CRN talks to CEOs of major cybersecurity vendors, including CrowdStrike, SentinelOne, and Netskope, about where they see AI agents going next and what they’re keeping a close eye on. Here’s what they had to say.
While it is clear that the adoption of AI agents will increase the demand for cybersecurity tools and services going forward, top cybersecurity CEOs pointed out that many uncertainties remain about how the technology will impact the workforce and the industry’s approach to security in the future. crn But RSAC 2026.
During a major industry conference in San Francisco last week, crn We spoke to the CEOs of 15 top cybersecurity vendors, including CrowdStrike, SentinelOne, and Netskope, about where they see AI agents going next and what they’re keeping a close eye on.
(Connected: 20 best AI and security products at RSAC 2026)
CEOs of other major players spoke crn These include Arctic Wolf, Proofpoint, Sophos, Mimecast, 1Password, Absolute Security, Savient, Huntress, Delinia, Orca Security, KeyFactor, and Zaffran Security.
Recurring themes include how agentic cyber is reshaping risk and creating more work for security teams, while also raising questions about whether AI will eventually replace or augment large portions of the workforce.
The CEO noted that importantly, solution and service providers will play a key role in enabling the next wave of agent adoption. crnWhich is likely to surpass even earlier technological changes in terms of accelerating cyber security developments. Overall, “this is probably the biggest market opportunity I’ve ever seen in my life,” said Sophos CEO Joe Levy.
What follows are insights from 15 top cybersecurity CEOs on the future of AI agents.
George Kurtz, Co-Founder and CEO, CrowdStrike
Huge cyber risk from uncontrolled agents: “When you look at the evolution of AI over the last few years – you had GenAI, then you had logic chains and then you had (AI) doing things. … I think (the risk) is really around the OpenClause model, where you have agents that are running in the context of the user on your desktop, with access to all your data files, everyone is plugging their credentials into Box and Dropbox and Google Drive to plug in their email and Every other thing that’s out there (if there are agents) that has access to shells, access to data and workflows – so how do you know what’s going on? There were a lot of malicious skills introduced in OpenCL in February – if you’re poisoning the memory, (attackers) don’t have the ability to attack the supply chain and just wait.
Tomer Weingarten, Co-Founder and CEO, SentinelOne
AI creates more security work: “If we think about these agents as fundamentally more employees, how do you scale your security operations? You’re not going to be able to hire that fast. Yes, we’re getting more automated. Yes, there’s more autonomy in the SOC – but you still need human supervision. And I think the scaling is going to come from the partner ecosystem. … We’ve been talking here for a couple of years, and every year I can say the same sentence – which is Is, ‘There aren’t enough people in cyber security.’ And this is going to be true every year. I don’t think that’s going to change with agents either. AI is not going to change this. Because AI for cybersecurity isn’t just something that helps enhance the workforce, as it does for every industry – it also creates more work for cybersecurity operators. So this is a segment where it is not only automating work, (but) it is also creating more work.
Sanjay Berry, Co-Founder and CEO, Netscope
CISOs are moving closer to agentic faster than ever before: “I think the biggest uncertainty is how fast can a CISO move to get the governance of their AI going, while maintaining the pace of adoption of that AI in their company? How do they not slow down the adoption of AI in their company, yet get it under control quickly? That’s their biggest challenge… I think what it did for the industry, it inspired[CISOs]to go ‘back to school’ to learn Every CISO and head of networks is learning about AI and how to use it faster than I’ve ever seen them. They may be scheduling a project for a year – either way, a year is not equal to a decade.
Sumit Dhawan, CEO, Proofpoint
Security teams should step aside: “The problem right now is that cyber teams are standing somewhat on the sidelines. … AI has become a priority for CEOs (and) cyber teams are not really in a position to slow it down. So again, as a result, they’re in a bit of a ‘wait and see’ (position) as it’s happening. Every vendor is enabling AI in their products, and every business user is bringing AI in, because there’s a top-line directive to adopt it as quickly as possible. … But (Businesses) want and need cyber – because they understand that all these new technologies can bring down their business if they don’t have the right security.
Nick Schneider, CEO, Arctic Wolf
The industry is still learning what can be automated on security: “You will see a steady increase in the leverage of agents within security operations. I think there are certain workflows that can be fully autonomous, and there are other workflows that will not be fully autonomous. And I think we’re still a ways away from not having a human-in-the-loop as a prerequisite for most organizations. Some of this has to do with the company’s risk curve. There are certain tasks or certain processes or certain workflows where it can be fully autonomous. Is – and if it’s wrong, or if it’s a little bit wrong, it’s not the end of the world. There are other places where, if it’s wrong, it’s a big deal and that’s where you still want some human validation, some human trust.
Joe Levy, CEO, Sophos
Unprecedented opportunities in security: “Hundreds of millions of businesses are going to undergo this transformation, and there is no sector or area or size of company that will be untouched by it. This is an economic wave that is going to sweep across the planet. How equipped do most organizations feel to deal with it? This is possibly the biggest market opportunity I have ever seen in my life.”
Christy Wyatt, CEO, Absolute Security
More intelligence needs to be transferred to the device: “The intelligence for all of this really needs to sit with the device itself. I think our approach is long-term. We talk about self-healing. It’s self-healing of the application, self-healing of the network connection, self-healing of the operating system and the device itself. And so obviously, it all happens at the endpoint. And as it all accelerates, long-term, you would want that device to understand its condition and be able to self-heal. Be able to.”
Sachin Nayyar, CEO, Savient
Structure for agents is uncertain: “We don’t know if we’ll have one account per agent, one account per user, one account per role. How the agent architecture will be set up – that hasn’t been revealed yet. So as we’re working with customers, we’re trying to figure out how many accounts we need to create to give an agent access. … So the agent architecture is undecided at this point. Companies are still working through that.”
David Fogno, CEO, 1Password
The role of stakeholders in spreading the word about AI: “Ninety percent of people in the country… don’t understand (AI). And those people run businesses. And so (there’s an opportunity around) helping people understand how to harness the future the right way and not put their head in the sand. … When you embrace it, it becomes something that can really add value to helping you. And so partners can really help people see it. But partners themselves have to lead by example.”
Mark van Zedelhoff, CEO, Mimecast
There may not be a new paradigm for AI security: “At the end of the day, you have to think about networks, endpoints, applications, data and identity. And then, of course, the humans who use it. Those are the principles of cybersecurity. They’ve been around forever. And so the question is, do we have to re-invent that stack (for AI)? Is AI a fundamentally new paradigm for securing? Or is it another technological revolution that needs to be secured with the old paradigm? I think it’s the latter. It’s a matter.”
Art Gilliland, CEO, Delinia
Compromising agents are inevitable: “The challenge is you don’t know all the places where AI breaches are going to hit you. I don’t know of any (cases) where an AI agent was taken over and that’s what caused a breach. But it’s going to happen. Because as more companies adopt it, it will be a weak spot in the organization. It will be a machine-speed, deeply integrated privileged user in your environment.”
Kyle Hanslovan, Co-Founder and CEO, Huntress
It’s not enough to keep humans “in the loop”: “(AI) requires humans, not just humans in the loop. So they have to guide the AI. They have to guide the detection research. And yes, some of these LLMs are really good at finding the low-hanging fruit. But here’s the wild part—as AI is really giving rise to creative, organized cybercriminals, it’s creating new tradecraft. … (This means) humans have to lead the solutions.”
Gil Geron, Co-Founder and CEO, Orca Security
There is uncertainty about whether AI can really replace workers: “It’s uncertain whether these tools will reach the maturity of simply co-piloting versus workforce replacement. What we’ve seen is that we’re helping everyone get better, including more junior safety engineers. And the question is whether there’s going to be a time when it’s actually going to replace the workforce and not just augment them. The effect of that is that if it’s a real replacement, you’re taking a lot of responsibility on yourself.”
Jordan Racki, CEO, KeyFactor
AI can reduce “Q-day” time: “How is AI impacting the quantum transformation? The (teams) that are building quantum computers are leveraging AI to speed up their timelines in quantum computers. And so while AI is beneficial in many ways, it’s actually shortening the timelines that organizations need to be ready for (for post-quantum encryption). AI is allowing those people to think more strategically and do more cross-referenceable analyses. Is giving, so that computers can be made faster.”
Sanaz Yashar, Co-Founder and CEO, Zafran Security
Employers are still unclear about the role of AI in the workforce: “I think the biggest uncertainty is, are we going to use AI instead of humans? Or are we going to be forced to augment the same workforce with AI? This is the dance we see companies[doing]with AI. On the one hand, they understand that without AI, they will be left behind. On the other hand, when they are using it, they have to show efficiency. How are they going to drive efficiency if they are not reducing manpower or services? So I think a decision needs to be taken.
