Securonix launches behavioural analytics for AI threats

Securonix has launched Advanced Behavioural Analytics, a product designed to detect threats involving both humans and AI agents.

The offering combines Agent and Entity Behaviour Analytics with User and Entity Behaviour Analytics to identify activity that is permitted within systems but appears abnormal in context. It is intended to help analysts spot behavioural changes across users, identities, assets, applications and AI agents, while keeping significant response actions subject to human approval.

The launch comes as companies give AI systems wider access to internal data, software tools and business processes. That shift has created a new security challenge for operations teams, which must determine whether unusual automated activity is legitimate, poorly configured or the result of manipulation.

According to research published by Securonix, 99.6% of 1,000 cybersecurity professionals surveyed said their organisations had increased budgets for AI cybersecurity automation over the past year. The same survey found that 96.6% said cybersecurity automation is important to their organisation, while 49% said they already use AI in behavioural analytics with human approval.

The system is built to flag cases in which an AI agent uses an unexpected tool, accesses data outside its assigned purpose or changes its operating pattern. It connects those changes to identity, authority, data use, asset sensitivity, timing and threat context, allowing analysts to review patterns across systems and over time.

Securonix argues that compromised accounts, insider activity and AI agents can all resemble normal behaviour when viewed as single events. By combining analytics across human and machine activity, the system is intended to give security teams a broader view of developing risk rather than isolated alerts.

Traditional security information and event management platforms were largely designed before AI agents became part of the attack surface. Securonix says its product extends behavioural baselines and risk detection across cloud environments, applications, identities and AI agents, with policies aimed at identifying established attack techniques including reconnaissance, exploit development, privilege escalation and data movement.

The wider platform also includes Sam, the company's AI SOC analyst, which can handle repeatable Tier 1 and Tier 2 tasks such as triage, investigation, evidence collection and case creation. An Agentic Mesh manages AI-supported workflows, while Agentic Guardrails keep actions policy-bound, visible and auditable.

For insider-risk investigations, the Insider Intent Agent adds contextual and corroborating evidence while preserving competing explanations. Analysts are expected to review each signal as part of a broader evidence set rather than treat a single message, search or behavioural deviation as proof of intent.

"AI is compressing the time between initial access and business impact, while security leaders remain accountable for the decisions made under that pressure," said Toby Weiss, Chief Executive Officer of Securonix.

"Securonix Advanced Behavioural Analytics helps analysts identify meaningful behavioural change earlier and gives them clear control over how AI supports the response. Consequential response actions remain subject to human approval," added Weiss.

External users of the company's threat analytics tools said behavioural context remains a persistent challenge, especially in cloud environments where workloads and identities change frequently.

"In a managed SOC, the hardest part isn't writing detection rules. It's understanding each customer's environment well enough to separate real threats from normal activity. That becomes even harder in cloud and Infrastructure-as-Code environments, where workloads appear and disappear and identities constantly change," said Darren Humphries, Group Chief Information Security Officer of Acora.

"Securonix Threat Analytics builds that behavioural baseline automatically, then uses risk scoring and evidence boards to show our analysts what matters and why. For Acora's customers, that means fewer false positives, faster investigations and quicker containment, without losing the customer context that makes detection effective," added Humphries.

Alongside the product launch, Securonix said it has obtained ISO/IEC 42001:2023 certification for its AI management system. The certification covers governance of AI across the company's product portfolio and, according to Securonix, followed an external audit with no major findings.

The standard is intended to provide independent validation of AI governance, risk management and human oversight. For customers, that can support procurement reviews and broader due diligence as security suppliers face greater scrutiny over how automated systems are managed and controlled.

Advanced Behavioural Analytics is available now.