Every company I talk to right now is trying to answer the same question:

How do we move faster with AI without taking on more risk than we can actually manage?

That’s not just a security question. It’s a leadership question.

I’m a CEO. I want our teams using AI everywhere it makes sense in how we build, operate, support customers, and compete. At the same time, I’m accountable for trust. For uptime. For reputation. For making sure innovation doesn’t outpace our ability to control it.

So, when CISOs and CIOs tell me they feel caught in the middle, pressured to enable AI adoption while being held responsible for the risks it introduces, I don’t see resistance. I see realism.

The AI enterprise changes the physics of risk

In our recent announcement, I said that modern networks have “changed the physics of cyber risk.” That wasn’t meant to sound dramatic. It’s meant to be literal.

Today’s enterprise is always on. Always connected. Always changing. AI systems, agents, and automations are making decisions and moving data at machine speed. Non-human identities now outnumber people. And everything including users, workloads, services, and AI are all tied together through the network.

At that point, the network stops being just infrastructure. It becomes the nervous system of the business, where identities act and data moves.

That matters because attackers understand this shift just as well as we do. They don’t need to “break in” anymore. They log in. They exploit trust. They blend in. And with AI, they can move faster than most traditional security processes can keep up with.  

This isn’t about smarter attackers, it’s about speed

I don’t think attackers suddenly got smarter. They got faster.

AI removes friction. It automates reconnaissance. It accelerates lateral movement. It compresses timelines. What used to unfold over days now happens in minutes.

Meanwhile, most security teams are still working with stacks that introduce delay by design:

• Tools that only see one slice of the environment

• Signals that don’t connect cleanly

• Manual triage and investigation at exactly the wrong moment

That’s why so many teams feel like they’re reacting instead of defending, even when they’re doing everything right.

Why this is so hard for CISOs (and why I empathize)?

The hardest part of this job today isn’t technical. It’s explanatory. Boards and executives want clear answers:

• Are we safer right now?

• Where are we exposed?

• Are the controls we’ve invested in actually working?

Those are fair questions. But when visibility is fragmented and signal is buried in noise, even the best security leaders are forced to answer with partial information and assumptions stitched together by hand. That’s not a failure of leadership. It’s a limitation of the model.

Endpoint security still matters, but it’s just not enough

Endpoint tools play an important role. We use them. Our customers use them.

But endpoints don’t represent the full enterprise anymore. They don’t show how identities behave across systems. And given the proliferation of non-human identities including machines, service accounts, and now AI agents that operate across systems at machine speed, endpoints don’t see service-to-service activity, SaaS privilege abuse, or how AI agents operate across environments. Endpoints also don’t reveal how attackers move between systems on a network, which is exactly where modern attacks live. Relying on any single vantage point creates blind spots, and blind spots are where confidence breaks down.

How I think about “preemptive” and “proactive” as a CEO

In our press release, we talk about preemptive security and proactive defense. Let me translate what that means in plain terms.

Preemptive security is about reducing exposure before something bad happens. It’s knowing where identities, trust relationships, and automation paths create risk so they can be fixed early, while you still have time.

Proactive defense is about stopping attacks as they start, not after they’ve already spread. It’s removing latency from how quickly defenders get answers, understand what’s happening, and act.

Both are really about the same thing: giving security teams back time and clarity in an environment that doesn’t slow down for anyone.  

Our role at Vectra AI

We don’t see ourselves as the hero of this story. We see ourselves as co-defenders.

Our job is to help teams see what’s actually happening across the enterprise as it moves across networks, identities, cloud, SaaS, and now AI agents. Then to surface clear, behavior-based signal that tells you what matters now. Not more alerts. Not more fancy dashboards.

Just accurate, trusted answers, faster.

Because when leaders have clarity, decisions get easier. Conversations with Boards get more grounded. And security becomes something that enables the business instead of constantly slowing it down.

Closing thought

AI isn’t optional. It’s the new operating model.

The companies that succeed won’t be the ones that avoid risk; they’ll be the ones that understand it well enough to move forward with confidence. That’s the balance I’m trying to strike as a CEO. And it’s the same balance I see CISOs, CIOs, and Boards navigating every day. You’re not wrong to feel the tension.

‍
You’re not alone in it. The goal isn’t perfection.
It’s resilience, built for a world that now moves at AI speed.