Building a Sovereign Trust Network, in Public
Why we are building Zambia's national rails for identity, payments and compliance — and how an AI build crew ships it with the brakes on.
Why this exists
Most digital infrastructure in our region is rented from somewhere else. Payments clear through rails we don't own, identity lives in silos that don't talk to each other, and the regulator is always the last to find out what happened. Unityzm is the attempt to invert that: one national network where identity, payments, invoicing, commerce and compliance share a single hash-chained ledger, no passwords are ever transmitted, and the regulator is a first-class participant rather than an afterthought.
This blog is where I write that build story down as it happens — the decisions, the dead ends, and the research behind them.
The shape of the thing
The platform is organised as a set of cooperating layers:
- Identity — sovereign sign-on with no password ever sent over the wire.
- Wallets & payments — settlement designed for African conditions, online and offline, with manual cross-pay deposits for the realities of mobile money.
- Commerce & logistics — a storefront, escrow, and a delivery network that settles end-to-end over the same rails.
- Compliance — forensic-ready audit on every move, with the regulator wired in from day one.
Each layer is real code with real tests, not a slide. Where something is still scaffolded, I say so.
The pace, in public
Progress is uneven on purpose — some weeks are research, some are a wall of shipping. Here is a recent stretch of weekly commits, the kind of honest signal this blog exists to show:
The W23 step is when the AI build crew came online — the rails that adjudicate what the agents may do.
That W23 step is not magic; it is the moment the governance was solid enough to let the build crew run. The chart above is rendered live from the post itself, not a screenshot.
Built by an AI crew, with the brakes on
A lot of this is built by specialist AI agents — a compliance officer, a tax specialist, a code architect, a reviewer — but autonomous action is adjudicated by rules, not vibes. Every move an agent wants to make passes through capability gates before it runs, and every action lands on the same audit ledger as everything else. The interesting engineering is not "the AI wrote code"; it's the governance that decides what the AI is allowed to do.
What you'll read here
The goal of this blog is the same as the goal of the platform: be legible.
Expect founder research, architecture notes, and honest progress reports. Some posts will be wins. Some will be "here is what broke and what we learned." That is the build story — and you're early.
Welcome.
Where to start — read the story in series:
- The Thesis (4 parts) — why we're building sovereign rails: the market gap, the first commit, one ledger, and no passwords.
- Building the Network (4 parts) — the rails themselves: USN addressing, sub-second settlement, fiscalized invoices, and the gateway.
- A Living Economy (4 parts) — the people on the rails: the citizen wallet, the social directory, trust by tier, and the agents that never sleep.
- Under the Hood (4 parts) — how it all works together: order-to-carrier, config-not-code, metered AI, and an honest fleet map.
- Trust, Metered (4 parts) — governance you can audit: the kill-switch, maker-checker, design-before-code, and an audit trail that can't lie.
- Delivery Efficiency (6 parts) — how one person shipped it with AI, with the receipts.