual property, EY.ai Value Blueprints, and agentic accelerators with Microsoft’s security infrastructure, threat intelligence, and AI capabilities. The companies say this will provide organisations with a framework for developing more adaptive cyber defence operations.
Connecting security across enterprise environments
EY plans to extend its agentic cybersecurity framework across multiple enterprise domains, including identity management, data, applications, cloud infrastructure, and operational technology.
The initiative is intended to address security environments where tools, processes, and systems often operate independently, making it difficult for teams to coordinate responses and maintain consistent security controls.
By introducing reusable agentic blueprints, EY aims to establish a repeatable approach to deploying AI-driven security capabilities across different parts of an organisation. Rather than relying on disconnected implementations, businesses could use common frameworks to coordinate security operations and continuously improve their defences.
The company also intends to integrate governance and security controls directly into the operating model. Continuous controls and AI agents could reduce manual workloads and operational overhead, while human judgement remains central to decisions that require approval or intervention.
EY deploys the technology internally
EY is also acting as what it calls “Client Zero,” deploying Microsoft security technologies for agentic cyber defence across its own global organisation.
The internal implementation will be used to test operating models, validate governance controls, and develop reusable deployment patterns. EY says the experience will help inform future client implementations and reduce the risks associated with introducing AI agents into cybersecurity operations.
The approach allows EY to evaluate how agentic security systems function within a large enterprise environment, including how agents interact with existing security tools, how actions are governed, and where human intervention remains necessary.
Naseem Tuffaha, Corporate Vice President of Security at Microsoft, said securing AI-powered enterprises requires a different approach to cyber defence as AI changes both business operations and the threat landscape.
The partnership aims to combine Microsoft’s security and AI capabilities with EY’s cybersecurity transformation expertise to help organisations move beyond traditional security operations and develop more adaptive, resilient defence strategies.
The challenge of autonomous cyber defence
The launch reflects a broader shift toward agentic AI in enterprise cybersecurity, where systems are increasingly expected to move beyond analysing information and recommending actions to actively participating in investigations and incident response.
However, expanding agent authority introduces new governance challenges. Organisations must determine which actions can be performed autonomously, which require approval, and how mistakes or unintended actions can be investigated.
EY’s approach places policy controls, traceability, and human accountability at the centre of its agentic cyber defence framework. As organisations begin integrating AI agents into security operations, the ability to manage their authority while maintaining visibility into their actions will be critical to adoption.
The Agentic Cyber Defence Centre represents EY and Microsoft’s effort to make AI-driven security operations more scalable, coordinated, and governed across enterprise environments.


