Mirantis Unveils Lens Agents: A Unified Platform for Governed Enterprise AI Operations

San Francisco, CA – May 12, 2026 – Mirantis, a leading specialist in open-source and AI infrastructure, today announced the launch of "Lens Agents," a groundbreaking software platform designed to provide controlled, secure, and scalable operation of AI agents across diverse enterprise environments. This strategic expansion of the widely adopted Lens ecosystem addresses a critical need for organizations grappling with the proliferation of AI tools operating outside centralized governance.

Lens Agents offers a singular, policy-driven framework for enterprises to deploy, secure, and scale AI agents whether they operate on individual desktops or within complex cloud infrastructures. The system is immediately available through an Early Access Program, inviting companies to integrate a wide array of AI agents – from popular desktop applications like Claude, Cursor, and Copilot to external autonomous agents built on various frameworks, and even platform-native agents created directly within Lens Agents – all under a robust umbrella of governance, traceability, and policy enforcement.

The move marks a significant evolution for Mirantis’ Lens product, which has long served as a leading Integrated Development Environment (IDE) for Kubernetes, boasting over a million users across development and operations teams. By simplifying cluster management and accelerating cloud-native development workflows, Lens has become an indispensable tool. Lens Agents extends this foundation, transforming Lens into a more comprehensive platform specifically tailored for enterprise AI operations.

The Uncontrolled Proliferation of AI: A Growing Enterprise Challenge

The rapid adoption of Artificial Intelligence across various business functions – from engineering and business operations to customer service – has introduced unprecedented efficiency and innovation. However, this growth has often occurred without adequate centralized control, creating significant challenges for enterprise IT and security teams. Miska Kaipiainen, Head of Product at Lens, succinctly captured this dilemma: "AI agents are now active in almost every company – the crucial question is whether they are also managed and controlled. Cloud platforms only govern what runs within their own cloud. But the vast majority of AI usage today occurs on laptops, with access to multiple clouds. None of this is adequately controlled. Lens Agents offers a single platform to manage everything: every agent, every model, every environment, strictly according to the customer’s rules."

This statement highlights a pervasive issue: the rise of "shadow AI." Employees, eager to leverage the power of AI, often integrate tools into their daily workflows without the knowledge or oversight of their IT departments. This creates myriad risks, including potential data leakage, compliance breaches, uncontrolled spending, and security vulnerabilities that can be exploited by malicious actors. The traditional perimeter defenses and cloud-specific governance models are proving insufficient in a world where AI agents can reside on individual endpoints, access sensitive corporate data, and interact with external services across multiple cloud providers. Lens Agents aims to close this critical gap by providing a unified control plane that extends beyond traditional boundaries.

A Chronology of Innovation: Lens’ Evolution to AI Governance

Mirantis’ journey to Lens Agents has been a deliberate and strategic one, building upon its established expertise in open-source and cloud-native technologies.

The Foundation: Lens as a Kubernetes IDE (2018 onwards)
Lens initially gained prominence as an intuitive and powerful IDE for Kubernetes. Its user-friendly interface simplified the complex world of container orchestration, making it accessible to a broader range of developers and operators. By abstracting away much of the underlying complexity, Lens empowered teams to manage Kubernetes clusters, monitor workloads, and debug applications with greater efficiency. This success established Lens as a critical tool for organizations embracing cloud-native architectures. The platform’s open-source nature fostered a vibrant community, contributing to its continuous improvement and widespread adoption.

Bridging to AI: The Lens MCP Servers (Recent Past)
Recognizing the growing convergence of cloud-native infrastructure and AI workloads, Mirantis took a crucial step with the introduction of Lens MCP (Multi-Cloud Platform) Servers. This development enabled AI assistants and tools to connect seamlessly with Kubernetes environments, allowing organizations to orchestrate and manage AI-related workloads alongside their existing containerized applications. The MCP Servers served as an initial bridge, facilitating the integration of AI capabilities within a controlled Kubernetes ecosystem. It laid the groundwork for managing AI computational resources and deployments, but the challenge of governing the agents themselves – their actions, access, and autonomy – remained largely unaddressed, especially for agents operating beyond the direct purview of a Kubernetes cluster.

The Strategic Leap: Lens Agents (Present)
Lens Agents represents the logical and necessary next step in this evolution. It expands the control plane from managing AI workloads within Kubernetes to governing the behavior and interactions of AI agents across all enterprise environments. This shift signifies Mirantis’ commitment to providing a holistic solution for the modern enterprise, where AI agents are not just another application but intelligent entities requiring sophisticated management and oversight. By moving from an IDE for infrastructure to a comprehensive platform for AI operations, Lens Agents positions Mirantis at the forefront of responsible AI adoption.

Comprehensive Control: Key Features of Lens Agents

Lens Agents is engineered with a robust suite of features designed to ensure complete governance and security for AI agents, irrespective of their deployment location or underlying framework. These capabilities are crucial for maintaining operational integrity, mitigating risks, and complying with stringent regulatory requirements.

1. Agent Identity and Access Control:
At the core of Lens Agents is the ability to assign a unique, verifiable identity to every AI agent. This identity is then linked to granular access control policies defined by enterprise rules. This means that each agent, whether it’s a desktop assistant or an autonomous bot, operates under a distinct persona with predefined permissions. Organizations can specify which systems, data sources, and external services an agent can interact with, preventing unauthorized access or lateral movement within the network. This feature is paramount for preventing "rogue agents" and ensuring that AI operations align with organizational security postures.

2. Isolated Execution Environments (Sandboxing):
To mitigate the risk of unintended actions or security breaches, Lens Agents provides isolated sandbox environments for agent execution. These sandboxes act as protective enclosures, preventing agents from directly interacting with the host system or other network resources beyond their designated scope. If an agent encounters malicious code or attempts an unauthorized action, the impact is contained within its sandbox, safeguarding the broader enterprise infrastructure. This isolation is critical for experimenting with new AI models or running agents that might access sensitive data without fear of system-wide compromise.

3. Server-Side Credential Injection:
A significant security vulnerability often arises from agents directly holding or accessing sensitive credentials (e.g., API keys, database logins). Lens Agents addresses this by implementing server-side credential injection. Instead of agents storing or directly possessing access keys, these credentials are securely managed and injected into the agent’s execution environment only when needed, and only by authorized server components. This "just-in-time" and "least privilege" approach drastically reduces the attack surface, ensuring that even if an agent’s execution environment is compromised, the underlying credentials remain secure and inaccessible to attackers.

4. Comprehensive Audit Trail:
Transparency and accountability are non-negotiable in enterprise AI. Lens Agents provides a complete and immutable audit trail, meticulously logging every interaction an agent has with enterprise systems, data, and external services. This includes details such as who initiated an agent’s action, what data was accessed, what decisions were made, and what outputs were generated. This comprehensive logging is vital for incident response, forensic analysis, compliance reporting, and establishing a clear chain of custody for AI-driven processes. The audit trail serves as the definitive record for understanding AI agent behavior and ensuring explainability.

5. Policy-Based Governance:
Lens Agents enables organizations to define and enforce central policies governing every aspect of AI agent behavior. These policies can dictate access rights, acceptable use cases, data handling protocols, and even levels of autonomy. For instance, a policy might restrict an agent from accessing customer financial data, or it might require human approval for any action that involves making external purchases. This centralized, declarative approach to policy management ensures consistent enforcement across all agents and environments, eliminating ad-hoc decision-making and standardizing AI operations.

6. Active Cost Control:
The computational demands of AI, especially for large language models and complex autonomous agents, can lead to spiraling costs if not carefully managed. Lens Agents integrates active cost control mechanisms, allowing organizations to set and enforce real-time usage and spending limits. Agents can be automatically paused or stopped when predefined budgets – at the organizational, team, or even individual agent level – are exhausted. This proactive approach prevents unexpected expenditures and ensures that AI resource consumption remains within budgetary constraints, making AI adoption more predictable and sustainable.

These features collectively empower enterprises to deploy AI agents with confidence, transforming them from potential liabilities into strategic assets that are fully integrated, secure, and compliant with corporate standards.

Official Responses and Strategic Vision

Miska Kaipiainen’s insights underscore the strategic imperative behind Lens Agents. His emphasis on the widespread, yet often uncontrolled, use of AI agents, particularly on individual devices, highlights a fundamental shift in how enterprises must approach security and governance. "The challenge isn’t just about managing AI in the cloud anymore; it’s about managing it everywhere it’s being used," Kaipiainen states. "Our goal with Lens Agents is to provide that universal control layer, ensuring that AI innovation doesn’t come at the expense of security, compliance, or financial stability."

This vision aligns with Mirantis’ broader commitment to open-source and cloud-native principles, adapting them to the burgeoning field of AI. By extending the trusted Lens platform, Mirantis is leveraging its expertise in simplifying complex infrastructure to tackle the equally complex domain of AI agent management. The company anticipates that Lens Agents will become an essential tool for organizations looking to harness the full potential of AI without inadvertently creating new operational and security risks.

Far-Reaching Implications: Accelerating Responsible AI Adoption

The launch of Lens Agents carries significant implications for enterprises striving to integrate AI responsibly and effectively. It promises to accelerate AI adoption by mitigating many of the associated risks that have historically deterred or slowed down enterprise-wide deployment.

Enhanced Compliance Frameworks:
Lens Agents is designed with enterprise compliance in mind, supporting critical regulatory frameworks such as SOC 2 Type 1 and ISO 27001. The platform’s robust audit trails, identity management, and policy enforcement capabilities provide the necessary evidence and controls to meet these stringent requirements. For organizations in highly regulated industries, this level of assurance is invaluable, turning a potential compliance nightmare into a manageable, auditable process.

Alignment with New Regulations (e.g., EU AI Act):
As regulatory bodies worldwide begin to formalize rules for Artificial Intelligence, platforms like Lens Agents will become indispensable. The EU AI Act, for instance, introduces comprehensive requirements for transparency, risk management, and human oversight for AI systems. Lens Agents’ ability to provide detailed audit logs, enforce policy-based governance, and allow for controlled autonomy directly addresses many of these emerging supervisory and accountability demands. By offering a framework for responsible AI operation, Mirantis helps organizations proactively align with future legal and ethical obligations.

Controlled Autonomy: The Balance of Innovation and Oversight:
One of the most powerful aspects of Lens Agents is its capacity to define and enforce "controlled autonomy." Enterprises can precisely determine the level of independence allowed for each agent – ranging from merely assistive roles that require human approval for every action, to fully autonomous operations within predefined, secure boundaries. This nuanced control allows organizations to incrementally increase an agent’s autonomy as trust is built and performance is validated, without relinquishing ultimate oversight. It strikes a crucial balance, enabling innovation through AI while ensuring that human control remains paramount when necessary.

Mitigating Shadow AI and Uncontrolled Costs:
By offering a centralized platform, Lens Agents effectively combats the proliferation of "shadow AI." It provides a sanctioned, secure channel for employees to utilize AI agents, bringing previously ungoverned usage under corporate control. The active cost control feature, in particular, will be a game-changer for CFOs and IT departments, providing real-time visibility and enforcement of budgets for AI consumption, turning what was once an unpredictable expense into a manageable operational cost.

In conclusion, Mirantis’ Lens Agents represents a pivotal step in the evolution of enterprise AI management. By providing a unified, policy-driven platform for governing AI agents across all environments, it empowers organizations to accelerate their AI initiatives with confidence, ensuring security, compliance, and cost-effectiveness. As AI continues to embed itself deeper into business processes, solutions like Lens Agents will be fundamental in shaping a future where intelligent automation is not only powerful but also responsibly managed and fully aligned with organizational objectives. The Early Access Program offers a critical opportunity for enterprises to begin shaping their AI governance strategies today.