As global investment in artificial intelligence surges, Africa is still searching for its own model, caught between enormous demographic potential and persistent structural constraints. High connectivity costs, limited computing power, and the scarcity of quality local data continue to slow the adoption of advanced technologies. Yet these limitations have also produced something remarkable: lean, highly optimized AI solutions built in environments of scarcity, but capable of competing with international models.

As the head of Zindi, the first pan-African data science competition platform, Celina Lee has seen how consistent practice, real-world problem-solving, and talent networks are accelerating the professionalization of an emerging sector.

In this conversation with Ecofin Agency, held on the sidelines of the Digital Transformation Summit in Cotonou in mid-November, she explains the levers needed to build an African AI market—from data structuring to skills development and public support—while also outlining why Africa does not need to replicate large models like ChatGPT to innovate effectively.

Ecofin Agency: Africa’s digital economy is growing, but adoption is slowed by the high cost of data, smartphones, and limited access to credit. Based on what you see on Zindi, how much do these barriers restrict the emergence of AI talent?

Celina Lee: Today, Zindi has nearly 100,000 users across the continent. They take part in competitions built around real-world problems and develop AI solutions. But the cost of connectivity and limited computing power are real obstacles.

We recently launched a competition that used roundabout surveillance videos to analyze traffic behavior. Even when compressed, the dataset remained massive—several hundred gigabytes. For many young Africans, downloading that much data is simply too expensive. Data is costly relative to average income, and access to a powerful enough machine is just as challenging.

That said, this environment has created an unexpected effect: it forces young people to innovate with extreme efficiency. In a competition we ran with the government of Mexico to map informal settlements using satellite images, the models proposed by African participants turned out to be far lighter and more optimized than the original ones. While others relied on large amounts of computing power, African data scientists found clever solutions that required far fewer resources.

Ecofin Agency: Have you measured Zindi’s impact on employability and access to global opportunities?

Celina Lee: We published a report with the Kenyan government analyzing the profiles and career paths of our users there. It showed that 18% of them obtained professional opportunities thanks to Zindi. We also found that participating in at least four competitions significantly increases the chances of getting hired.

Across Africa, our surveys show that 85% of users say they acquired new skills that improved their careers or professional growth. This confirms that regular practice and exposure to real problems truly accelerate learning.

Ecofin Agency: Many African companies remain hesitant to adopt AI due to low digital maturity and fragmented infrastructure. What realistic use cases can create immediate value for SMEs, especially in agriculture, finance, and logistics?

Celina Lee: What I see is that everything must start with the business problem, not the technology. In some cases, the right solution will be AI; in others, it will not. We need to move away from this idea that AI must be used everywhere.

For companies with low digital maturity, a first step might simply be helping employees use tools like ChatGPT to work more efficiently—writing, analysis, marketing, financial reporting, and so on.

Next come solutions already built into existing software, available through licenses. Only once a company reaches a more advanced level should it consider integrating AI at the core of its business model.

What stands out about African companies is their pragmatism. Economic constraints are strong; they adopt technology only when it brings clear, measurable value. They should not be pushed into “doing AI” for its own sake. They need support in progressing at a pace that fits their real needs.

Ecofin Agency: Africa faces a chronic shortage of quality local data, which directly limits innovation. What is needed to build an ethical, shared, large-scale data ecosystem? And who should lead it—governments, the private sector, or regional organizations?

Celina Lee: At Zindi, we have organized more than 500 competitions, each based on local datasets, mostly African. This demonstrates that most of the data an enterprise needs often comes from the enterprise itself. A company can train an AI model on its own history—its customer interactions, for example—without relying on external data.

However, for general-purpose models like large language models, the situation is different. They require massive, diverse datasets, similar in scale to the entire Internet. Very few African languages currently have enough digital content to support that.

To build a coherent African data ecosystem, structured collaboration is essential: governments must set ethical frameworks, companies must generate and share sector-specific data, and regional bodies must harmonize standards and ensure interoperability.

A common misconception with the rise of ChatGPT is the belief that huge datasets are always necessary to train effective models. In reality, AI is not just about large general-purpose models. A small or medium-sized business can build a much simpler model that performs a narrow set of tasks extremely well. On Zindi, we are seeing rapid growth in small language models—compact, specialized systems requiring far less data and computing power.

This trend is particularly promising for Africa: it makes innovation possible without massive infrastructure and fits the continent’s economic realities. Africa does not need to replicate ChatGPT. Smaller, specialized models can be far more effective.

Ecofin Agency: Several governments now want to develop “sovereign” African AI models. Is this economically realistic, or could these projects become symbolic, disconnected efforts?

Celina Lee: I do not think it is purely symbolic. It is achievable—provided the approach is pragmatic. Take Nigeria, for example. They recently launched N-ATLAS, a multilingual model covering five Nigerian languages, and made it open source.

They did not train it from scratch on a massive dataset. They started with an open model, Llama, and fine-tuned it with a smaller set of local data. This is a realistic path, and it works.

There is a full spectrum of possibilities: on one end, large foundational models that Africa cannot feasibly reproduce today; on the other, small highly specialized models that perform extremely well. The challenge is finding the right position on that spectrum.

Discussions are emerging about building a foundational model in Swahili, which has a far larger digital corpus than many other African languages. And countries like Benin have begun collecting voice data to develop models adapted to local realities. These initiatives show that the effort is already underway.

What matters most is that these projects remain aligned with practical use cases: linguistic inclusion, public services, agriculture, education. If they respond to real needs, they will be useful and sustainable.

Ecofin Agency: Zindi plays a major role in nurturing talent, but Africa still faces the risk of brain drain. How can the continent retain its best data scientists?

Celina Lee: First, we must be realistic: if a young person receives a great opportunity, they should take it. We cannot block them or blame them. The issue is not that they leave; it is that the local market does not offer enough prospects.

Governments need to invest in the entire ecosystem: support start-ups—often the first entry point into the workforce—encourage research and development, and create programs to train and integrate young graduates. The private sector is still reluctant to take risks, whether that means investing in AI or training early-career professionals.

If governments build a clear path to employment, young people will stay. They want to stay. They want to build at home. What they need is a real market that can absorb their skills.

Ecofin Agency: If the continent manages to provide reliable connectivity, affordable devices, quality data, and basic digital skills, what could it achieve in the next decade?

Celina Lee: The gains would be enormous. The continent is full of energy, creativity, and innovation. Young Africans are incredibly motivated. They tackle real problems, often in very difficult conditions, and they are just waiting for the environment to allow them to go further.

With reliable connectivity, affordable services, and access to electricity, the potential would be limitless. We have seen this in Kenya and Nigeria: ecosystems can grow very quickly when they are supported. I believe AI will follow the same trajectory. And what is exciting is that Africa can invent applications impossible elsewhere because its realities are unique. The continent could contribute something entirely new to the world.

Ecofin Agency: Which sectors should be prioritized for AI deployment in the next five years?

Celina Lee: On Zindi, agriculture stands out clearly. It is the sector where AI can create the fastest and most significant impact—yield forecasting, disease detection, irrigation optimization, weather analysis.

Climate change is also central. It is not an industry in itself, but it is a cross-cutting challenge where AI can make a major difference.

Then come traditional commercial uses: finance, retail, logistics, manufacturing. Wherever AI reduces costs or increases productivity, companies will adopt it.

Fiacre E. Kakpo