African Labour Intelligence

Evidence infrastructure for work that systems cannot see

Africa's labour markets are not under-skilled. They are under-measured. AYLI converts real job-site activity into trusted, auditable Provisional Economic Actor Records — so informal and semi-formal work can finally be seen by the financial and institutional systems that could serve it.

View sample EAR Explore the architecture
EAR — evidence, not a certificate Observe first Provisional governance
Core Positioning
We are building the missing evidence infrastructure that lets informal workers and micro-enterprises become visible to the economic systems that could otherwise serve them — lenders, contractors, and development programmes that today have no reliable way to see them.

This is foundational infrastructure underneath formal markets, not another consumer app on top of them.

What we build

AYLI — a multimodal evidence engine for informal and semi-formal work

Ndau Africa builds AYLI, a system that ingests ordinary job-site media and contextual inputs, translating them into structured, verifiable Economic Actor Records (EARs). An EAR documents what was observed on site, what trade activities that supports, which capability hypotheses are reasonable, how confident the protocol is, and where human review is required.

Is

AYLI is

  • Institutional evidence infrastructure
  • Human-reviewable by design
  • Provisional and auditable
  • Grounded in visible job-site media
  • Built for real, messy operating environments
Is not

AYLI is not

  • A traditional certificate issuer
  • A subjective worker ranking app
  • An unverified gig job board
  • An automated underwriting score
Institutional Access

What an Economic Actor Record makes possible

AYLI establishes the trust layer required to bridge informal trade capability with formal capital and institutional procurement. These are the core use cases our pilot infrastructure serves.

Lenders & Microfinance

Verify work history before a first loan

Review empirical, evidence-backed activity for operators with no formal credit bureau file, bypassing the need for self-reported paper resumes.

Contractors & Enterprises

Vet informal labour with objective proof

Inspect verifiable job-site performance history for subcontractors and teams, relying on auditable visual records rather than unverified word-of-mouth.

Programmes & Development Finance

Measure true economic activity

Ground economic development initiatives, grants, and impact reporting in rigorous multimodal work evidence instead of survey estimates.

Execution Sequence

Four strict stages — nothing skips ahead to a skill claim

01

Work submissions

Informed consent, raw media capture, activity context, and optional client confirmation.

02

Observation Engine

Extracts only visible actions, tools, materials, techniques, and evidence quality.

03

Capability reasoning

Conservative hypotheses with confidence metrics, evidence checklists, and rigorous checks.

04

Economic Actor Record

A living, reviewable evidence record engineered for institutional consumption.

Current Status

A live pipeline running on Google Cloud & Gemini

AYLI actively processes real-world job-site video into structured EARs—integrating multimodal observation, capability mapping, client validation, and automated quality control end to end.

Gemini multimodal Observation before inference Adversarial audit pass Client confirmation Tested on core trades Provisional governance
Open sample EAR Inspect JSON Specimen
record_typeeconomic_actor_record
capability_tierprovisional
evidence_anchoredtrue
unsupported_claim_checkpass
human_reviewrequired_if_flagged
not_a_credit_scoretrue
Methodology

Trust and verification over volume

01

Evidence-anchored observation

The model must explicitly cite visible actions and equipment before forming any hypothesis.

02

Unsupported-claim detection

An independent adversarial pass challenges statements that exceed the captured media.

03

Human-in-the-loop

HITL reviews escalate weak evidence. AI accelerates review without replacing human expertise.

04

Data governance

Consent at capture. Pilot records remain provisional, guarded by strict institutional safeguards.

Leadership & Origin

The Founder Story

Arnold Muza

Arnold Muza

Founder & CEO — Ndau Africa

"The people literally building Africa's infrastructure have been left without a recognized Ndau—without a rightful place, standing, or belonging in the formal economy."

Arnold Muza was raised by his grandparents in Chipinge, Zimbabwe, while his mother worked in South Africa to support their family. In his native roots, the word Ndau carries a profound meaning: “our place” or “our land.”

Leaving formal school following Form 4, Arnold arrived in Johannesburg with no traditional qualifications and forged his own path—teaching himself hotel management and running an establishment by the age of 22. Today, he operates a home renovations business, working daily beside master welders, electricians, and builders whose technical capability is extraordinary, yet entirely invisible to formal economic systems.

Ndau Africa was born from that lived reality. Arnold built AYLI's first end-to-end multimodal extraction pipeline himself on Google Cloud and Vertex AI to restore economic dignity—creating the missing trust infrastructure that grants Africa’s informal workforce their rightful Ndau.

Architectural Rigor

AYLI isn't a theoretical concept; it is actively engineered and stress-tested in production environments:

  • Engineered natively on Google Cloud Platform and Vertex AI.
  • Utilizes advanced Gemini multimodal models for frame-by-frame trade analysis.
  • Built by a founder who combines frontline construction expertise with hands-on systems architecture.
  • Designed from day one for compliance, POPIA alignment, and zero-compromise data governance.
Institutional Access

Partner with Ndau Africa

Partnerships

Lenders, microfinance institutions, general contractors, and development funds.

partners@ndauafrica.com →

Research & ML

Applied AI researchers, computer vision specialists, and data governance experts.

research@ndauafrica.com →

General Inquiries

Press, media inquiries, and general correspondence.

hello@ndauafrica.com →