AI and Semi-Formal Work: The Hidden Jobs Story in Emerging Markets
The majority of people in Africa, South Asia, and much of Latin America are not in formal salaried jobs. Most earn a living in the informal economy as farmers, traders, drivers, or small shopkeepers. Others operate in what we are calling the “semi-formal” layer: mobile money agents, delivery couriers, community health workers, or gig workers on digital platforms who earn money by providing services to formal enterprises. Then there is the much smaller formal sector, including office jobs in the government, financial institutions, customer service centers, and retail, as well as blue-collar jobs in manufacturing, construction, and agriculture.
The informal sector is the largest in most countries. In Sub-Saharan Africa, estimates put informal work at roughly 85 percent of total employment. In India, informality is also pervasive; the India Employment Report notes that around 90 percent of workers are in informal employment. In Latin America and the Caribbean, informality is lower but still sizable, at 47.6 percent in mid-2024 according to the ILO’s latest overview.
Semi-formal workers: where AI lifts performance
The semi-formal layer is harder to pin down in official statistics, but it is probably growing faster than the other layers. In 2024, there were about 28 million registered mobile money agents worldwide and roughly 10 million active each month, with most growth in Sub-Saharan Africa. This group represents as much as 2 – 3 percent of the labor force, and is growing quickly. India alone has about 12 million gig workers in 2025, the majority of them in delivery and ride-hailing, a number projected to reach 23.5 million by 2030. Though this number is still modest, it is fast-growing, accelerated by the rapid growth in digital infrastructure and platforms.
The typical semi-formal worker is a younger person with a high school education, often with a smartphone. The role itself relies on a combination of digital tools and human trust, positioning these workers to benefit from AI augmentation when AI can act like a co-pilot: helping a driver choose faster delivery routes, reminding a health worker about treatment steps, or flagging risks that a mobile money agent should double-check. The human still does the job, but with AI support, they can handle more cases, make fewer mistakes, and potentially earn more in the process.
The health sector offers some of the clearest evidence of the potential for productivity gains. Sub-Saharan Africa has roughly 637,000 paid community health workers (CHWs) and about 3.7 million volunteers who bring basic care to households that may never see a trained clinician. When those workers get digital decision support, outcomes improve. A 2019 randomized trial of a community health promoter model in Uganda found a 27 percent reduction in under-five mortality, achieved at very low cost per person. Looking ahead, AI-based diagnostic and triage systems can build on this foundation to further increase the reach and effectiveness of CHW programs.
Semi-formal workers already sit at the junction of platforms and people, and the companies they work for are often already highly digital with good access to AI resources and talents. AI adoption in this group will likely be high, allowing them to convert more orders, spot more health risks, and help more farmers improve their harvests.
Because semi-formal workers usually serve the last mile, AI adoption at this level also has the advantage of bringing AI productivity gains to end-users across the whole geography, including people with low education and few digital skills. This could drive fundamental and broad-based improvements in productivity, wellbeing and GDP growth. AI adoption in the semi-formal workforce may be both good business and good policy.
Formal sector: where replacement risks rise
The formal sector faces a tougher reality. The IMF estimates that around 40 percent of global employment is at risk from AI, with exposure highest in white-collar service industries such as finance, IT, and customer support. The risk is particularly visible in outsourcing hubs such as India and the Philippines but threatens white collar workers across the globe, especially those in junior roles. These are high-productivity sectors with relatively well-paid workers, so job losses have the potential to destabilize a nascent middle class.
Emerging markets are already seeing adjustments. South African banks use AI for compliance and fraud detection, and Kenyan firms are piloting AI chatbots. Rather than actual job loss, we’ll likely see a failure of these sectors to grow the number of jobs, further reducing the share of formal jobs in the overall economy.
Informal sector: incremental changes
The informal sector is dominated by economic activities that are low-productivity and low-income, such as smallholder farming and last-mile retail. Because of their low capital and technology components, productivity gains due to AI are quite remote and are likely to arise largely through the support of semi-formal workers. Though this could change in the medium term, it would involve large-scale transformations such as mechanization in agriculture. We are not ready to speculate on whether AI will speed this up, and for now would expect that productivity and employment effects in the informal sector will be low in the short to medium term.
Since the informal sector employs the vast majority of workers across Africa and the developing world, this could put a significant brake on economy-wide productivity improvements due to AI, even if the semi-formal sector is growing.
Will the economy gain or lose jobs?
The number of jobs in the economy will depend on three moving parts: informal workers will experience limited changes, semi-formal frontline workers will see their performance lifted by AI, and formal workers will face sharper risks of replacement. The outcome depends on how governments, investors, and innovators choose to act. At BFA Global, this is the focus of our work: helping build ecosystems where technology improves livelihoods rather than displaces them. Through the Catalyst Fund and the Jobtech Alliance, we support startups that use AI to expand access to health, finance, and gig opportunities.
Our approach spans multiple fronts. We’ve published a GenAI guide and toolkit with playbooks, prompt libraries, and case studies to help founders integrate AI into operations such as smarter job matching, fraud detection, and demand generation. We run masterclasses that make AI adoption accessible and actionable, enabling immediate integration into workflows. We’re piloting AI-powered tools for worker training and coaching, building assistants that reinforce onboarding and skills development, and developing financial modeling agents that strengthen women’s economic resilience. We’re also supporting platforms to adopt AI-enabled interview preparation and self-service training solutions that reduce costs while expanding access to quality learning. Together, these initiatives demonstrate how AI can directly improve livelihoods in emerging markets by lifting worker performance, expanding opportunity, and enabling inclusive growth.
If adoption is steered toward augmentation and inclusion, AI can raise incomes, spread opportunity, and strengthen resilience for millions. If not, the shift could deepen inequality. The choice is still open, and the time to shape it is now.
