By Veerakumar Natarajan
Over the past two years, Generative AI has taken centre stage. In Kenya, like elsewhere, tools that write content, summarise reports and answer questions have captured attention. They help teams move faster and reduce routine work, this value is real says Veerakumar Natarajan.
But Generative AI is only one part of the AI story. For most Kenyan organisations, the bigger opportunity lies in how generative and agentic AI work together and when to use each. The goal is not replacement, It is progression.
Start With the Basics: Data and Workflows
AI is only as good as the data and processes behind it. This is where many organisations struggle. Across sectors, manual work is still common. Data sits in silos. Systems do not talk to each other. Records are inconsistent. These gaps make it hard to jump straight into advanced AI use cases. Even powerful models will fail if the inputs are weak.
That is why the smartest first step is often simple digitisation and automation. Routing customer tickets. Reconciling mobile money transactions. Managing field reports. Processing sensor data. These may sound basic, but they deliver fast results.
They cut errors, improve consistency and show leaders how information really flows. Most importantly, they create clean data. That foundation is what AI needs to work well.
GenAI Creates, Agentic AI Acts
Once workflows are stable, AI becomes more effective. Generative AI helps people create. It drafts, explains and speeds up thinking and communication. Agentic AI does something different, it takes action.
Agentic systems can suggest steps, check rules and then execute tasks within set limits. This matters in regulated sectors like banking, insurance and public services, trust and accountability are critical.
Think of a loan process. An agentic system can review an application, confirm KYC checks, validate limits and only then approve or escalate. Decisions are faster, but controls remain in place. That balance is key.
Why Local Context Still Matters
Global AI models are powerful, but they often miss local detail. Kenyan regulations, language use, business norms and sector terms are unique. Without this context, AI outputs can be inaccurate or risky. This is where contextual and sovereign AI models add value.
These systems are tuned with local data and aligned to local rules. They do not replace global models. They complement them. The result is AI that understands the Kenyan market and operates within it.
AI Can Support Local Innovation
AI also opens the door to broader participation. With no-code and low-code tools, advanced capabilities are no longer limited to large firms or experts. Small businesses, NGOs and county governments can build automations that fit their needs. A micro-insurer can improve claims handling. A county office can speed up citizen services. An agritech startup can better support farmers.
Innovation becomes decentralised, solutions come from people closest to the problem.
For leaders, the message is clear, start with workflow automation, fix the data. Use Generative AI where it boosts productivity. Introduce agentic AI when systems are ready for secure, auditable action. Add contextual models to ensure local relevance.
This step-by-step approach is more sustainable than chasing the latest model. Kenya’s AI future will not be defined by hype. It will be shaped by organisations that focus on fundamentals. Those that invest in data, design smart workflows and deploy AI responsibly will see real returns.
AI works best not as a flashy tool, but as a dependable partner one that helps teams do better work, every day.
