By Chris Maclay

This is the last of a four-parter about job creation in Africa. It looks at which digital jobs can still be created in Africa in the age of generative AI. Part 1 focuses on how jobtech platforms in Africa, when they’re at their best, are truly creating new jobs for Africa. Part 2 looks at how digital work platforms need to create work for global companies rather than just selling it, and how we need to build up digital trade routes. Part 3 explains how we’re creating roles for Africans by supporting African GBS providers to set-up shop in markets of demand.

Humans get scared about technological shifts. They always have. 

Samuel Butler (1863)

African countries had been hoping to ride the demographic dividend wave to become ‘the labour basket of the world’. But today, as the world wrestles with the impact of AI on labour markets, we are left questioning whether the bus has left the station before we in Africa have even reached it.

This blog does not seek to comprehensively explain the impact of AI on African digital labour markets, but provides one framework for thinking about the global knowledge work roles that will continue to exist over the coming years, and that could be delivered from Africa. As the nature of digital labor is being rewritten in real-time, Africa needs to strategically position itself for the roles that will exist.

What generative AI really means for the labour market in Africa

If you were to visualise a ‘typical job’ in a country like Kenya, where 83% of employment takes place in the informal sector, it would probably look something like that of a mama mboga.

Above: A mama mboga – a saleswoman vending vegetables typically from a temporary store or sheet (Source: Farmerstrend)

The advent of generative AI isn’t really going to change much for the day-to-day work of the mama mboga (though behind-the-scenes efficiencies in AI-enabled logistics could reduce her costs, etc).

The good news is that this means that the continent may not face the complete societal upheaval that the global North is slowly (and soon, fast) going to face as AI replaces much of the knowledge work where over half of people are employed.

Yet most African countries’ visions for economic transformation relied on moving more people into high paying knowledge work jobs.

Some of us were thrilled about the post-COVID era when all of a sudden any job could be delivered from Africa. ‘Africa’s is going to leapfrog manufacturing,’ we said, ‘ and move straight into services’. Knowledge work was the future of work in Africa.

That enthusiasm was heavily tempered by the launch of ChatGPT in late 2022. Since then, AI has begun rapidly automating routine knowledge work tasks, reducing demand for many entry-level roles once seen as gateways for young professionals, and once seen as primed for outsourcing to ‘the world’s new labour basket’. 

Below: Percentage of tasks in outsourcing roles susceptible to automation by AI

Customer Service:

IT-enabled Services

Finance & Accounting

Source: Genesis Analytics and Caribou Digital

The number of jobs available for African freelancers on platforms like Upwork has already reduced by 21% since 2022. This trend is likely to continue, particularly for entry-level roles. 

While AI will certainly create new opportunities and make others more efficient, there can be no doubt that generative AI will cause less global roles to be available for Africans than without the advent of AI. At least for the next decade.

So what’s left? The AI-aligned Labour Stack

This is not to give up on knowledge work, or the potential of Africans to contribute to global labour markets. The growth of digital work, ITO (IT outsourcing) and ITES (IT-enabled services) will be important to labour markets in Africa, whatever scale it takes. Therefore, we need an understanding of what those roles might be, and see if we can position Africa and Africans effectively for those roles. 

At the Jobtech Alliance, we’ve been lucky to work with a number of innovators building jobtech solutions at the intersection of AI and human capacity. As we look at the solutions being built, we see three main categories of digital work emerging. 

The table below applies this framework to a few of the Jobtech Alliance portfolio to present the distinct roles in the ICT value chain, before we explain these archetypes in more detail:

	Architect	Implementer	Human in the Loop
Execo	Works with clients to design integration of AI into contract lifecycle management	Dev team builds LLM and conducts initial scraping	Double-checks scraped data to train and improve model
Elewa	Works with customers to design project eg. app development	Works with client to implement project	Ongoing dev maintenance and support

Architects: As we wrote in Part 2 of this series, global digital work won’t come to Africa unless it is ‘made’. As the table above demonstrates, the role of the architect goes beyond ‘just selling’, but plays a significant role in ‘creating the work’ for both implementers and humans-in-the-loop.

Most global companies have a strong reversion of the status quo of traditional local hiring. If we want roles, functions, and tasks to come to Africa, we need ‘architects’ to support global companies to identify specifically how to effectively outsource these functions. 

As we wrote in Part 2, “The core theme is not one of selling to an existing latent demand for outsourcing; it is creating the work itself for the African market.” 

An Architect is responsible for the strategic work of product or process design, and choosing how work can be packaged to be done by Africans, by AI agents, or a combination of the two. Which roles still need to be done by humans in the company locally? How could the work be packaged to be delivered from abroad? Which tasks can be done by AI agents? Which ongoing functions could better-be served by outsourced labour? 

As Part 3 explained, Jobtech Alliance portfolio company Elewa is able to generate work for its team of engineers in Africa by working with customers (mostly start-ups in Europe) to design their project; from structuring an app development process to a digitization journey. Similarly, as Part 2 explained, Jobtech Alliance portfolio company, Robin, uses the entry point of ‘integrating AI’ to package work to be done by AI and humans-in-the-loop. As founder Phyl Georgiou explained, “What we call the ‘architecture’ phase is really our way of uncovering the customer’s true needs—it’s a strategic step that allows us to design and deliver the right product.” While companies are seeking opportunities to automate and build efficiencies with AI, they still need ‘architects’ to strategically plan how they do so, and how job descriptions will be unbundled and rebundled given how AI changes who does what work.

And by doing so, those architects not only create work for themselves, but function as entry points for packaging work for implementers and humans in the loop, as we’ll cover next. 

Implementers: There was, and will continue to be, a role for African developers implementing ‘tech projects’. These will increasingly involve building or adapting/implementing AI solutions such as agents. While the architects are responsible for identifying what role AI solutions could play in business process flows, the implementers are responsible for building the AI solutions.

Jobtech Alliance portfolio company, Execo, helps global companies in areas like contracting, contract lifecycle management, and operations support by combining AI with high-performing global teams. After global teams have architected (designed and sold) the project with the client, a team of 30+ engineers (or ‘implementers’) in Kenya conduct part of the work on these projects, such as building code to scrape data off the web. Then a team of hundreds of humans-in-the-loop quality check and amend that data (human-in-the-loop roles – more below). 

In the same way that the architect can create the work for the implementer, these two can create ongoing work for humans in the loop, as will be discovered next.

One dynamic of this framework, is that the role of implementers will likely naturally dissolve over time, with either the competent Implementers becoming architects, or humans in the loop being able to implement off-the-shelf solutions or no-/low-code tools.

Humans in the loop: This category represents what we traditionally would have thought of as ‘outsourcing work’; the ongoing knowledge work required to support global companies with their labour requirements. This ongoing labour function is likely to be less repetitive than historical BPO roles, but could be incredibly vast in type; engineering work to maintain tech products, customer success roles, communications, or even AI data annotation or training. These roles will increasingly involve utilization of AI, or indeed oversight of AI agents, quality control or escalation responsibilities.

While the quantity of these roles over the coming decade will be less than it might have been before the advent of AI, and these ‘HITL’ individuals will likely be utilizing AI tools to achieve efficiencies themselves, there will still be large quantities of roles for humans in the loop. 

Jobtech Alliance portfolio company, Tana, provides global companies with quality assurance engineers, data analysts, and technical support engineers; perfect examples of human-in-the-loop roles. 

Where Africa Wins

While this framework provides an initial mental model, and it will necessarily keep evolving given the pace of AI development, we hope it provides some framing for how to think about the opportunities that will emerge and continue to exist for IT outsourcing in Africa during this period of profound upheaval to the labour market.

Architect roles are naturally the most resilient to AI themselves, and effectively control the downstream value chain. Controlling this leverage point is how African platforms move from being ‘just’ delivery partners to demand shaping. It is therefore critical that Africans are able to take on these roles, both as individuals and as African GBS providers. While being small in number, they are high value and have significant potential downstream value creation.

One challenge of this in global markets, however, is that these roles can often require proprietary insight into the workings of the hiring companies in those markets. This can make it difficult for Africans without global work exposure to get into these roles; it takes global experience to get global experience. Potentially we might initially lean on African diaspora to initially unlock these opportunities for Africans back home. Moreover, we need to provide opportunities for implementers to grow into architects.

AI won’t destroy all digital work. But it will change where value sits. The labor of the future won’t be in typing—it will be in orchestration. Designing the system. Running the agents. Adding judgment where the machines fall short. That’s what we need to train for. And that’s what African jobtech platforms need to enable.

As we’ll cover in our soon-to-be-published sector scan on the future of digital work in Africa, we believe that there are particular opportunities when platforms consider (among other things):

• Presence in global markets, to play a bigger role in architecting implementation and HITL roles to be done from Africa.
• How best to blend humans + AI in knowledge work.
• Unique human-in-the-loop benefits, such as leveraging demographic or geographic diversity.

We don’t need to fear AI. But we do need to rebuild our playbook. The jobs are still out there. We just need to rewire how we find them, package them, and deliver them from Africa.

The author is Mercy Corps’ Program Director at the Jobtech Alliance