Cloud Range Rolls Out Validation Range for Secure AI Testing

Cloud Range on Tuesday launched its AI Validation Range, a cyber range platform designed to help organizations securely test, train, and validate AI models and agentic AI before deployment in production environments.

Balancing AI security and human judgment

According to Cloud Range, its new platform is designed to address the rapid adoption of unmanaged AI tools and agentic AI, especially those that have been integrated within organizations but have yet to be evaluated for safety.

AI Validation Range allows organizations to verify AI performance and reliability before deployment by testing and measuring how models respond to real adversarial inputs and uncertainty. 

“For years, Cloud Range has helped organizations know how to perform under real attack conditions. Applying that same simulation rigor to AI allows organizations to measure how AI agents and models perform side by side with human defenders, using the same scenarios, tools, and pressures,” said Cloud Range CEO Debbie Gordon. 

“That comparison is critical to understanding where AI truly strengthens security and where human judgment still matters most,” she added.

How the AI Validation Range tests agentic AI in SOC environments

For organizations integrating agentic AI into SOC, cyber defense, and offensive security workflows, the new platform supports training agents on real systems and observing how they interact with live infrastructure and security controls. 

Cloud Range says this controlled approach to developing AI agents and assessing models gives security and engineering teams clearer insight into AI reliability, decision logic, and failure modes, allowing them to establish guardrails and reduce risk. 

Some of its key capabilities include:

• Adversarial AI Testing: Simulate real-world cyber attacks to evaluate how AI models and agents detect, respond, and adapt under hostile conditions.
• Agentic SOC Training: Condition AI agents on how to defend against real cyberattacks, workflows, and response actions in a safe, non-production environment.
• Operational Readiness Validation: Measure AI performance and implement security controls to determine production readiness and identify gaps before deployment.
• Governed, Repeatable Experiments: Support controlled testing and repeatable scenarios to enable consistent validation, tuning, and improvement over time.
• Secure, Isolated Range Environment: Protect production systems and model data integrity while enabling high-fidelity simulations and training exercises.

How Cloud Range supports secure AI deployment in cybersecurity

Cloud Range says businesses can use its catalog of real-world attack simulations and licensed security tools to safely test AI models for data leakage, logging behavior, and unintended outputs in realistic IT and OT/ICS environments.

In addition, they can use the platform to train agents on offensive security objectives such as vulnerability discovery to scan networks and validate real threats. The platform also supports defensive use cases such as identifying malicious behaviors and accelerating alerts.

With these capabilities, Cloud Range says it aims to help security teams gain clearer visibility into how AI systems behave, where safeguards are needed, and how responsibility should be shared between technology and human personnel.

“This enables organizations to operationalize AI with confidence, aligning innovation, security, and accountability before AI becomes embedded in mission-critical workflows,” Cloud Range said in an official statement.

Last year, we spoke with enterprise leaders to explore how agentic AI and mounting ROI pressures could reshape channel strategies in 2026. Learn more about their insights and other technology predictions for the year ahead.