Where Nigerians learn to ship AI, and get paid for it.

Decisions already made

The other two streams

Ninety-day target for this stream: two sprints signed or one sprint plus a pilot in negotiation. Not ten. The honest year-one membership picture, if Thursday works, is about 80 paying Builders by month three and 250 Builders plus 20 Studios by month six. Do not model ten thousand members. Model four hundred people who ship.

1. PHASE 0

   Lock the shape

   Days 1–7

   A sentence, a room, a list of thirty humans, a person who owns Thursday.

   • Homepage sentence goes live. About page keeps the longer pan-African line.
   • Four roles named on one page. Community lead is a name, not a hope.
   • WhatsApp Community open, rules pinned.
   • Thirty voice notes sent. Use the script in This week.
   • Paystack plans created, not announced.
   • Builder-in-residence starts the thinnest version of one founder ship.

2. PHASE 1

   First cohort, first ships

   Weeks 2–4

   Forty seats. Thursday becomes a habit. Three founder ships set the bar.

   • Open Builder. First four weeks free for whoever attends night one, then ₦4,500. Cap 40.
   • Week 2 is online only. Week 3 is Kaduna in person if the venue has said yes. Week 4 is Abuja only if fifteen names are real.
   • Ten prompts published on OpenLibrary with the taxonomy filled in.
   • Malam banners and trial codes go on the two flagship courses.
   • By week 4 you want 20 people in the room. The floor is 12. Under 12, do not open another city.

3. PHASE 2

   Public proof

   Weeks 5–8

   A demo day a buyer could sit through without being embarrassed.

   • Cohort-one demo day. Main room in Kaduna or Abuja. Lagos as a watch party if the names are real. Livestream, then a short cut for Arewa or Dole Ajimu.
   • Eight to twelve ships, each with the case-study fields. If you have fewer than six, delay the marketplace.
   • Open the curated marketplace with those ships. Manual payouts are fine.
   • Five buyer conversations already on the calendar. Two written proposals out. One verbal yes is a win.
   • Kano night, co-branded with Arewa, if Kaduna attendance held.

4. PHASE 3

   Money that is not the founder

   Weeks 9–12

   Members pay, or a buyer pays. Preferably both. Publish the proof either way.

   • Convert cohort one. Expect 25–40% to pay. Under 15%, the night is entertainment. Change the format before you tell anyone to “engage.”
   • Run one paid Team Sprint, even at the ₦600,000 reference price, or take the kickoff on a tool pilot.
   • Publish a 90-day proof page: members, ships, naira members earned, languages, cities. Small numbers are allowed. Invented numbers are not.
   • Hackathon one: 48 hours, theme from the reference buyer, ₦200,000 plus a paid extension if the tool is real.
   • Diaspora Circle opens to twenty people you can actually call. Their fees fund student seats. Say the ratio you can truly cover.

Rooms, power, and the network

• One physical city a month. Kaduna first, because the team, Kadroot, and Dole Ajimu are there. Then Kano with Arewa. Then Abuja or Lagos, whichever has fifteen warm names. Do not fly to an empty room.
• Everyone else is on the same Meet. Notes are text, the same night. Video is optional and small.
• If the network or the generator fails, the night continues in WhatsApp. It still ends with: what will be tryable next Thursday.

Posts you can send

1. 01Hausa market price note

   Pilot ₦400,000–₦1.2m, or a prompt pack at ₦5,000

   Agri · Kano or Kaduna · Hausa · last week’s sheet works offline

   A farmer names a crop and a market. The tool returns a short price range and who to call, as text or a voice note. Data is a weekly sheet from five Arewa members, not a scraped fantasy. WhatsApp bot or a tiny form.

   First buyer. A trader association or an agri extension NGO.

   Watch. If the sheet is not updated, say so. A confident wrong price is worse than silence.

2. 02Kadroot nursery assistant

   Free to those sellers. The case study is the payment. A Studio member later charges other markets ₦80,000 to set the same thing up.

   Commerce · Kaduna · English and Hausa · online, last stock cached

   A seller says what they have in plain language. The tool drafts a Kadroot listing and a customer reply using real categories from the nursery marketplace.

   First buyer. Kadroot sellers. The first users are already in the portfolio.

   Watch. Do not invent stock. If it is not in the seller’s message, it is not in the listing.

3. 03JSS basic-science quiz

   School setup ₦150,000–₦400,000. Sprint if they want the staff trained.

   Education · Kaduna · English with a Hausa glossary · printable, so offline after export

   A teacher names a topic from the scheme of work. The tool returns ten questions, answers, and a short Hausa gloss of terms students mix up. Answer-key habit comes from Postcode Academy, not from a new LMS.

   First buyer. One secondary school, then a state basic-education desk if the first teacher will say it worked.

   Watch. A teacher has to mark a sample before you show it to a principal. Wrong science is a reputation event.

4. 04Pidgin customer desk

   ₦80,000–₦150,000 per shop.

   Commerce · Lagos or Kaduna · Pidgin and plain English · text, low bandwidth

   A shop owner forwards a WhatsApp complaint. The tool drafts a calm reply and flags refund versus explain. It does not send the reply on its own.

   First buyer. The shop. Studio members resell the setup.

   Watch. Never auto-send. A bad refund promise is a real loss.

5. 05Postcode explainer for SMEs

   Leads for the Academy. A training sprint if an agency wants staff taught.

   Commerce · Lagos, Kano, Kaduna · plain English · text

   An owner describes a street. The tool explains how to get and use a digital postcode using Postcode Academy’s actual lessons, then links to post.malam.app.

   First buyer. SMEs who already ask you this, and any postcode or state ICT contact you truly have.

   Watch. Do not invent addressing rules. If the curriculum does not say it, the tool does not say it.

6. 06Arewa heritage voice card

   Project fee. The teller is paid for the session even if the card is never used.

   Culture · Kano or Kaduna · Hausa, English summary · audio stays with the teller

   A 60-second story becomes a title, a short English summary for diaspora, and a card for Arewa and Dole Ajimu. Nothing posts until the teller approves the summary.

   First buyer. Arewa, a cultural NGO, or diaspora family associations.

   Watch. This is the public ethic for every later dataset. Approval before publish. Pay before record.

7. 07Clinic reminder

   ₦300,000–₦800,000 setup, or fold it into a sprint.

   Health · one clinic · Hausa and English · the register never goes into a public prompt

   Name, date, language. The tool drafts an SMS or a voice-note reminder. No diagnosis. No symptoms. The template runs on data the clinic holds.

   First buyer. A small clinic, an NGO, or an LGA health desk.

   Watch. If patient data would have to be pasted into a public model, you do not ship it. Say that in the case study.

8. 08Ward letter drafter

   ₦200,000–₦600,000, or a sprint if they want staff to draft the next ten themselves.

   Governance · Kaduna or Kano · English letter, Hausa summary · text

   A community leader describes a request — a borehole, a classroom. The tool drafts a formal letter and a Hausa summary for the ward meeting, in a format a local champion already uses.

   First buyer. A councillor’s office, a community association, or a governance NGO.

   Watch. Do not claim it was submitted. Do not add accusations the speaker did not make.