Step into Labs →

Systems that
run | the desk

We build trading, settlement and data systems for energy and finance teams in Africa, and the AI agents that keep them running.

Lusaka, Zambia · Energy first, built to extend

Built, shipped and working

Watch our systems run, or pick one from the dock.

Trading deskFileViewWindowMon 09:41
Trading desk · P&L and risk
Trading desk P&L and risk screen with illustrative figures

What each system does, and how

Three products in use or in pilot today, and the agent that keeps their data honest underneath all of them.

Live on a Southern African trading desk

Trading desk

One system for a power-trading desk, from booking a trade to getting paid, with profit and risk visible as it moves.

  1. 01Book
  2. 02Settle
  3. 03See

Closed pilot · ZICTA Top 100 Innovations 2026

KaSomeFin

Saving that happens without deciding to save: every mobile-money payment is rounded up and the change set aside on its own.

  1. 01Pay
  2. 02Round up
  3. 03Grow

Live platform · Energy management

EMS Zambia

Energy management for commercial and industrial sites, saving clients ZMW 550K–2M a year with payback in 2–4 months.

  1. 01Measure
  2. 02Validate
  3. 03Act

Runs under every project · Live in 2026

Fidèle

The data-integrity agent underneath everything we build. Garbage in, garbage out, so Fidèle stops the garbage.

  1. 01CheckEvery table, every sweep: structure, duplicates, links, freshness, ranges, units.
  2. 02ExplainEach problem is traced to the row and the value, with why it matters.
  3. 03Approve and documentThe table's owner approves the fix; every change is written to the log.

How we work

We only take on problems that pass all four tests of the 4U filter, then take them from brief to shipped system in five stages.

Unworkable

The existing solution has broken down and is no longer fit for purpose.

Unavoidable

The pressure compounds the longer it's ignored. It can't be deferred.

Urgent

Time is the constraint, not capital. Speed of delivery is the value.

Underserved

Overlooked by the global market, because the context is African.

  1. 1

    Intake

    The brief goes through the 4U filter. Only qualifying problems proceed.

  2. 2

    Research

    Context analysed, constraints mapped, local realities respected.

  3. 3

    Prototype

    A first working version that fails fast and learns faster.

  4. 4

    Validate

    Real-world testing under the conditions it will actually run in.

  5. 5

    Ship

    Deployed, documented and maintained. A system worth keeping.

A problem worth
building around?

Bring the unworkable, the unavoidable, the urgent, and the underserved. We'll bring the engineering.

info@charismaticcodes.com · Lusaka, Zambia