Post 1: Building Prospergenics – Understanding What Communities Actually Need
Nederlands:
Wanneer Goedbedoelde Hulp Voorbij Gaat aan Echte Behoeften
Er zijn duizenden development programma’s in Oost-Afrika. Velen falen. Niet door gebrek aan budget of goede intenties, maar omdat ze oplossingen bouwen voor problemen die ze denken te kennen, zonder te luisteren naar wat communities werkelijk nodig hebben.
Bij Prospergenics begonnen we anders: niet met een plan, maar met vragen. Niet met aannames, maar met luisteren. En net als bij Art Revisionist, waar we AI gebruikten om kunsthistorische misinformatie te corrigeren, gebruikten we AI om iets veel fundamentelers te begrijpen: wat hebben mensen echt nodig om te groeien?
Het Probleem met “We Weten Wat Ze Nodig Hebben”
Traditionele development programma’s werken vaak top-down:
– Westerse organisaties besluiten wat “goed” is
– Ze implementeren voorgedefinieerde programma’s
– Ze meten succes aan hun eigen metrics
– Communities worden “ontvangers” in plaats van partners
Het resultaat? Laptops die ongebruikt blijven omdat er geen elektriciteit is. Training in skills die de lokale markt niet vraagt. Microkredieten met voorwaarden die de realiteit van kleine ondernemers niet begrijpen.
We realiseerden ons: voordat je een platform bouwt, moet je begrijpen wat je eigenlijk aan het bouwen bent.
Van Aannames naar Data
We begonnen met systematisch verzamelen van informatie, vergelijkbaar met hoe we bij Art Revisionist primaire bronnen verzamelden:
Community Conversations (Primaire Bronnen)
– 47 interviews met jonge ondernemers in Kenia
– 23 gesprekken met lokale coaches en mentors
– 12 focus groups in verschillende regio’s
– Directe observatie van bestaande training programma’s
Existing Programs Analysis (Secundaire Bronnen)
– Evaluaties van 30+ development programma’s
– Success/failure rates van verschillende interventies
– Marktonderzoek lokale economieën
– Skill gap analyses arbeidsmarkt Oost-Afrika
Conflicting Information (Wat We Ontdekten)
– Development rapporten zeggen X
– Community members ervaren Y
– Onze aannames waren Z
– De werkelijkheid bleek iets heel anders
Wat We Hoorden (En Wat Het Betekende)
“We willen geen handouts, we willen skills die ons geld opleveren”
→ Betekenis: Training moet direct verdienpotentieel hebben, niet algemene educatie
“Ik kan niet 3 maanden fulltime naar school – ik moet werken”
→ Betekenis: Programma moet part-time, flexibel, en inkomsten-genererend tijdens leren
“AI en programmeren lijken interessant, maar is er werk voor?”
→ Betekenis: Skills moeten lokale + internationale markten bedienen (remote work mogelijk)
“Microkredieten vragen collateral die we niet hebben”
→ Betekenis: Alternatieve financiering gebaseerd op skills en potentieel, niet bezit
“Community support motiveert ons meer dan lone entrepreneurship”
→ Betekening: Collaborative learning model, niet individueel competitief
Het Structureren van Communities’ Stemmen
Net als bij Art Revisionist, waar we informatie systematisch organiseerden, creëerden we een gestructureerde kennisbank:
Needs Taxonomy:
– Immediate needs (food, shelter, basic income) → Can’t train hungry people
– Short-term needs (quick income, 1-3 months) → Fast skills with immediate earning
– Medium-term needs (career building, 6-12 months) → Professional development
– Long-term needs (wealth building, 1-5 years) → Entrepreneurship, asset building
Skills Mapping:
– What skills are teachable online? (Programming, design, AI tools, marketing)
– What skills have local demand? (Solar tech, agriculture tech, mobile money)
– What skills enable remote work? (Development, content creation, virtual assistance)
– What skills compound? (Digital literacy → specific tools → entrepreneurship)
Barrier Documentation:
– Financial barriers (training costs, equipment, connectivity)
– Time barriers (work schedules, family obligations)
– Knowledge barriers (language, prerequisite skills, confidence)
– Social barriers (gender expectations, age discrimination, geographic isolation)
De Doorbraak Insights
Ergens in maand 3 van research vielen de puzzelstukken op hun plaats:
Insight 1: Education + Income Must Be Simultaneous
Je kunt mensen niet vragen te stoppen met werken om te leren. Het programma zelf moet inkomsten genereren vanaf dag 1.
Insight 2: AI Democratiseert Skills
Met AI tools kan iemand met basiskennis professioneel werk leveren. Een beginnend designer met AI image generation kan betaalde projecten doen. Een junior developer met GitHub Copilot kan productieve code schrijven.
Insight 3: Community > Competition
Westerse entrepreneurship is vaak lone-wolf model. Oost-Afrikaanse cultuur is inherent collaboratief. Het programma moet dat versterken, niet tegengaan.
Insight 4: Microcredit Should Follow Skills, Not Precede Them
Traditioneel: eerst lenen, dan proberen business te starten
Beter: eerst skills opbouwen, dan microcredit voor scaling
Insight 5: Remote Work = Geographic Arbitrage
Iemand in Kenia die voor Nederlandse klanten werkt verdient 5-10x lokaal loon. Maar alleen als ze skills hebben die internationale markt vraagt.
Van Luisteren naar Bouwen
Op basis van deze research tekenden we de contouren van Prospergenics:
Core Principle: Earn while you learn through AI-augmented skills
Model: Part-time cohort learning with peer support
Skills Focus: Digital tech (programming, AI tools, design, content)
Revenue Model: Student earnings + microcredit + community wealth building
Success Metric: Not “graduates” but “sustained income increase”
Maar informatie verzamelen is pas stap 1. De volgende uitdaging: hoe maak je deze kennis actionable? Hoe bouw je een AI-systeem dat niet alleen weet wat communities nodig hebben, maar ook kan helpen dat te leveren?
Dat is waar de AI kennisbank om de hoek kwam kijken. Maar daarover meer in de volgende post.
Lees meer over ons model: Prospergenics
Zie ook hoe we vergelijkbare research methodologie gebruikten: Art Revisionist
—
English:
When Well-Intentioned Aid Misses Actual Needs
There are thousands of development programs in East Africa. Many fail. Not from lack of budget or good intentions, but because they build solutions for problems they think they know, without listening to what communities actually need.
At Prospergenics we started differently: not with a plan, but with questions. Not with assumptions, but with listening. And just like at Art Revisionist, where we used AI to correct art historical misinformation, we used AI to understand something much more fundamental: what do people really need to thrive?
The Problem with “We Know What They Need”
Traditional development programs often work top-down:
– Western organizations decide what’s “good”
– They implement predefined programs
– They measure success by their own metrics
– Communities become “recipients” instead of partners
The result? Laptops that remain unused because there’s no electricity. Training in skills the local market doesn’t demand. Microcredits with conditions that don’t understand small entrepreneurs’ reality.
We realized: before you build a platform, you must understand what you’re actually building.
From Assumptions to Data
We started by systematically gathering information, similar to how we collected primary sources at Art Revisionist:
Community Conversations (Primary Sources)
– 47 interviews with young entrepreneurs in Kenya
– 23 conversations with local coaches and mentors
– 12 focus groups in different regions
– Direct observation of existing training programs
Existing Programs Analysis (Secondary Sources)
– Evaluations of 30+ development programs
– Success/failure rates of different interventions
– Local economy market research
– Skill gap analyses East African labor market
Conflicting Information (What We Discovered)
– Development reports say X
– Community members experience Y
– Our assumptions were Z
– Reality turned out to be something else entirely
What We Heard (And What It Meant)
“We don’t want handouts, we want skills that earn us money”
→ Meaning: Training must have direct earning potential, not general education
“I can’t go to school fulltime for 3 months – I need to work”
→ Meaning: Program must be part-time, flexible, and income-generating during learning
“AI and programming sound interesting, but is there work for it?”
→ Meaning: Skills must serve local + international markets (remote work possible)
“Microcredits ask for collateral we don’t have”
→ Meaning: Alternative financing based on skills and potential, not possessions
“Community support motivates us more than lone entrepreneurship”
→ Meaning: Collaborative learning model, not individually competitive
Structuring Communities’ Voices
Just like at Art Revisionist, where we systematically organized information, we created a structured knowledge base:
Needs Taxonomy:
– Immediate needs (food, shelter, basic income) → Can’t train hungry people
– Short-term needs (quick income, 1-3 months) → Fast skills with immediate earning
– Medium-term needs (career building, 6-12 months) → Professional development
– Long-term needs (wealth building, 1-5 years) → Entrepreneurship, asset building
Skills Mapping:
– What skills are teachable online? (Programming, design, AI tools, marketing)
– What skills have local demand? (Solar tech, agriculture tech, mobile money)
– What skills enable remote work? (Development, content creation, virtual assistance)
– What skills compound? (Digital literacy → specific tools → entrepreneurship)
Barrier Documentation:
– Financial barriers (training costs, equipment, connectivity)
– Time barriers (work schedules, family obligations)
– Knowledge barriers (language, prerequisite skills, confidence)
– Social barriers (gender expectations, age discrimination, geographic isolation)
The Breakthrough Insights
Somewhere in month 3 of research, the puzzle pieces fell into place:
Insight 1: Education + Income Must Be Simultaneous
You can’t ask people to stop working to learn. The program itself must generate income from day 1.
Insight 2: AI Democratizes Skills
With AI tools, someone with basic knowledge can deliver professional work. A beginning designer with AI image generation can do paid projects. A junior developer with GitHub Copilot can write productive code.
Insight 3: Community > Competition
Western entrepreneurship is often lone-wolf model. East African culture is inherently collaborative. The program must strengthen that, not oppose it.
Insight 4: Microcredit Should Follow Skills, Not Precede Them
Traditional: first borrow, then try to start business
Better: first build skills, then microcredit for scaling
Insight 5: Remote Work = Geographic Arbitrage
Someone in Kenya working for Dutch clients earns 5-10x local wages. But only if they have skills the international market demands.
From Listening to Building
Based on this research, we outlined Prospergenics:
Core Principle: Earn while you learn through AI-augmented skills
Model: Part-time cohort learning with peer support
Skills Focus: Digital tech (programming, AI tools, design, content)
Revenue Model: Student earnings + microcredit + community wealth building
Success Metric: Not “graduates” but “sustained income increase”
But gathering information is only step 1. The next challenge: how do you make this knowledge actionable? How do you build an AI system that not only knows what communities need, but can help deliver it?
That’s where the AI knowledge base came around the corner. But more on that in the next post.
Read more about our model: Prospergenics
See also how we used similar research methodology: Art Revisionist
—
Publicatiedatum: [DAY 1]
Tags: Community Development, AI Education, East Africa, Entrepreneurship, Prospergenics
Categorie: Impact