The Agentic SOC Alliance has launched seven CISO-led working groups to develop guidance for autonomous security operations. Dr Edward G Amoroso will coordinate the coalition's work.
The groups aim to create a framework for organisations moving agentic security operations centres from pilot projects into production.
The initiative brings together enterprise security leaders, researchers, academics, systems integrators and technology providers to define how an AI-driven security operations centre should function. Members cited in the announcement include CrowdStrike, LangChain and ExtraHop.
Amoroso, Chief Executive Officer of TAG Infosphere and an NYU research professor, will oversee the effort. He previously spent 31 years at AT&T, where he served as Senior Vice President and Chief Information Security Officer.
Operating model
The coalition is focused on what it describes as a new security operations centre, or SOC, operating architecture built around context, a control layer called the harness, and interchangeable AI models. The aim is to give autonomous agents access to real-time evidence while subjecting their actions to governance, guardrails and human approval.
The alliance argues that companies face unresolved questions as they introduce AI agents into security operations, including how much authority to grant them, how to verify their actions, and how to ensure they can work safely across multiple tools and data sources.
"The shift to autonomous security operations is happening now, and enterprises need an architecture they can trust as agents begin taking actions with significant consequences," said Dr Edward G Amoroso, Chief Executive Officer of TAG Infosphere and coordinator of the Agentic SOC Alliance. "The winning architecture will be open, governed and designed for a world where the agents, models and security technologies continue to evolve. The Alliance can make that architecture practical by building it, validating the outcomes alongside industry veteran security leaders and creating a path to trusted production-scale autonomous security operations."
ExtraHop founded the alliance earlier this year to define and standardise a common architecture for AI-driven security operations. Interest quickly grew among security leaders looking to test the model and involve their own suppliers and teams.
"AI is finally letting defenders solve problems we've never been able to solve in the history of cyber defense, but trusted autonomy is the foundation we as an industry need to build together first," said Greg Clark, Chief Executive Officer of ExtraHop. "When we founded the Alliance, we offered an open reference architecture as a starting point. What happened next was remarkable. Within days, security leaders wanted to know more about the proposed architecture and bring their vendors and teams together to create operational prototypes. Now with Ed stepping in to coordinate the Alliance and CISOs helping shape its direction, we can unite as an industry on an open, reliable and interoperable architecture that everyone can build on for years to come."
Working groups
The working groups cover several layers of SOC design and operation. They include a SOC Reference Architecture group, a Human Layer group, an Agentic AI Layer group, an Infrastructure group, a Data and Context group, a Validation group, and an Industry Engagement group.
The SOC Reference Architecture group is tasked with coordinating the overall framework and publishing documents and tools. The Human Layer group will focus on tasks that remain with analysts, including detection engineering, case ownership, threat hunting and SOC management.
Other groups will address how AI agents support or carry out detection, response and analysis; how orchestration and workflow management should be handled; how telemetry and intelligence feeds are managed; and how organisations test systems through activities such as red teaming, penetration testing and vulnerability management.
Several alliance members said a shared framework is needed because no single supplier can answer the broader questions raised by AI-led security operations.
"Agents need access to models, context and telemetry, tooling and orchestration, security controls, governance and trust. Without a common approach to ensure those interconnected security layers can interoperate, reason together, and operate safely, autonomy can become a liability," said Jim Routh, Chief Trust Officer at Saviynt.
GuidePoint Security said the work is also intended to help companies understand how existing security systems fit into an AI-based operating model.
"Security leaders are being asked to bring AI into the SOC alongside the security functions they already run. The Alliance's working groups are addressing how SIEM, EDR, NDR and IAM integrate, how to obtain AI models and applications, and how to measure results against industry peers. Their output will come together in one Agentic SOC framework, so leaders can plan upgrades and investments against a shared standard," said Mark Thornberry, Chief Business Officer of GuidePoint Security.
World Wide Technology said the coalition's output could help vendors and integrators align deployments across different customer environments.
"Agentic SOC will reach production across enterprises when vendors and integrators build from one framework. We are working with the Alliance's working groups to design an overall reference architecture that holds across multiple environments and validate it against our experience with how enterprises actually deploy," said Chris Konrad, Vice President of Global Cyber at World Wide Technology.