SuperCom has introduced SafeGenie, a cybersecurity product suite designed to monitor and control autonomous AI agents in real time. The product combines patent-pending agentic AI security methods with kernel-level endpoint security technology from Safend, SuperCom’s cybersecurity subsidiary. It is intended to help enterprises govern AI actions, prevent sensitive-data exposure and enforce security policies as AI agents gain access to business applications and information.
AI Agents Create a New Endpoint Security Challenge
Enterprise AI adoption is moving beyond chatbots that answer questions toward agents capable of carrying out multi-step tasks across applications, files and business systems. These agents can operate with permissions granted to employees or service accounts, potentially accessing sensitive information and executing actions faster than human supervisors can review them.
That autonomy creates a security challenge for tools originally designed to protect users, devices and conventional applications. An AI agent may follow malicious instructions embedded in an email or document, access information beyond its intended purpose, or make an unauthorized change while performing an otherwise legitimate task. Unapproved tools and agents running under personal accounts can further complicate enterprise oversight.
SuperCom says its new SafeGenie suite is designed to address these risks by evaluating the context and intent of an agent’s actions and allowing, flagging or blocking them before execution completes. The company describes the technology as a control layer for agentic AI, with additional protection for conventional generative AI use, including prompt security.
The launch comes as security teams reassess how to govern AI systems that can act on behalf of users. Gartner has forecast that spending on securing AI will rise from approximately $2.8 billion in 2026 to nearly $7.7 billion in 2028, according to figures cited in SuperCom’s announcement. That forecast reflects a broader market opportunity, but it does not establish demand for SafeGenie specifically.
Kernel-Level Controls Target Agent Actions
SafeGenie’s architecture builds on Safend’s endpoint security technology, which operates at the operating system’s kernel level. The kernel manages fundamental system resources and provides a layer beneath ordinary applications. Security controls implemented at this level can potentially observe or restrict certain activity closer to where it occurs, rather than depending exclusively on network monitoring.
SuperCom says SafeGenie combines this endpoint foundation with new, patent-pending methods developed to address agentic AI risks. Its approach is intended to enforce organizational policies locally, including for supported endpoints that are offline or running AI agents on-device.
That local enforcement could matter for organizations using coding agents, desktop assistants or other AI tools that interact directly with files and applications. Network monitoring alone may not provide complete visibility into activity that takes place locally, particularly when tools operate outside approved enterprise infrastructure.
However, the announcement does not disclose the technical details of SafeGenie’s agent-detection methods, supported operating systems, integration requirements or independent test results. The extent to which it can distinguish legitimate agent actions from harmful ones will depend on policy configuration, endpoint coverage and how the system handles different agent architectures.
SuperCom President and CEO Ordan Trabelsi described the product as a way to preserve AI productivity while stopping high-risk actions. The company argues that combining new agent-focused controls with Safend’s existing endpoint technology gives it a foundation for serving both current cybersecurity customers and new enterprise accounts.
Initial Rollout Targets Existing Safend Customers
SuperCom plans to introduce SafeGenie first through Safend’s existing customer base and distribution channels across the United States, Europe and Asia. The company then expects to extend availability to its government customers and additional enterprises.
SafeGenie will be sold as a separate subscription alongside existing Safend licenses and maintenance agreements. This gives SuperCom a potential route to market through established cybersecurity relationships, including customers in government, defense and enterprise environments.
The company says Safend’s endpoint technology has been deployed across millions of endpoints and thousands of customer environments. That installed base could provide an initial distribution advantage, although the announcement does not disclose subscription pricing, launch dates, supported platforms or expected revenue.
For customers, the main consideration will be whether SafeGenie can integrate with existing endpoint protection and security operations processes without introducing excessive alerts or disrupting legitimate automation. Organizations will also need clear policies governing which AI agents are approved, what permissions they receive and which actions require human authorization.
Market Landscape
Enterprise AI security is expanding beyond conventional endpoint protection, identity management and network monitoring. Autonomous agents can interact with files, applications and business systems using delegated permissions, creating risks that span prompt injection, excessive access, data leakage and unauthorized actions.
Gartner forecast in August 2026 that spending on securing AI would reach nearly $4.8 billion in 2027 and $7.7 billion in 2028, up 68.7% from 2026 to 2027. Gartner also expects more than half of successful attacks on AI agents by 2029 to exploit access-control weaknesses and prompt injection.
SafeGenie enters this market with an endpoint-focused approach that combines kernel-level controls with agent-specific security methods. Its commercial success will depend on the breadth of agent coverage, the quality of its policy enforcement, compatibility with enterprise environments and independent evidence that it can stop harmful actions without blocking legitimate work.
Top Insights
- SuperCom has introduced SafeGenie, a security suite designed to monitor, flag and block risky actions by autonomous AI agents.
- The product combines patent-pending agentic AI security methods with kernel-level endpoint technology from SuperCom subsidiary Safend.
- SafeGenie is designed to enforce organizational policies locally, including on supported endpoints running offline or using local AI agents.
- The initial rollout will target Safend’s existing customers and distribution channels across the United States, Europe and Asia.
- SafeGenie will be sold as a separate subscription, with broader availability planned for government customers and additional enterprises.
Power Tomorrow’s Intelligence — Build It with TechEdgeAI
