Azeezat Raheem, a Silicon Valley expert in software and systems engineering, shared insights in an interview with The Nation on how African tech ecosystems, particularly Nigeria—the continent’s largest market—can proactively prepare for the era of agentic AI (autonomous decision-making AI systems).
She emphasized the importance of addressing unique regional challenges, adopting agile development practices, improving data collection practices, and leveraging emerging opportunities.
Azeezat highlighted that agentic AI represents the next major frontier for technology, with leading companies such as Microsoft, Nvidia, and Alphabet eagerly anticipating its implementation in 2025 and the budding African Market, particularly the Nigerian tech ecosystem is a fertile ground for building localized solutions that can solve problems in agriculture, healthcare, and education. She nudged tech companies and ideation hubs to provide solutions that would rival the counter solutions provided without recourse to local needs.
She said to think of “AI as a chef preparing meals. If the chef only knows European/Asian recipes, they won’t satisfy customers craving jollof rice or injera. Local flavours (data and context) make the AI relevant and useful”.
She also mentioned that Data is the fuel for AI and argued for localized data collection, storage, and governance.
She said: “Without clean and plentiful fuel, even the most advanced engine (AI) won’t run effectively. Africa needs its refined fuel source rather than relying on imported or irrelevant data that would not translate to immediate solutions to our local problems.
For example, in Nigeria, platforms like Helium Health digitize healthcare data, enabling better decision-making in hospitals and clinics. More digitization is needed for effective implementation of Agentic AI across all other sectors”
When asked if tech companies can achieve this without recourse to other sectors of the economy, she responded that it can never operate in isolation. She explained that “Agentic AI solves complex, multi-step problems by following a four-step process: perceive, reason, act, and learn.
It gathers and processes data from various sources to understand its environment, uses advanced reasoning powered by language models to generate solutions and coordinate specialized tasks, and executes plans by integrating with external tools. Guardrails ensure tasks are performed accurately and safely.
Over time, it learns and improves through feedback, making it more effective and adaptable for tasks like customer service, decision-making, and operational efficiency. Hence, Agentic AI can be used within and across entwined sectors.
Collaboration in the form of partnerships among governments, academia, and private sectors to build Agentic AI capabilities is crucial for its actualization and efficiency. A good example of collaboration in East Africa is the Rwanda AI Labs supported by the government and private partners, which explores AI applications in healthcare and agriculture.
A closer example is the Nigerian government’s Federal Ministry of Communications, Innovation & Digital Economy partnership with Google which focuses on upskilling talent and building AI technical solutions for local challenges.
Discussing the question of Ethical AI, she emphasized that African companies must prioritize ethics from the inception of Agentic AI design to ensure systems align with societal values and cultural norms. Africa’s diverse languages, traditions, and community-oriented ethos require AI solutions tailored to its unique context, as seen with platforms like Masakhane, which develop natural language tools for African languages, and fair implementation of AI across all sectors and regions particularly in education and health to prevent bias. Customers are increasingly aware of their privacy rights, driven by regulations like Nigeria’s NDPR and South Africa’s POPIA, but these laws often lack the depth to address ethical AI challenges such as algorithmic bias and accountability. Governments must expand these frameworks to include clear guidelines for ethical AI, ensuring fairness, transparency, and protection from misuse. By embedding ethics and cultural awareness into AI, Africa can foster trust, drive inclusive innovation, and set a global example for responsible AI development.
When asked how Agentic AI intersects with traditional software development which has gained traction in Africa, particularly in Nigeria, she reiterated the role of Agentic AI and stated “Agentic AI is like a personal assistant who anticipates your needs. For farmers, it’s like a helper who monitors the weather and recommends planting times without being asked and even recommends essential pre-planting operations.”
She further explained that the development of Agentic AI intersects with software development by enabling the creation of intelligent, context-aware software systems that anticipate user needs and provide actionable insights.
For example, software developers can integrate AI models with real-time data sources like weather APIs and IoT sensors to recommend optimal planting times for farmers. This requires building secure and intuitive platforms, such as mobile apps and APIs that translate complex AI capabilities into simple, user-friendly interfaces.
Additionally, developers embed feedback loops to continuously improve AI performance and ensure ethical use of sensitive data, such as location and crop yields. By bridging advanced AI technologies with practical, user-focused solutions, software development makes Agentic AI accessible and impactful across industries.
Speaking about her published article, on continuous integration and deployment strategies, she explained that Agentic AI as these practices are essential for efficiently developing, deploying, and maintaining complex AI systems.
Raheem explained Agentic AI relies on multiple components, such as data processing, reasoning engines, and action interfaces, which benefit from the automation and iterative delivery provided by CI/CD pipelines.
Azeezat stated: “In my article, I delved into advanced CI/CD strategies that facilitate rapid iteration and ensure consistent delivery of updates. I also emphasized the significance of tools like Docker and Kubernetes for seamless deployment across various environments.
Additionally, I highlighted the importance of tracking key metrics, such as deployment frequency and mean time to recovery (MTTR), to maintain system reliability and foster user trust.
These practices address common challenges like dependency management and test reliability, laying the foundation for robust and adaptive agentic AI systems capable of meeting evolving user needs and dynamic external conditions”.
She further elaborated on the advanced innovations within software development operations (DevOps), such as containerization and orchestration, which align closely with the scalability requirements of deploying resource-intensive AI models.
Raheem noted that predictive analytics powered by AI can further enhance CI/CD pipelines by optimizing deployments and anticipating potential failures.
Moreover, fostering collaboration and a DevOps culture is crucial in Agentic AI development, as it involves integrating data security, ethical considerations, and domain-specific adaptations across interdisciplinary teams.
“These CI/CD practices not only streamline the development and deployment process but also establish a sustainable framework for innovation”, Azeezat said. “By enabling scalability, reliability, and cultural alignment, these strategies ensure that Agentic AI systems are well-prepared to tackle real-world challenges in competitive and ever-evolving technological landscapes”. She concluded.
She said: “My advice is simple the African tech ecosystem should adopt agile development practices, adopt latest CI/CD practices to scale Agentic AI development, align technology with cultural values, and train talent in advanced tools in AI and ML including Docker. I also call on African governments to strengthen data privacy and AI ethics regulations to drive responsible innovation”.
