The problem, in one image: a hotel keycard that opens your room, the gym, the rooftop bar, and — because nobody ever checked — the manager’s office safe. That’s most corporate networks. One badge, one login, and suddenly you’re everywhere.
The Lesson I Give Before Every Lead
I teach sensual bachata at my studio, Cyber Dance. Before I let a new student lead their first close-embrace turn, I tell them something that surprises people: “Don’t trust your partner’s balance. Feel it, every half-second, and adjust.” A bad leader assumes the follower will be there. A good leader checks — through frame, through weight, through breath — that the connection is still real, turn after turn. The moment you stop checking is the moment someone falls.
I once had a beginner ask me, half-joking, “Why can’t I just trust her to follow? Isn’t that the whole point of leading?” I told him: trust isn’t a starting condition. It’s a continuous negotiation you renew every eight counts. That sentence, almost word for word, is the entire philosophy of Zero Trust — and I didn’t learn it in a SOC. I learned it on a dance floor.
The Badge That Opened Everything
Now the corporate version, and it has a name everyone in this industry knows: Target, 2013. Attackers didn’t break down the front door. They stole the credentials of a small HVAC contractor — a vendor who serviced refrigeration units. That single set of credentials, meant for a narrow, boring purpose, turned out to have a path into Target’s payment network. Forty million card numbers later, the postmortem read like a tragedy in one sentence: nobody ever asked “why does the air-conditioning guy have a route to the cash registers?”
That’s the core failure of the old model, the one we used to call “castle and moat.” Get past the perimeter — a VPN, a firewall, a badge at reception — and the inside of the network trusts you by default. One login was the whole performance. No one checked the frame again once the turn began.
Zero Trust, Without the Buzzwords
Strip away the vendor decks and the acronyms, and Zero Trust is a very old, very human idea dressed in architecture: trust nobody by default — including people, devices, and applications already inside your network — and re-earn that trust for every single action, every single time.
It’s not a product you buy. It’s a posture you build, made of a few honest questions asked continuously instead of once at the door:
- Who exactly is asking? Not “someone with valid credentials,” but this specific identity, on this specific device, at this specific moment.
- Do they need this, right now, for this task? Least privilege isn’t a policy line — it’s the difference between the HVAC vendor reaching the thermostat and the HVAC vendor reaching the cash registers.
- Is the context still normal? Same laptop, same geography, same behavior pattern as five minutes ago — or did something just shift?
- What happens if I’m wrong? If this identity turns out to be compromised, how far can the damage travel before something notices?
That last question is the one boards actually care about, because it’s the one with a dollar sign attached. IBM’s Cost of a Data Breach research has shown for years that breaches involving lateral movement — an attacker moving from the initial foothold deeper into the network — cost significantly more and take longer to contain than breaches that stay put. Zero Trust, done properly, is the architecture that shortens that journey to almost nothing.
What “Continuous Verification” Looks Like on a Tuesday
In practice, in the enterprise environments I assess, it’s rarely one dramatic control. It’s a layered set of unglamorous habits:
- Microsegmentation — networks cut into small rooms instead of one open floor, so a compromised laptop in finance can’t casually stroll into R&D.
- Privileged Access Management (PAM) — admin rights issued just-in-time, for a specific task, then revoked. Not a master key living in someone’s pocket for three years.
- Conditional, session-aware access — the system re-asks “is this still you?” when the signals change: new device, new country, 2 a.m. instead of 2 p.m.
- Explicit trust boundaries around AI agents — the newest and, frankly, the most under-examined risk. An AI agent with standing access to your CRM and your email is a follower who will do exactly what the last instruction told them, even if that instruction arrived hidden inside a poisoned document. I’ve written about that ghost before; Zero Trust is precisely the discipline that keeps it from having the run of the house.
The Boardroom Translation
When I bring this to a board, I don’t say “Zero Trust Architecture.” I say: right now, one compromised password can reach almost anything in this company. I want to build it so one compromised password reaches almost nothing. That sentence has a business owner nodding before the CISO finishes it, because it answers the only question that matters in the room: how much does a mistake cost us, and how fast does it spread. Formula 1 doesn’t trust its brakes because the car looks fast. It trusts them because every stop, every lap, someone re-measures.
Three Things to Take Back to Your Desk
- Audit your standing access, not your incidents. Ask: who has access they haven’t used in 90 days? That’s your attack surface waiting quietly.
- Map one “HVAC vendor” in your own environment. Every org has a narrow-purpose account with an unnecessarily wide reach. Find it before someone else does.
- Treat AI agents like new hires with root access. Give them scoped, revocable permissions and log every action — enthusiasm is not a control.
- Re-verify, don’t just re-badge. A login at 9 a.m. doesn’t mean the same identity is trustworthy at 9 p.m. Build systems that ask again.
Back to the Studio
That beginner who asked me why he couldn’t just trust his partner to follow — I gave him the turn again, slower this time, and told him to feel for her weight at every count instead of assuming it. He caught her the second time. Not because he trusted more. Because he checked more, and checked well enough that the checking became invisible — which, on a dance floor and in a network, is what real trust actually feels like: earned so continuously it stops feeling like effort at all.
I write From the CISO’s Diary from 20 years across BDO, ThriveDX, the IDF and my own ventures. Let’s connect: Omri Sagron on LinkedIn.
