News for today:
- [Claroty] Native AI security agent for cyber-physical. Claroty Claire is a new AI security agent for mission critical cyber-physical environments. Claire is powered by the world’s most advanced CPS language model, trained on over a decade of industry expertise and the largest CPS data lake. This innovation marks a paradigm shift in organizations’ ability to ensure safety, uptime, and availability in an AI-driven world … Harnessing the power of the world’s most comprehensive CPS language model, comprising detailed information from over 6,500 unique OEMs and medical device manufacturers, deployed in 20K+ sites, across 50+ sectors and 60+ countries worldwide, and backed by award-winning threat research from Claroty’s Team82. Claroty
- [Check Point] Anti-Mythos play. Check Point launched Agentic Exposure Validation for Exposure Management, providing a way for organizations to assess vulnerabilities and exploits before threat actors do. Agentic Exposure Validation (AEV) uses AI agents that reason like attackers across the organization’s specific environment, correlating exposure data, asset context, live exploit research, threat intelligence, and protection coverage to determine whether an exposure is truly exploitable. Rather than relying on static severity scores, AEV follows a safe proving loop: it analyzes the relevant asset or CVE, enriches findings with live Check Point threat intelligence, checks whether existing controls already block the path, and builds a targeted validation that mirrors attacker reasoning without disruptive techniques. It then either proves the exposure with direct evidence, pivots to a new attack path when blocked, or discards the threat altogether. AEV is a critical validation capability within Continuous Threat Exposure Management (CTEM) programs, helping organizations move from discovery and prioritization into confident, evidence-based exposure reduction at AI scale. Check Point
- [Sophos] 89 seconds for SOC response. Sophos said that the use of agentic capabilities within its MDR over the past year has shrunk threat response to 89 seconds. The production data from the past twelve months sets a new benchmark for managed security operations: [1] 89 seconds from case creation to fully automated response. This metric measures how fast the Sophos Central Defense System acts on cases AI is authorized to resolve, translating directly into faster response and stronger resilience against attacks that move at machine speed; and [2] 52% of MDR cases closed end-to-end by AI, without human intervention required, inside boundaries continuously calibrated by analysts. This metric measures the volume of work AI is doing autonomously, not just alert triage or threat containment. Sophos
- [KnowBe4] Securing Microsoft Teams. KnowBe4 launched KnowBe4 Messaging Security for Microsoft Teams, for detection and remediation of content threats in Teams, such as phishing and social engineering. Organizations can now detect and remediate threats directly within Microsoft Teams. This also provides organizations with previously unavailable visibility into risks associated with external communications. KnowBe4
- [KnowBe4] New capabilities in KnowBe4 Prevent (previously Egress). KnowBe4 announced two new capabilities for its Prevent offering, for outbound email security: a DLP rule builder, and content analysis for misdirected emails. On the second: The AI-powered feature detects when data is about to be sent to the incorrect person. It provides a context-aware warning to the user before the email is sent, turning a near mistake into a teachable moment. Available June. KnowBe4
- [Forcepoint] Claude Compliance API integration. Forcepoint announced its integration with Claude Enterprise via the Claude Compliance API. Customers can secure data and AI before agents act on sensitive information, applying the same Forcepoint data discovery, classification intelligence and single-policy framework that already governs AI workflows, SaaS, websites, networks, email and endpoints. A single AI data security view covers Claude Enterprise, Microsoft 365 Copilot, ChatGPT Enterprise, shadow AI usage and AI agents. Users can find data as it is created, adapt when risk changes and protect everywhere data is used, across the most widely deployed AI environments. Forcepoint
- [ArmorCode] Anya Agents – new agentic AI framework. ArmorCode announced a new agentic AI framework – called Anya Agents – for AI-driven security at enterprise scale. Anya Agents [delivers] reusable, role-aware AI workers that are grounded in insights from ArmorCode’s Context Risk Graph, combining tool findings, CVE databases, asset inventories, software supply chain data, threat intel feeds, and business context. Rather than functioning as generic chatbots, each agent is scoped for a specific security workflow with preconfigured prompts, defined context, and bounded actions to produce consistent outcomes. Customers are now able to infinitely augment their security teams to tackle the tsunami of vulnerabilities coming from frontier AI models. ArmorCode