April 2026 Summaries

AI agents running in AWS environments pose significant security challenges due to a lack of visibility into their operations, models, and data reach, creating a potential "blast radius" during security incidents. This complexity arises from multiple teams, such as developers, platform teams, and data engineers, deploying agents on Amazon Bedrock AgentCore without a unified governance view. Veza addresses this gap by offering comprehensive visibility and governance for both Amazon Bedrock Agents and AgentCore, covering various frameworks like LangGraph and OpenAI Agents SDK. It models critical components like agent runtimes, gateways, memory, execution tools, and models to enhance access control. Veza helps security teams answer critical questions about agents' outbound and inbound access, enabling precise security queries and governance of AI lifecycle management. By mapping the complex web of IAM policies and permissions into a clear graph path, Veza ensures only authorized personnel can manage and invoke these agents, safeguarding against data exfiltration and unauthorized actions. Veza AI Agent Security with Amazon Bedrock AgentCore support is available for existing customers, integrating seamlessly into current AWS setups.

The 2026.4 release from Veza introduces AI-powered enhancements and expanded integrations to improve security workflows and governance across IT and HR platforms. Key updates include new integrations with Freshservice and Personio, optimized AWS IAM batch extraction, and enhanced Kubernetes RBAC policy visibility. The update also introduces AI-driven features, such as Access AI explanations for dashboards, AI-suggested queries, and NHI governance with suggested owners for unowned identities. The platform now supports more precise lifecycle management with real-time provisioning policy status and Google Workspace OU targeting. Enhancements in access reviews include consolidated risk scores and better due date visibility for multi-level reviews, while lifecycle management features offer actions like Get Node From Graph for workflow execution. Additionally, improvements in AI Agent Security and Access Intelligence usability aim to streamline processes and provide deeper insights.

AI agents in ServiceNow are now more transparently governed thanks to the integration of Veza's AI Agent Security, which extends its Access Graph capabilities to these agents. Previously managed like prompts, these agents are now treated as identities, inheriting roles and ACLs and operating as ServiceNow users with the ability to access tables and records. Veza's solution provides automatic discovery of AI agents, tools, skills, and configurations without additional deployment, enhancing visibility into who can invoke and manage these agents and what data they can access. This integration supports critical security questions, such as identifying the scope of agents' access and the roles and privileges they inherit. By capturing the operational model configurations and binding them to specific models, Veza facilitates oversight of AI capabilities and their compliance with organizational policies. This advancement offers security and IAM teams improved oversight through features like access reviews, blast radius assessments, and rules and alerts, ensuring that AI agents are managed alongside human and service accounts within ServiceNow environments.

Anthropic's Project Glasswing, featuring the Claude Mythos model, marks a significant advancement in cybersecurity by detecting long-standing software vulnerabilities, such as a 27-year-old flaw in OpenBSD and a 16-year-old bug in FFmpeg, which eluded traditional scanning tools. This rapid detection capability enhances cyber resilience but also highlights the need for comprehensive governance of AI agents, emphasizing identity management and access control to mitigate potential risks. As AI agents become integral to platforms like AWS Bedrock, Microsoft Azure AI Foundry, and Google Cloud Vertex AI, ensuring secure interactions between agents and resources is critical. Veza's Access Graph Architecture offers a solution by treating AI agents as first-class identities, allowing organizations to manage effective permissions and limit the blast radius of any compromised agent. The future of enterprise security hinges on two pillars: the hardening of software through initiatives like Project Glasswing and the implementation of least privilege agent identities through access control frameworks, ensuring a secure AI revolution.

Veza's Access AI is an innovative tool integrated into the Veza query experience, offering security analysts immediate insights into identity risks by explaining query results in plain language and providing actionable recommendations. This AI-driven feature contextualizes risk data, highlighting the specific risk profile, potential security implications, and trends over time, thereby transforming static findings into dynamic, actionable insights. With Access AI, security teams can engage in interactive conversations to better understand the significance of the data, prioritize actions, and make informed decisions without the need for external consultation or additional tools. Complementing Veza Actions, which allows for direct remediation like disabling accounts or launching access reviews, Access AI bridges the gap between identifying risks and taking decisive action, enhancing the overall efficiency and effectiveness of identity security management.

Veza's 2026.3 release introduces significant updates across various product areas, enhancing integration capabilities, security measures, and lifecycle management. The release includes new integrations with UiPath and Google Artifact Registry, improvements in access intelligence, and expanded governance capabilities for on-premises and AI agent infrastructure. Key features include the ability for security teams to remediate risks directly from query results, granular control over workflows for identity governance and administration (IGA) teams, and new public APIs for enrichment rules, AI agent management, and more. Enhanced features also offer better integration with Azure, support for OAuth2 authentication in ServiceNow, improved visibility for Workday Security Group inheritance, and added support for various database authentication methods. Additionally, access reviews now feature extended API capabilities for grace periods, a new Action Allow List for review modifications, and improved reporting consistency. The release also focuses on lifecycle management by introducing selective entitlement removal, LDAP as a source and target system, and accelerated identity table population, along with new access request enhancements that allow for seamless automation without intermediary ITSM tickets.