By Johnson Idesoh, Group Chief Information and Technology Officer at Absa
For the past three or four years, the world has lived through what is probably the fastest adoption of a general-purpose technology in modern economic history, with systems becoming embedded in ordinary working life so quickly that companies, governments, and regulators have often struggled to understand the implications at the same speed that employees and consumers were already adapting their behaviour around them. And the economic implications became impossible to ignore almost immediately.
Analysts now estimate that AI could contribute trillions of dollars to global GDP over the next decade, although what has arguably mattered more in the short term is the speed with which corporate behaviour changed once executives realised this was no longer a speculative technology discussion but something capable of altering fundamentals like cost structures, labour requirements, customer engagement models, and competitive positioning simultaneously. And yet there is already a growing sense, particularly among the companies building these systems, that the generative AI boom may ultimately be remembered as only the beginning of a much larger technological transition: agentic AI and autonomous AI agents.
If generative AI introduced systems capable of producing human-like outputs from prompts, agentic AI pushes further into decision-making and execution, allowing software not merely to respond to instructions but to pursue objectives with a degree of independence that earlier systems simply did not possess. In practical terms, rather than asking an AI to complete a single task in isolation, a user might assign a broader objective altogether, such as improving customer onboarding, resolving compliance bottlenecks, or identifying unusual transaction patterns – all areas becoming increasingly relevant for financial institutions such as Absa -after which the system determines for itself which steps need to be taken, which software tools to access, which sub-tasks to prioritise, and how to adapt if the initial approach fails – with minimal human intervention.
The question now is what this next phase of AI means for Africa.
Every major technological transition tends to expose the same underlying divide in the global economy: the countries and companies building foundational systems accumulate disproportionate economic power, while those arriving later often settle into the role of consumers and downstream adopters of technologies designed elsewhere. There is a risk that AI follows a similar pattern, particularly as the infrastructure underpinning it consolidates within a relatively small number of countries and firms.
And yet Africa’s technology story has often unfolded differently from conventional expectations. The continent’s digital evolution has rarely followed the linear path seen in developed economies, partly because institutional gaps and infrastructure constraints created pressure for entirely different forms of innovation. Mobile money, for example, emerged in environments where large segments of the population had never fully entered traditional banking system, a shift Absa has witnessed across multiple African markets over the past decade, and fintech ecosystems expanded rapidly because millions of consumers were already solving practical problems around payments, identity, access, and informality long before regulators and incumbent institutions had fully adjusted to the scale of the shift underway. Combined with one of the youngest populations in the world and accelerating smartphone adoption, this created conditions where new forms of digital behaviour spread far faster than institutional systems were prepared for.
In many respects, Africa became accustomed to technological leapfrogging precisely because existing systems were incomplete, and the same conditions that once spurred mobile-first innovation across the continent may now create an unexpected advantage as AI moves into its next phase.
The countries that will benefit most are likely to be those investing early in people, governance, and locally relevant application. Africa’s young population represents a major advantage here, but only if young people are equipped to shape these technologies themselves rather than experience them primarily as consumers of systems designed elsewhere. It will also require stronger coordination between banks, regulators, telecommunications companies, technology firms, and policymakers as autonomous systems begin integrating more deeply into everyday economic activity, something Absa believes will be critical to ensuring AI adoption remains trusted, secure and inclusive. Most importantly, agentic AI in Africa has to solve African problems in ways that feel practical and grounded, whether around financial inclusion, small business growth, fraud prevention, or expanding access to reliable services.
The generative AI boom may have introduced the world to machine intelligence. The agentic era will test which societies are able to apply it in ways that actually widen human and economic possibility.
