A modern cyber range provides teams with a safe, realistic environment to practice responses to attacks, helping them see how people, tools, processes, and AI perform under pressure. This is crucial because disruptions in manufacturing are not just about data.
The State of Agentic Cybersecurity
Q&A with Jason Allnut, OT & Cyber SME | SimSpace
Tell me more about Yourself and SimSpace?
I have nearly 20 years in the security industry and currently serve as OT & Cyber SME at SimSpace. Our goal is to be the AI Proving Grounds for cybersecurity. We provide a realistic, intelligent cyber range where organizations can train people, test security tools and AI agents, and validate workflows before they are trusted in live operations. The organization was established by pioneers from U.S. Cyber Command and MIT Lincoln Laboratory, and its mission has consistently been grounded in a fundamental principle: readiness must be demonstrated, not presumed.
Why is Cyber Range technology key for today’s industry?
Manufacturers are managing intricate environments that integrate IT, OT, cloud, identity, endpoint, and physical systems. Protecting such a diverse setup requires more than just tabletop exercises. A modern cyber range provides teams with a safe, realistic environment to practice responses to attacks, helping them see how people, tools, processes, and AI perform under pressure. This is crucial because disruptions in manufacturing are not just about data—they can quickly escalate into operational, financial, and safety issues.
Is AI making it more or less difficult to protect today’s companies?
In our opinion, we believe it’s both. AI can help security teams move faster, analyze more signals, and improve response. At the same time, it raises the bar because attackers can use AI to scale reconnaissance, phishing, malware development, and adaptive campaigns. What we’ve seen in analyst and customer conversations is that everyone agrees companies should use AI. It’s now about deployment and trust. The trust comes into play when companies, especially security teams, can prove how their AI performs before it’s deployed into real security operations.
Why is AI making things both more and less difficult in your opinion?
We think one of those reasons is that AI introduces new failure modes. A human analyst can be trained and evaluated. A security tool can be tuned and tested. AI agents need the same rigor. SimSpace’s recent report found that 78% of security leaders have high confidence in their defenses, yet teams scored as low as 30% in security exercises. It also found that 73% of organizations are already using AI agents in the SOC, while only 29% conduct continuous simulation testing. That leads to a lack of trust and frankly understanding of what is going into production environments. Would you trust a pilot who has never flown in a simulator or on practice training routes? Why would you expect something less from security when it’s so important?
Why is that important?
People will tell you and it’s instinctive: AI cannot be trusted on promise alone. Before a manufacturer allows AI to assist with operational decisions, incident response, detection engineering or threat hunting, it needs to know how that AI behaves when conditions are messy. If manufacturers are really worried about AI integration and the effects it could have its going to be on it's impact to availability, as keeping the processes running keeps the money flowing. A cyber range gives companies a safe place to fail, tune, and improve before those systems are exposed to production pressure.
How is the industry responding?
The industry is moving quickly toward AI-assisted security, but validation is lagging. SimSpace’s report shows that companies are adopting AI agents faster than they are testing them. That creates a confidence gap. Security leaders believe AI can improve performance, but many do not yet have the proving grounds needed to measure whether those systems actually improve readiness.
You recently had a survey. Can you tell me more about that?
SimSpace’s report, The State of Agentic Cybersecurity, combined survey data from global CISOs and senior security leaders with readiness data from SimSpace environments. The research looked at how organizations are training human operators and testing AI agents to work together in realistic conditions. The key finding is clear: confidence in AI-driven security is running ahead of measurable performance.
What does that say about the industry and the future of cybersecurity?
It says cybersecurity is entering a proof-driven era. Companies can no longer rely on certifications, annual exercises, or vendor claims. They need continuous mission rehearsal, realistic testing, and objective readiness data. For manufacturers, this is especially important because resilience depends on whether teams can keep operations moving while defending against real-world attacks.
In the next 10 years, where do you see the industry heading?
Cybersecurity is projected to evolve into a more continuous, measurable, and AI-driven discipline. Leading organizations will regard readiness as an ongoing operational practice rather than an annual event. AI agents will be incorporated into security teams; however, they will also necessitate training, testing, and accountability, akin to other elements of the defense system. The industry is transitioning from presumed readiness to demonstrable readiness.
The content & opinions in this article are the author’s and do not necessarily represent the views of ManufacturingTomorrow
Featured Product
