News for today:
- AI agent identity change controls and identity incident response. Cayosoft updated its Guardian 7.2 platform, with new capabilities around AI agents and incidents that have identity drivers / derivates. Cayosoft Guardian … brings AI agent identities into existing identity threat detection and response (ITDR) workflows, giving security teams visibility, reporting, alerting, and automated rollback, without adding another dashboard. Cayosoft also launched a new Identity Forensics & Incident Response (IFIR) service, a purpose-built incident response offering designed specifically for Microsoft hybrid identity environments to address the growing operational and security risks created by complex environments and the rapid rise of non‑human identities. Cayosoft
- Operationalizing AI agents across the enterprise. Hexaware launched Agentverse, with more than 600 ready-to-deploy AI agents, across use cases for customer experience, financial services, manufacturing, and more. The platform integrates with core enterprise systems including CRM platforms, IT service management tools, knowledge repositories, data platforms, telephony systems, and collaboration applications. Through advanced orchestration, AI agents retrieve contextual knowledge, interpret process documentation, automate conversations, and execute operational actions directly within enterprise workflows. Built-in governance capabilities including role-based access controls, audit trails, observability, and policy guardrails ensure agents operate securely and in alignment with enterprise compliance requirements. Hexaware
- On the business benefits of better identity security. Research commissioned by Ping asserts that organizations with continuous, contextual identity verification have a whole range of better outcomes. Based on a global survey of 794 organizations, IDC found that enterprises meeting verified trust criteria delivered statistically significant performance advantages across core enterprise metrics: Customer registration conversion: 51% higher; Compliance readiness: 44% stronger; Fraud losses: 43% lower; and Workforce onboarding time: 47% faster. Ping wants organizations to make a shift from identity [as] a perimeter authentication event to a runtime control plane governing every authorization decision. Ping Identity
- Agentic network security. Tufin announced a new offering – the Tufin Agentic Network Security platform – with ready-to-run AI agents for network security included. Tufin is launching a collection of AI agents designed to take on meaningful network security work for teams already stretched too thin — freeing scarce security expertise to focus on higher-order risks, critical decisions, and defending the enterprise against an evolving AI threat landscape. The agents: Compliance, Posture, Application Deployment, and Policy Recertification. Tufin
- Validating cyber capability. Immersive updated its Immersive One platform, to enable access to validated real-world cyber capability outcomes for leaders. Immersive One advances organizational resilience across three critical domains: securing AI adoption, validating real‑world enterprise defence capability, and delivering framework‑aligned evidence to regulators, boards, and executive leadership. Immersive
- Securing AI and the data it uses. Varonis released its Atlas AI security platform, for visibility / discovery, AI runtime guardrails, and detection of suspicious activities, among others. Atlas covers the entire AI security lifecycle — from discovery and posture management to runtime protection and compliance — in a single solution. It connects to virtually any AI system that organizations build or run: hosted AI platforms, custom LLMs, agentic frameworks, chatbots, and embedded AI. Through integrations with the Varonis Data Security Platform, Atlas brings data context that standalone AI security tools can’t match. Available immediately. Varonis
- Multimodal AI foundation model for preventing AI-driven attacks. Abnormal AI launched Attune 1.0, its new single, unified model for cybersecurity, which brings together its previous multiple models. Unlike earlier detection systems that treated identity, behavior, and content as separate signals, Attune 1.0 utilizes a unified multimodal architecture. By learning these modalities jointly, the model better understands how signals reinforce or contradict one another to reveal patterns that attackers work hardest to hide. From inception, Abnormal has focused on knowing what’s normal in customers’ environments, enabling it to detect and block novel, AI-driven attacks when a deviation in behavior is found. Abnormal AI