How African Industries Benefit from Open Source, Domain-Specific Models & Platforms

Enterprise leaders and developers can end up talking about AI in a very abstract way, using phrases like “unlock our true potential” or “give us valuable insights into our business”. It’s true, AI can do those things. But when it comes to actually following through, the trick is to be as specific as possible.

Enterprises across Africa face the challenge of adopting AI at scale and keeping up with their competitors and industries at large, but they also need adoption to deliver company-wide impact and open the door to future value generation. According to research by PwC, African organisations demonstrate strong intent in adopting AI, but are falling behind the rest of the world in terms of unlocking measurable returns.

Let’s take a step back and think about the fundamentals. Enterprises need models, tools and platforms that let them shape AI according to their digital environments and not rely on foreign, costly architectures. Models must be mapped to specific use cases, and not the other way round, and platforms must enable enterprises to move workloads between those models as they need. Being able to do so, regardless of the use case, helps enterprises maximise their control over their systems and set themselves up for measurable success.

‘Model behaviour’ for enterprise AI adoption

Looking back at the first few years of the AI revolution, the focus was, understandably, on generative AI (GenAI) and companies leveraging large language models (LLMs) to reason and generate content.

What made LLMs like ChatGPT, Claude and Gemini capture the imagination of business leaders, and eventually their IT budgets, was their inherent versatility. Here was a technology that would automate business functions, reduce costs and drive productivity and overall efficiency. And today, the use of enterprise LLMs globally extends beyond isolated deployments to comprehensive ecosystems that combine data, workflows and decision-support systems.

However, LLMs and frontier models only tell half of the enterprise AI adoption story. Like how organisations across Africa leverage different IT environments, including hyperscaler, on-premises and edge, many will end up using a mix of proprietary, open source and domain-specific models. This will include small language models (SLMs) that require fewer computational resources, which organisations can integrate with their own secured datasets and are thus compliant with local regulations.

Make no mistake, there is a place for LLMs in every enterprise alongside smaller models. But that combination reflects a need for IT platforms that not only allow for streamlined management, but also let enterprises build, deploy and move models between environments as they need them to. For businesses across Africa, ‘model behaviour’ for enterprise AI is open, flexible and consistent.

Built for purpose and for the industry

Unlocking measurable returns with AI requires organisations to understand what objectives they want to achieve using the technology and how the technology needs to be adapted based on those objectives. While Africa’s workforce has embraced AI, according to PwC’s Africa Workforce Hopes & Fears Survey 2025, 64% of African workers report they have used AI tools in the past year and 72% expect productivity gains within three years. But adoption isn’t just defined by using ChatGPT at one’s desk.

Financial services remain the best industry example. Earlier this year, the Central Bank of Nigeria formally incorporated AI into its money-laundering framework and mandated institutions to deploy automated systems to detect financial crime. In practice, this involves organisations analysing data like a geographical distance from a previous account transaction or use of a PIN and determining whether that transaction is fraudulent. That process requires a sensitive dataset and is thus executed with the use of a specifically trained model.

Agriculture is another example. Recent reporting shows that AI models developed overseas are failing to analyse data related to Africa’s agricultural production. In response, scientists have collected data and developed models to map and classify crops and estimate yields, resulting in models that account for the context of local farmers and deliver more accurate results.

These examples show the versatility of model development and how enterprises, regardless of their function or the products and services they offer, can apply AI to enhance existing processes. The possibilities of AI are limitless, so long as it’s built for purpose and compliance.

The foundation makes all the difference

AI adoption in Africa does not happen in a vacuum, nor is it the result of a single proprietary solution. Enterprises are on a path to discover how new technologies work best for them; and in the process of making them work, they need to prioritise platforms that are backed by resilient hardware ecosystems and accessible architectures.

The value of open source solutions like Red Hat AI is that they prioritise flexibility, portability and interoperability. Enterprises can develop and deploy AI solutions across hybrid cloud environments, easily build and customise models and connect them to data, and ensure collaboration among developers thanks to a consistent user experience. By grounding their AI strategy in open source, enterprises gain the structural agility and cost efficiencies needed to move models out of experimentation and into production easily and at scale.

Getting specific with what businesses want AI to do for them is the first step to having the technology deliver value, and that goes for the models they build, deploy and adapt. And even if that’s still uncertain, by prioritising open platforms that accelerate innovation, businesses in Africa are set up for wherever their AI journey takes them.

By Oluwafiropo Tobi Ogundare, Regional Sales Lead for West Africa and Mauritius at Red Hat