Zero Networks Cuts AI Hype With Control

Zero Networks has launched AI Segmentation to target enterprise security gaps. It looks to give enterprises real control over the risks from AI agents. The company says this is about agents’ access to sensitive systems, the speed of attacks and difficulties with compliance. These are issues that IT security teams will recognise as AI agents are increasingly accessing enterprise systems and data.

Benny Lakunishok, CEO and Co-Founder, Zero Networks, said, “Most vendors are out there selling AI hype. We’re not. Zero Networks puts enterprises in control of AI – full stop. This isn’t just visibility. It’s real control.

“Real-time, deterministic control over AI agents, combined with AI-driven visibility and an integrated Compliance and Risk Engine that continuously scores risk, maps activity to frameworks like NIS2 and CIS, and flags what actually matters. While others are still watching dashboards, Zero is enforcing outcomes – stopping lateral movement and preventing threats from becoming business problems.”

Lakunishok want to move from detection to enforcement. Detection happens after the breach, while enforcement stops the breach first. For Zero Networks, the approach is to assume the breach will happen, but deny the movement.

Reinforcing Least Privilege Access

Least privilege is about minimising what can be accessed. While it has been a staple of security plans for years, privilege bloat means it is rarely enforced. That opens the door for excessive privilege, which enables attackers using stolen credentials to move through an organisation.

Zero Networks is using several capabilities to protect the organisation through the network rather than the endpoints. One of these is AI Lateral Movement control, which relies on identity controls and network-based Least Privilege. Operating at the network level allows it to remove excessive connectivity. This restricts the ability of an AI agent to move around the network.

There is also an AI-Powered Compliance Engine. It queries live network activity and analyses billions of connections. By assigning dynamic risk scores, security teams are able to see real-time behaviour and where risk is coming from.

Other capabilities

There are three features that Zero Networks has grouped under AI Visibility and Control. These focus on better management of SaaS, AI agents and protecting LLMs.

• AI SaaS Control: Governs which cloud AI services employees, users, and devices can access at the network layer. Organisations control access to sanctioned tools such as ChatGPT, Gemini, Copilot, and others while automatically blocking unsanctioned or shadow AI services across the environment.
• AI Agent Control: Provides visibility and enforcement for every AI agent operating in the environment. Organisations can identify which agents are running, what they’re accessing, and how they communicate, while enforcing the same identity-based controls applied to users and devices, with strict least-privilege boundaries on every interaction.
•  LLM Protection: Secures model infrastructure at the network layer so only authorised systems can reach it. By segmenting access to LLM environments, Zero Networks helps prevent tampering, poisoned data inputs, hidden backdoors, and compromised models from accessing critical resources.

The combination of these five features gives security teams much more granularity in what they can do. They use natural language queries to discover what is happening on the network. It can also use that data to create a graphical representation of who is accessing what on the network. It is easier to see bottlenecks and unexpected access visually.

Another capability is the ability to analyse behaviour across connections. This will help spot when an AI agent is trying to access areas that it should. Security teams can build baselines of expected activity and use that to discover anomalies. With UEBA coming back into fashion, this is a key tool for spotting potential risk.

Enterprise Times: What does this mean

Breaches are a fact of life for security teams. Despite trying to shift left to a more preventative approach, many are stuck in reactive mode. There are many reasons why, but solving them is not about spending more money on tools and people. Both of these just mask the issue but don’t shift the security perspective.

Zero Networks is looking at long-established approaches such as network segmentation and least privilege access. Combining the two of these to create a network-based least privilege solution is an interesting approach. Arguably, it is something that should have been done some time ago, but it takes time to set up.

The key component is getting control of all the privileges users have, and they bloat over time. With the arrival of AI agents acting as users, the issue of privilege has been in the spotlight. Ensuring that all agents have a unique identity is a starting point. Controlling the privileges each gets is the next step. This announcement assumes the first and uses the second, combined with network controls to limit risk.

It will be interesting to see how different organisations respond to this announcement. Will customers use this as a focus point to create more restricted agents? How will users respond to that? Will it create a chance to review user privileges and gain control of those? There are many more questions to be asked, so it will be interesting to see what Zero Networks says about adoption in a few months’ time.

The post Zero Networks Cuts AI Hype With Control appeared first on Enterprise Times.