SuperCom Ltd. (Nasdaq: SPCB) has introduced SafeGenie, a cybersecurity product suite designed to monitor and control actions taken by autonomous AI agents in real time. The company said the software is intended to allow, flag or block an agent’s activity before it completes, giving security teams a control point for agents that operate with access to enterprise systems and data.
SafeGenie combines new agent-security methods with endpoint technology from Safend, SuperCom’s wholly owned cybersecurity subsidiary. According to the company, the suite evaluates the context and intent behind an agent action, then applies an organization’s policies to decide whether the action should proceed. It is also designed to address conventional generative-AI use, including prompt protection.
The product addresses a practical gap in enterprise security operations. Traditional endpoint and network controls were largely built around human users, devices and approved applications. AI agents can research, move data and perform multi-step tasks on their own, potentially using the permissions of the employee or service account that authorized them. That creates a need to review not just the model or prompt, but individual actions taken at runtime.
SuperCom said SafeGenie works at the endpoint, at the kernel level beneath the applications and agents it governs, rather than relying only on network-traffic inspection. The company said each decision follows an organization-defined policy instead of an AI model’s judgment. It also said the local design is intended to protect supported endpoints whether they are online or offline, including endpoints running AI agents locally.
For security teams, the distinction is important because an agent can make changes or access information much faster than a conventional review process can intervene. A system that evaluates and blocks a high-risk action before completion could give organizations an additional enforcement layer for data movement, tool access and other sensitive steps. SuperCom’s claims about the product’s efficacy have not been independently verified, and its usefulness will depend on policy design, integrations and deployment coverage.
The initial rollout will focus on integrating SafeGenie with Safend’s infrastructure and offering it to Safend customers through direct and distribution channels in the United States, Europe and Asia, the company said. SuperCom then plans to offer it to government customers and other enterprises. It said SafeGenie will be sold as a separate subscription alongside existing Safend licenses and maintenance agreements.
The vendor said Safend’s endpoint technology has been deployed across millions of endpoints and thousands of customer environments. That existing footprint may provide a route to deployment, but the company did not give a specific availability date or identify which agent platforms will be supported first.
The immediate enterprise benefit is a proposed mechanism to put policy enforcement closer to the endpoint where an AI agent acts. Organizations considering the product would still need to define permitted behavior, manage identities and monitor the controls themselves. As autonomous agents become more common in business workflows, that runtime-control layer could become a more routine part of endpoint and AI governance programs. The company did not disclose subscription pricing. The company also did not identify an exact launch timetable.