Recivo
Live build · LagosVol. 01 · May 2026

Skill is a verb.
Prove it.

Recivo is WAEC for the informal economy. A voice-and-simulation skill exam that turns a trader’s hands into a verifiable credential, and routes payment through Squad rails so doing the work on-platform makes you more hireable the next time.

Field-validated with 24 workers and 6 employers across Computer Village, Ladipo, Tejuosho, Aba.

Tunde Adebayo, phone repair technician, Computer Village
Recivo · Verified#RC-9012-3456
Tunde Adebayo
Phone repair · Computer Village, Ikeja · 12 yrs
Knowledge
76/100
Trust
71/100
Iṣẹ́
73/100
24 Squad-confirmed jobs58% repeat customers
Diagnostic replay
06:42
41M
Informal workers in Nigeria
NBS, 2023
≈100%
With no portable credential
1 in 4
Off-platform jobs that ghost monthly
field interviews, n=20
<60s
Squad rails settlement
Transfer API
01 — The problem
Forty-one millioninvisibles.

Nigeria’s informal sector employs more workers than the entire formal workforce. They are not poor by lack of ability — they are poor by lack of proof.

01

Skill is invisible.

A phone technician with twelve years of experience has no certificate. No employer outside his WhatsApp circle has any way to know.

Structural failure
02

Trust is local.

Reputation lives in physical neighborhoods and word-of-mouth. A welder in Aba cannot accept a job in Lekki because nobody in Lekki has heard of him.

Structural failure
03

Payment is broken.

Clients ghost after work is done. Workers chase fees for weeks. Informal work pays 30% less than formal work — risk priced into every transaction.

Structural failure
02 — What’s been tried

Why nothing else verifies the work itself.

  • Jiji / OLX
    Classified listings
    No verification. Stars from strangers.
  • LinkedIn
    Self-reported résumés
    Built for paper-credentialed users.
  • Lynk (KE)
    Admin verification
    Verifies ID, not skill.
  • WhatsApp
    Where work happens
    Trapped inside groups. No payment safety.

If the customer can see I passed the test before hiring me, I don’t need to beg for the job. The test is begging for me.

Emeka · phone repair · Computer Village

02 — Diagnostic Interview

Not a skill exam.
A diagnostic interview.

A drag-and-drop cannot prove a worker can physically execute a repair — that takes muscle memory, smell, sound. What it canprove is that they can correctly diagnose, reason about, and articulate the fix. That’s exactly what an employer needs to know before hiring. Real execution is verified later, by Squad-confirmed completed jobs. We don’t over-claim.

Stage 010:00 – 0:30

Presentation

A 3D fault appears on screen. The voice agent — in Pidgin, Yorùbá, Igbo, Hausa or English — frames the customer brief and asks where the worker would begin.

Stage 020:30 – 3:00

Diagnosis dialogue

The worker reasons aloud. The agent probes — Why this tool first? How will you eliminate motherboard? Branches across a 30-node decision tree per fault.

Stage 033:00 – 5:00

Interactive demo

Drag the battery out. Probe the multimeter onto the contacts. Re-seat the ribbon. Every action emits a typed event that the deterministic scorer grades against the correct-path graph.

Stage 045:00 – 6:00

Resolution

The worker explains the full fix to the customer in their own words. The reasoning model scores articulation — the heaviest-weighted dimension, because it&rsquo;s what employers actually evaluate.

Sample exchange — Tecno Spark, no power
Recording · language: Pidgin
T-00:24
AI
Customer bring this phone come. E say e drop am, now e no dey come on at all. Battery dey, charger dey. Wetin you go check first?
TU
I go open the back first, check if the battery still dey sit well. If the drop knock am loose, na quick fix.
AI
Good. How you go know say na battery problem, no be motherboard?
TU
I go put multimeter on the contact. If e show three-point-seven to four-point-two, battery alright — I move to display ribbon and the charging port.
Scorer note · Reasoning: isolates cheapest variable first; cites correct voltage band. +6
Rubric — phone repair
  • Diagnostic reasoning
    Isolate causes systematically
    30%
  • Procedural accuracy
    Right tool, right sequence
    25%
  • Articulation
    Can explain what & why
    20%
  • Safety awareness
    Discharge, isolation, data
    15%
  • Edge-case handling
    Respond to curveballs
    10%
Every score includes a text justification the worker can dispute. Dual-model scoring with auto-flag on Δ > 8 points. Weekly 5% human audit against master-technician reference.
03 — The Iṣẹ́ Score

Two halves, compounding.

Iṣẹ́ (Yorùbá: work, craft) — pronounced ee-shay. Knowledge is what you know. Trust is what you’ve done. The weighting moves from one to the other as you compound on-platform.

Composite formula
Iṣẹ́ = α · K + (1 − α) · T
α = max(0.30, 1 − jobs / 20)

A first-day worker is judged entirely on Knowledge. After 10 jobs, half-and-half. After 14+, Trust dominates — capped at 70% so skill never becomes irrelevant.

Weighting curve

How α decays as you compound.

Knowledge (α)Trust (1−α)
0%50%100%05101520COMPLETED JOBSα floor = 0.30

Knowledge — fixed for six months

From the Diagnostic Interview. Voluntary retake at ₦500 — empowering, not punitive. Workers come back when they think they’ve improved.

Trust — compounds with every Squad-confirmed job

Inputs: completed jobs (log curve), star ratings, structured feedback tags, repeat-customer ratio, dispute rate, time-decay.

Going off-platform passively outranks you

We don’t punish workers who transact off-platform. Their Trust just stops compounding while everyone else’s climbs.

Veterans can&rsquo;t coast on past glory

Trust decays over time. A great mechanic who hasn’t worked in a year drops in rank. Knowledge keeps the bar real.

04 — Squad APIs · structural, not decorative

Remove Squad,
the product disappears.

Trust Score literally cannot exist without Squad-confirmed transactions. Per-job Virtual Account escrow is a financial primitive that turns “matching” into trust infrastructure — three APIs doing core platform work, not decorative payment collection.

Per-job VAs

Each accepted job spawns a job-scoped Virtual Account. Employer funds escrow on hire.

Solves “client ghosted me”. Reframes Recivo as trust infrastructure.

Virtual Accounts

Provisioned at signup — a GTCO-backed VA tied to the worker’s number.

The spine of the Trust Score. Every payment in = verified job event.

Transfer API

On confirm, funds release from job escrow to the worker’s primary VA.

Sub-60-second payout. Worker sees money instantly — beats WhatsApp + bank.

Lifecycle of a single job

Funded → released → score updated.

  1. T+0s
    Employer funds job VA
  2. T+3s
    Worker notified, funds locked
  3. T+confirm
    Employer taps release
  4. <60s
    Transfer API → worker VA
Recivo fee: 3% on release. Worker takes home 97%, same day.
₦12,000 → ₦11,640 in 47s
05 — Who we built for

Three faces. One stuck market.

Recivo is designed against named, interviewed humans — not personas we invented. Field validation across Computer Village, Ladipo Auto Market, Tejuosho, and four Lekki fashion houses.

Tunde, 38, Phone repair technician
Phone repair technician
Tunde, 38
Iyana-Ipaja → Computer Village, Ikeja

“I’ve been fixing phones for twelve years. Everybody in this market knows me. But if you ask me to go work for Slot or 3CHub, they don’t know me, so they won’t hire me. I’m stuck here.”

Income
₦80k–₦220k / month · volatile
Phone
Tecno Spark 10 · 4GB RAM
Languages
Yorùbá, Pidgin, basic English
Pain point
Lost ₦100,000 last month settling a false damage claim — no intake record
Adaeze, 29, Fashion entrepreneur
Fashion entrepreneur
Adaeze, 29
Lekki, manufactures in Aba

“I’ve hired three tailors in eighteen months who claimed to do French seams. Two couldn’t. ₦400,000 in spoiled fabric. I’d pay a 15% premium for someone I can trust.”

Hires/yr
≈18 tailors across orders
WTP
+15% for verified skill
Channels
WhatsApp groups · Jiji · referrals
Pain point
No way to test claimed specialty before commissioning
Mr Bello, 47, HR officer, logistics SME
HR officer, logistics SME
Mr Bello, 47
Apapa · 40-truck fleet

“Mechanic interviews are theatre. Candidates memorise answers. A bad mechanic hire costs us ₦600,000 in repair errors over six months. I need to watch them reason through a real fault.”

Fleet
40 trucks · mixed diesel
Cost of bad hire
≈₦600k / 6 months
Current method
Trade-test on owned truck
Pain point
Theatrical interviews; no fair assessment
06 — Scale path

Three trades to fifteen.
Six weeks at a time.

Adding a trade is a content problem, not an engineering one. Once the simulation engine, voice agent, and rubric framework are in place, a new trade needs 3D assets, master-craftsman rubrics, and a decision tree — six weeks, start to launch.

Phase 1
Months 1–6
3trades
Computer Village + Tejuosho
Verified workers18,000
Phase 2
Months 6–12
6trades
+ Aba, Kano
Verified workers50,000
Phase 3
Months 12–24
15trades
Every state capital
Verified workers250,000
Squad · GTCOComputer Village Phone Dealers AssociationFADAN · Fashion Designers Association of NigeriaNAMA · Auto Mechanics AssociationFG Innovation Enterprise InstitutionsLagos State Skills Acquisition CentreNaijaVoices Corpus · Common VoiceSquad · GTCOComputer Village Phone Dealers AssociationFADAN · Fashion Designers Association of NigeriaNAMA · Auto Mechanics AssociationFG Innovation Enterprise InstitutionsLagos State Skills Acquisition CentreNaijaVoices Corpus · Common Voice
The bet, restated

Skill is a verb.
We help them prove it.

Recivo is Squad’s path into 41 million Nigerian wallets that no bank has ever touched. The diagnostic is the moat. The Trust Score is the credential. The rails are the proof.