What a telecom core network taught me about running a company.

Before Norvan, I spent eleven years inside telecom core networks — the systems that route the world’s phone calls and data. Not the apps. The core: the machinery that cannot fail, because when it does, a city loses its ability to communicate.

That world runs on a discipline most businesses never encounter, and it changed how I see what a company is.

A telecom core network has a property that sounds mundane and is actually profound: universal input, uniform processing. Signals arrive from wildly different sources, over different protocols, in different shapes — and the system feeds them all through one coherent pipeline that receives, validates, prioritizes, and routes every one. Nothing enters unseen. Nothing is processed twice by accident. Nothing falls through a gap, because there are no gaps — the architecture doesn’t permit them.

Now look at how a company handles its inputs. A lead arrives by email, another from the website, another from a referral at dinner, another from an event. Each enters through a different door, handled by a different person, tracked in a different place — or not tracked at all. The gaps aren’t a flaw in execution. They’re built into the structure. The company has no single pipeline that sees every input. So things fall through, constantly, and everyone treats it as normal.

The core network would never tolerate this. And it doesn’t have to be tolerated in a business either — it’s an architecture choice, not a law of nature.

There’s a second idea from that world I think about constantly: the observation hierarchy. In a core network, raw events at the bottom get processed, layer by layer, into higher and higher meaning — until an operator sees a single status light that says “healthy” or “intervene.” When something turns red, they don’t drown in raw data; they drill — from the alarm, to the subsystem, to the exact component, to the precise failed event. Forty alarms resolve to one root cause. The intelligence to do that lives in the architecture, not in the operator’s memory.

A company has every one of those layers — a board, management, departments, the daily work, the raw data underneath — and almost none of the processing between them. Information moves up by being manually compiled and filtered, losing fidelity and time at every step, until leadership is reading a blurred, weeks-old picture and calling it a report. There’s no drill-down. When something’s wrong, no one can trace it in three clicks from “revenue is soft” to the exact stalled deal and the exact missed follow-up and the person who owns it.

That’s not because it’s impossible. It’s because no one built the layers.

The deepest lesson, though, is about failure. In a mission-critical system, when something stalls, the architecture doesn’t wait for a human to notice. It reminds, it reroutes to a backup, and if no one acts, it executes the fallback itself — and logs exactly what it caught. The system has volition about its own integrity. It does not quietly let things break and hope someone files a ticket.

Most companies let things break and hope someone files a ticket.

None of this is a metaphor. The architecture that keeps a call from dropping — universal capture, layered intelligence, drill-down to root cause, automatic recovery — is precisely the architecture an enterprise needs and almost never has. Nous is that architecture, transplanted from the network core to the business. Not inspired by it. Built from it.

The phone call doesn’t drop because someone decided, decades ago, that dropping it was unacceptable and engineered a system that wouldn’t. The same decision is available to any company willing to treat its own intelligence as mission-critical.