Forget AI Moonshots. Focus on Automating the Boring Stuff First.

We are wasting a lot of time and money trying to solve the 0.00000001% of problems that are intellectually interesting but irrelevant to most real-world use cases.  

We chase AGI-like autonomy, semantic perfection, and context-aware magic while ignoring the huge value in automating the 99.999% of repetitive, obvious tasks that bog teams down every day.

Nowhere is this clearer than in the contrast between fully autonomous driving systems and hands-free driving assistance features now being developed by top car companies.

And in security, it shows up in how we think about cloud visibility, observability, and Zero Trust.

Fully autonomous AI tech misses the point

Full self-driving (FSD) systems are a technical moonshot. They want to solve the general driving problem: any road, any condition, no hands.  

That sounds amazing — until you realize how unnecessary it is. Most people spend nearly all their drive time doing two things:  

• Staying in their lane
• Not hitting the car in front of them

On the other hand, hands-free driving systems have figured this out. Instead of building an all-knowing driving AI, it automated the boring part.

On mapped highways, with good weather and reasonable lines, it works hands-free. It doesn’t need to handle snowbanks, traffic cops, or weird intersections. It does the job.

And that’s the point. The hardest problem isn’t always the one worth solving. The better question is: what can I take off your plate, right now, that you’re doing over and over again, poorly, with too much effort?

What cybersecurity can learn from self-driving cars

In cybersecurity, particularly around segmentation, we see the same pattern repeating.  

There’s a temptation to generalize the problem — to build a system that can dynamically and perfectly enforce least privilege everywhere, for every workload, user, and device, at all times.  

The appeal of a fully autonomous, AGI-like Zero Trust engine is strong.

But just like FSD, pushing too far too fast, especially in high-stakes environments, can lead to fragile systems, unexpected breakage, and teams pulling the plug after a bad experience.

The appeal of a fully autonomous, AGI-like Zero Trust engine is strong.

The better model, at least today, is the hands-free driving model. Scope the problem tightly, define the safe and automatable zone of control, and help users drive better and safer in the 99% of environments where they operate most.

What security teams actually need right now

Only a small percentage of enterprises are enforcing comprehensive default-deny policies across their internal environments.  

While default-deny is common at the perimeter, very few organizations have extended that rigor laterally within the network or across cloud-native workloads.

In contrast, a large number of organizations are actively investing in telemetry, visibility, observability, and actionable insights. The market is clearly signaling that understanding exposure, not enforcing it, is the primary and immediate need.

That’s where we are focused today, using correlated telemetry and system signals to highlight meaningful security insights such as:

• “This internal service is accepting connections from the internet.”
• “This deprecated system is still communicating with production.”
• “This workload is bypassing segmentation policy via indirect access.”

These are insights that require no policy enforcement to deliver immediate value. And in many cases, they surface risks that teams didn’t know existed.

Automation that supports humans, not replaces them

The most useful automation today isn't autonomous, it's assistive. We’ve seen this proven in every other domain:

• GitHub Copilot doesn’t write code alone. It accelerates the human.
• Grammarly doesn’t write stories. It helps tighten the prose.
• Ford BlueCruise doesn’t drive the car. It reduces fatigue and error.

Security teams want the same kind of help. The goal is not to replace analysts or admins, but to make them faster, safer, and more consistent.

Illumio Insights does exactly that. It automatically correlates low-level system state and signals, elevates what matters, and gives teams the context they need to act on threats with confidence and speed.

Moving toward full enforcement — but smarter

Ultimately, segmentation is about containment, reducing blast radius, limiting exposure, and stopping lateral movement.

Declarative policy enforcement is an essential part of that journey. And we’re committed to continuing to make that easier, less high risk, and more scalable.

But today, we’re meeting customers where they are. And most are still in the visibility phase. That’s not a failure — it’s a necessary and productive step.

We’re making that phase smarter and more valuable with Illumio Insights by providing the kind of context that lets organizations make progress, even before full enforcement is turned on.

Think of it as assisted security driving. The system sees more than you do, tells you when you're drifting, and alerts you when you're speeding toward a cliff, even if it doesn't hit the brakes for you yet.

Boring security is the best security

The future of Zero Trust isn’t an all-or-nothing switch. It’s a continuum, and most of the industry is still near the beginning.

The job today isn’t to force enforcement before teams are ready. It’s to illuminate what’s happening, highlight what matters, and build confidence along the way.

If that sounds boring, that’s because it is. But boring is exactly what security should be: predictable, observable, and safe.

That’s where we’re investing — in helping our users get the most value out of the environments they already have, with insights they can trust, and automation that supports progress without risk.

We’ll get to full enforcement. But we’ll get there safely, one insight at a time.

Learn more about Illumio Insights today.