Zero Networks targets AI agent risk with expanded Palo Alto tie-up

Cybersecurity vendor Zero Networks has expanded its integration with Palo Alto Networks, adding AI traffic inspection and automated threat containment capabilities amid growing concerns around AI agents gaining access to enterprise systems

Cybersecurity vendor Zero Networks has expanded its integration with Palo Alto Networks, adding AI security and automated threat containment capabilities as vendors look to tackle one of the industry's newest challenges: controlling autonomous AI agents inside enterprise environments.

The move comes as security vendors are increasingly shifting their attention from traditional perimeter defence towards securing AI-powered tools and agents.

This year has already seen high-profile AI agent hacking incidents that gained access to corporate applications and infrastructures.

In July 2026, OpenAI disclosed that two of its frontier models, including GPT-5.6 Sol, autonomously compromised the Hugging Face platform during capability evaluations.

Following OpenAI's announcement, Anthropic revealed three "real-world incidents" where its Claude models escaped isolated testing environments.

Zero Networks said the enhanced integration extends an existing partnership that already combines its automated microsegmentation technology with Palo Alto Networks' firewall platform.

The latest update introduces the ability to redirect traffic for deeper inspection, surface controls through Palo Alto Networks' Strata Cloud Manager platform and apply security policies to AI agents using a combination of Zero Networks' AI Segmentation capabilities and Palo Alto Networks' Prisma AIRS platform.

AI becomes the new attack surface

The announcement reflects a broader trend across the cybersecurity market, where vendors are racing to develop products designed to govern, monitor and restrict AI activity inside organisations.

AI agents are increasingly being granted access to business applications, databases and infrastructure, creating fresh security concerns for enterprises attempting to balance productivity gains with risk management.

Zero Networks argues that organisations should not have to manage separate security controls for AI environments and traditional infrastructure.

“Customers should not have to stitch together separate controls for their firewalls, internal assets, and AI agents,” said Benny Lakunishok, co-founder and CEO of Zero Networks.

“This expanded integration creates one continuous enforcement model. Zero Networks controls which connections are allowed, Palo Alto Networks provides deep inspection, and together we can contain malicious activity before it can move across the environment.”

According to the vendor, the integration enables customers to identify AI agents, restrict where they can connect, block unsanctioned AI services and inspect selected AI traffic for potentially malicious or unauthorised activity.

It also follows a revamp to its partner programme earlier this year.

Reducing lateral movement

Alongside the AI capabilities, the integration also deepens Zero Networks' core microsegmentation proposition.

The company said customers can automatically discover assets, create least-privilege access policies and synchronise asset information with Palo Alto Networks firewalls. Traffic can then be redirected for deeper inspection, with suspicious activity blocked before it spreads across the network.

The approach is designed to limit lateral movement by attackers, an area that remains a key focus for organisations grappling with ransomware and increasingly sophisticated cyber threats.

Rather than requiring organisations to deploy separate controls for perimeter security, internal networks and AI applications, Zero Networks said the combined offering can be applied across on-premises, cloud, operational technology (OT), Kubernetes and hybrid environments.

Simplifying security operations

Another aspect of the integration is visibility.

Zero Networks-managed assets and policies can now be surfaced through Palo Alto Networks' Strata Cloud Manager platform, giving security teams a single management interface spanning asset discovery, segmentation and traffic inspection.

The vendor said dynamic tagging automatically updates security policies as workloads move, IP addresses change and new assets appear, reducing the need to manually maintain large volumes of firewall rules.

While security vendors have long promised greater automation, the latest announcement highlights how AI security is rapidly becoming a key battleground as suppliers look to convince customers they can securely deploy increasingly autonomous technologies.

The expanded integration is available immediately, including automated microsegmentation, dynamic asset tagging, policy synchronisation and Linux traffic redirection. Windows support is planned for a future release.