AppViewX and Teleport Launch Products to Secure AI Agent Identities as Enterprise Deployments Surge

On June 16, AppViewX and Teleport both released products to manage the identity and access of AI agents in the enterprise. AppViewX announced Agent Identity Security, a new module for its machine identity platform. Teleport launched two new capabilities, LLM Proxy and Delegated Identity, as part of the public beta of its Beams framework. Both aim to close a gap that security teams say leaves agents with unchecked privileges and limited auditability.

Gartner projected in 2025 that by 2028, at least 25 percent of enterprise breaches will involve AI agent abuse, up from less than 1 percent in 2024. The two product launches come as organizations race to govern agents that act autonomously on sensitive data and infrastructure.

AppViewX Targets Governance and Post-Quantum Readiness

AppViewX said its Agent Identity Security, built on a decade of PKI and certificate lifecycle management, discovers every AI agent in an environment, tracks its associated LLMs, MCP tools, and credentials, and enforces policies based on frameworks such as the NIST AI RMF, EU AI Act, and SOC 2. The product includes a feature called Guardian Agent, an AI-powered security companion that detects anomalous agent behavior and guides remediation. CEO Archit Lohokare described agents as “the largest workforce most enterprises never hired” and argued the solution lets organizations deploy them at scale without sacrificing security.

• Provides a centralized AI Bill of Materials (AIBOM) for all agents.
• Enforces task-based least-privilege access at the agent identity level.
• Integrates with existing PAM and IAM tools for consistent enforcement.
• Generates audit-ready compliance evidence for boards and regulators.
• AppViewX claims the platform is also designed to support post-quantum cryptographic agility.

Teleport Focuses on the Agent-to-LLM and Agent-to-Resource Gaps

Teleport’s Beams beta attacks a different part of the problem. The LLM Proxy inspects and controls what instructions an agent sends to a large language model, while Delegated Identity binds the agent’s actions to a human or service principal so every operation can be traced and audited. Teleport already manages access for infrastructure such as servers, Kubernetes clusters, and databases. The new capabilities extend that model to the two most consequential points in an agentic workflow: the prompt and the resource request.

“Much of AI innovation to date has focused on the model, not the pipeline around it,” said Ev Kontsevoy, CEO of Teleport, in a statement. “Beams gives security teams the ability to observe and control what an agent is told to do and what it touches.” The product is in public beta as of June 16.

Both companies are responding to a reality underscored by Equinix in a separate blog post on the same day: ClickOps is dead, and agentic AI killed it. The post argued that old manual approaches to connectivity and provisioning cannot handle the scale of distributed AI workloads. Identity management for agents, the piece suggested, must be fully automated and policy-driven.

Neither AppViewX nor Teleport disclosed pricing or general availability dates. Teleport expects to exit beta in late 2026. AppViewX said Agent Identity Security is available immediately as a module within its existing platform.