African Reality Is Not Messy. It Is Intelligence

Created by yours truly — OpenAI

Your Sunday school teacher may have skipped this part, but David was basically saying, “This tech stack does not fit my use case.”

As usual, I do not feel like a writer if I do not show you how what is written in the Bible still correlates with what is happening now.

So, allow me to take you to the man who wrote Psalms.

Not to talk about my favourite Psalm — “The Lord is my shepherd; I shall not want” — but to show you that David had his cup filled without wearing Saul’s armour.

Before David faced Goliath, he first had to survive Saul’s armour.

And this part of the story, my friends, is the section that will make you M-Pesa me your Sunday offering because your Sunday school teacher may not have told you that sometimes the best way to win is by using your own tools.

We often focus on the giant, the stone, the sling, and the victory. But before David stepped onto the battlefield, someone tried to dress him in a system that was not built from his experience.

The Bible says:

“And Saul armed David with his armour, and he put an helmet of brass upon his head; also he armed him with a coat of mail. And David girded his sword upon his armour, and he assayed to go; for he had not proved it. And David said unto Saul, I cannot go with these; for I have not proved them. And David put them off him.” — 1 Samuel 17:38–39, KJV

And the above is the word of God.

Thanks be to God 🙏

Lets pay attention to this like

“I cannot go with these; for I have not proved them.”

David was not being arrogant. He was being intelligent.

He understood that something can look powerful and still be unsuitable. Something can be respected by kings and still become a burden on the wrong body.

Saul’s armour was not useless. It was just untested in David’s reality. So David removed it.

He went back to what his environment had taught him.

1. A sling.
2. Stones.
3. Speed.
4. Precision.
5. Distance.
6. Observation.

The quiet intelligence of a shepherd who had learned danger in the field before he ever met danger on a battlefield.

David did not defeat Goliath by borrowing Saul’s system. He won with intelligence shaped by his own reality.

And maybe that is where Africa needs to begin its conversation about Data and AI.

When Borrowed Intelligence Becomes a Burden

For a long time, Africa has been handed Saul’s armour.

• Imported systems.
• Imported models.
• Imported assumptions.
• Imported definitions of intelligence.
• Imported ways of measuring trust, risk, education, productivity, identity, talent, and opportunity.

And when our reality does not fit those systems, we are told the problem is us.

• Our markets are too informal.
• Our data is too messy.
• Our addresses are not clear.
• Our languages are too many.
• Our systems are not mature enough.

But maybe we have been asking the wrong question. Maybe the question is not:

Why does Africa not fit imported systems?

Maybe the real question is:

Why do imported systems fail to understand African intelligence?

Because African reality is not empty. It is full of intelligence.

Informal Systems Are Intelligence

A market in Nairobi, Lagos, Accra, or Kampala may look chaotic to someone searching for supermarket aisles, barcodes, formal receipts, and clean inventory systems.

But sit there long enough and you will see order.

• You will see credit without banks.
• Reputation without rating platforms.
• Customer segmentation without CRM.
• Supply chain routing without dashboards.
• Price optimization without spreadsheets.
• Risk scoring without formal credit bureaus.

A mama mboga knows who can take tomatoes on credit and who cannot. She knows who pays on Friday, who delays after school fees week, who is struggling silently, and who has built enough trust to be helped today.

That is not guesswork. That is intelligence. It is relational intelligence. It is market intelligence. It is survival intelligence.

Kenya’s Jua Kali sector shows this clearly. In 2019, it was responsible for 762,100 of the 840,600 new jobs created in the country. That is not a small informal corner outside the real economy. That is the economy breathing through informal systems.

Inside it, social networks move information, opportunity, apprenticeship, trust, supply, and reputation. That is distributed intelligence. We just rarely name it that way.

The problem is that many imported systems do not know how to read it.

They look for structure in the wrong places.

And because they cannot see the structure, they call it disorder.

African Adaptation Is Intelligence

M-Pesa is one of the clearest proofs that African adaptation is not weakness.

It succeeded because it understood the environment. Not everyone had a bank account. But many people had mobile phones.

Not everyone trusted institutions. But many people trusted agents, networks, relationships, and daily transaction behaviour.

M-Pesa did not wait for African customers to first become legible to Western banking systems. It worked with the behaviour already present in the market. It became powerful because it was shaped by African reality, not because it ignored it.

The same logic applies to credit.

Traditional credit scoring often assumes that lack of formal banking history means high risk. But in many African contexts, financial trust lives elsewhere: mobile money movement, chama discipline, airtime patterns, daily transactions, family obligations, and community reputation.

That is why systems like M-Shwari and Tala became important. They looked at mobile money velocity and micro-behaviours instead of waiting for formal credit histories. They proved that informal cash flow is not invisible. It only becomes invisible when the system is trained to look elsewhere.

This is the danger of borrowed intelligence. It does not always fail because the technology is weak.

It fails because the worldview is incomplete.

Local Markets Contain System Logic

An open-air market may look messy from a distance. But every market has memory.

• It remembers who supplies good stock.
• It remembers who lies.
• It remembers who pays late.
• It remembers who should be trusted with credit.
• It remembers which days people buy more.
• It remembers when school fees affect purchasing power.
• It remembers when rain changes supply.
• It remembers when politics, fuel prices, holidays, and funerals shift demand.

That is system logic.

A boda boda rider navigating Nairobi traffic is not operating in chaos. He is processing speed, road behaviour, weather, police presence, customer urgency, fuel cost, risk, and instinct in real time.

That is intelligence.

A small trader managing stock without a dashboard is not backward. She is reading demand, seasonality, supplier reliability, cash flow, customer mood, and household pressure.

That is intelligence.

A family surviving on irregular income while balancing rent, food, school fees, transport, emergencies, and business capital is not financially illiterate. They are running a high-frequency risk model every day.

That is intelligence.

So why should our AI ignore this?

When Systems Cannot Read Our Reality

This is where AI becomes dangerous when it is disconnected from context.

Not because AI is evil. But because intelligence without context can become confident blindness.

A logistics platform built around formal addresses may struggle in places where people navigate by churches, kiosks, schools, gates, trees, and boda stages.

“Turn left at the Safaricom kiosk, call me when you reach the blue gate” is not bad data. It is a different spatial system.

That is why companies like OkHi and Copia had to build around African navigation realities instead of forcing African customers into imported address logic.

The same problem appears in education.

An AI system trained on Western examples may confidently tell African children that there are four seasons, while many West African contexts are shaped mainly by wet and dry seasons.

That is not a small mistake. It teaches children that intelligence comes from elsewhere, even when their own environment is telling them something different.

It appears in recruitment. A model may confuse access to elite institutions with actual talent.

It appears in facial recognition. A system may struggle to see African faces accurately.

It appears in policy. A framework copied from formal economies may collapse when placed inside economies where informality carries employment, trade, survival, and innovation.

This is Saul’s armour. It may be advanced. It may be expensive. It may be globally respected.

But if it has not been proved in our reality, it can weaken us.

Our Behaviours Should Shape Our AI

Africa does not need to reject global technology.

That would be foolish.

David did not reject weapons because weapons were bad. He rejected weapons that did not fit the battle he understood.

That is the maturity we need.

We can learn from the world without disappearing inside the world’s assumptions.

We can use global AI without surrendering African context.

Because systems reveal environmental pressure.

Silicon Valley intelligence evolved under certain assumptions: stable electricity, high-speed internet, credit cards, formal addresses, individualism, abundant compute, and high institutional trust.

African intelligence evolved under different pressures: unreliable infrastructure, multilingual communities, informal work, daily income cycles, low institutional trust, strong social networks, and constant adaptation.

Those pressures did not produce absence. They produced intelligence.

So African AI should understand daily cash cycles, not only monthly salaries.

It should understand landmarks, not only postal codes.

It should understand local languages, Sheng, pidgin, mother tongue, code-switching, silence, tone, and context.

It should understand community trust, not only individual identity.

It should understand that a person is often not an isolated user, but a node in a family, a chama, a SACCO, a church, a WhatsApp group, a market, and a community.

That is not just culture. That is architecture.

The Counter-Narrative Has Already Started

The good news is that Africa is not only complaining about imported intelligence.

We are building.

Masakhane has been building African language NLP through a grassroots network of researchers, coders, and volunteers. Lelapa AI has built tools like InkubaLM and Vulavula to process African languages such as isiZulu and Sesotho.

This matters because language is not just translation. Language carries memory, humor, authority, emotion, worldview, and belonging.

When AI cannot hear us in our own languages, it cannot fully serve us.

When AI cannot understand our markets, it cannot fully support our businesses.

When AI cannot read our trust systems, it cannot fully measure our risk.

When AI cannot survive our infrastructure, it cannot fully work for our people.

The future of African AI must therefore be grounded.

• Low-bandwidth.
• Multilingual.
• Community-aware.
• Context-sensitive.
• Built for constraint.
• Trained on reality.

Not as a lesser version of Silicon Valley intelligence. But as a different evolutionary branch of intelligence.

One shaped by different pressures, and therefore capable of solving problems that imported models may not even perceive.

The Sling, the Stone, and the Future

The future of AI will not only belong to those with the biggest models.

It will belong to those with the most grounded models. The models that can work under constraint. The models that understand context. The models that can serve people whose lives are not clean, predictable, formal, or perfectly documented.

And this is where Africa has something powerful to offer the world.

AI trained only in stable environments may become impressive but fragile.

AI shaped by African reality can become adaptive, resilient, multilingual, low-bandwidth, community-aware, and deeply human.

David did not win by becoming Saul.

Africa will not win by becoming Silicon Valley with warmer weather.

David won because he trusted the intelligence his environment had already built in him.

Africa must do the same.

Not by rejecting the world.

But by refusing to enter the future wearing systems that do not fit our reality.

Because African reality is not a gap in the dataset. It is not noise. It is not disorder. It is not waiting to be rescued by imported intelligence.

African reality is data.

African reality is logic.

African reality is intelligence.

And it is worthy of intelligent systems.