Artificial intelligence is no longer a distant promise for entrepreneurs in developing economies. It is already embedded unevenly, imperfectly, but unmistakably in how Africa’s micro, small, and medium-sized enterprises (MSMEs) market their products, manage customers, transport goods, design offerings and handle finances. What remains unresolved is not whether AI matters, but whether the conditions exist for it to work at scale.
A new publication from UN Trade and Development (UNCTAD) makes that distinction clear. Artificial intelligence, it argues, is increasingly shaping how entrepreneurs in developing countries start, manage and grow their businesses. Yet its impact depends far less on the sophistication of the technology than on the surrounding ecosystem skills, governance, finance, data and trust.
Across Africa and other developing regions, MSMEs and start-ups are already deploying AI-powered tools across a wide range of activities, including marketing, customer service, logistics, finance and product design. Large language models, in particular, are emerging as foundational building blocks, allowing smaller firms to adopt AI quickly and at relatively low cost.
The appeal is straightforward. AI offers entrepreneurs the ability to raise productivity, cut operating costs, improve decision-making and reach markets that once felt out of bounds. For firms operating with thin margins and limited staff, even modest efficiency gains can be transformational.
But the report is careful to temper optimism with realism. AI adoption remains uneven and in many cases shallow. Many entrepreneurs report a limited understanding of the business value of AI what problems it can solve, how it fits into long-term strategy and how to implement it step by step. For smaller firms with limited resources, a lack of managerial understanding and technical talent slows implementation, sometimes by months or years.
“AI can be a powerful tool for entrepreneurs, but its benefits are not automatic,” said Arlette Verploegh Chabot, who leads UNCTAD’s work on entrepreneurship development. “Supportive ecosystems, the right skills and clear rules are essential to help entrepreneurs understand where AI adds value and how to integrate it into their businesses in a meaningful way.”
That message resonates across African markets, where enthusiasm for digital tools often runs ahead of institutional capacity. In Lagos, Nairobi, Accra and Kigali, founders are experimenting with chatbots, automated marketing tools and AI-assisted design. Yet beyond these hubs, adoption drops sharply a reminder that technology alone does not bridge structural gaps.
Connectivity Remains the First Hurdle
AI depends on connectivity, data and computing power. In that sense, the digital divide remains the most basic constraint. Only 27 per cent of people in low-income countries are online, compared with more than 90 per cent in high-income economies. In least developed countries, roughly 65 per cent of the population remains disconnected, keeping AI out of reach for millions of potential entrepreneurs.
This matters because AI tools, even relatively simple ones, rely on stable internet access and digital literacy. Without those foundations, the promise of automation and intelligence remains theoretical.
Where connectivity does exist, skills quickly become the binding constraint. Many entrepreneurs see AI as complex, opaque or poorly aligned with their immediate business needs. The report finds that this is often less about resistance to technology than about uncertainty over costs, over risks, and over whether AI will genuinely deliver returns.
Women entrepreneurs face an even steeper climb. According to the report, women are about 25 per cent less likely than men to adopt AI, a gap driven largely by confidence and familiarity rather than a lack of interest or ability. Younger entrepreneurs report widespread use of AI tools but say training on ethics and responsible use is insufficient, a warning sign as AI systems become more deeply embedded in business decisions.
Why Regulation Matters More than Speed
As governments scramble to respond to rapid advances in AI, regulation has become a focal point of debate. For entrepreneurs, the issue is not whether rules exist, but whether they are clear, predictable and proportionate.
“AI regulation is not about slowing down progress,” said Amandeep Singh Gill, the UN’s top official overseeing cooperation on digital and emerging technologies. “It is about making sure innovation happens responsibly, whereby clear rules help build trust and create a fair playing field.”
For MSMEs, poorly designed regulation can be as damaging as no regulation at all. Rules created with large multinational firms in mind often overwhelm smaller businesses, raising compliance costs and discouraging experimentation. The UNCTAD report highlights the value of risk-based approaches, sector-specific rules aligned with national development priorities and regulatory sandboxes that allow firms to test AI solutions safely while regulators learn alongside innovators.
In African economies seeking to attract investment while nurturing local enterprise, that balance is delicate. Too little oversight risks eroding trust; too much can suffocate innovation before it takes root.
Lowering the Cost of Experimentation
Access to affordable technology and finance is another decisive factor. AI adoption often requires upfront investment in software, infrastructure and critically, skills. Many MSMEs struggle to access credit or equity financing, particularly in markets where interest rates are high and venture capital remains concentrated in a handful of cities.
Open-source platforms and collaborative initiatives are emerging as practical solutions, enabling entrepreneurs to test and deploy AI without heavy upfront investment. Global technology firms are also positioning themselves as partners in this transition.
Natasha Crampton, Microsoft’s Chief Responsible AI Officer, argues that lowering barriers requires more than cheap tools. “True AI adoption” depends on accessible infrastructure, skills development and trusted governance ecosystems that allow entrepreneurs to learn, experiment and grow rather than simply consume technology.
Hiroaki Kitano, Chief Technology Fellow at Sony Group Corporation, advocates a phased approach to adoption. Firms, he suggests, should start with off-the-shelf tools, then collaborate with start-ups and partners to access specialised expertise, before gradually building in-house capabilities. For resource-constrained businesses, this sequencing reduces risk while building confidence.
Such approaches are already visible in parts of Africa’s tech ecosystem, where partnerships between start-ups, universities and multinational firms are filling skills gaps. Yet the report notes that without deliberate policy support, these initiatives risk remaining isolated pockets of progress.
Data: The Quiet Constraint
AI systems are only as good as the data they are trained on. In many developing countries, relevant local and sector-specific data are scarce, fragmented or prohibitively expensive. This limits entrepreneurs’ ability to tailor AI tools to local markets and realities from language and consumer behaviour to supply-chain dynamics.
Open data initiatives and data-sharing frameworks can help close this gap, provided they are designed with privacy and trust in mind. For African entrepreneurs serving informal markets or underserved communities, access to reliable data can be the difference between generic automation and meaningful innovation.
From Awareness to Capability
Across UNCTAD’s entrepreneurship network, one theme recurs. Skills are the decisive factor in AI adoption. Not just technical skills, but strategic ones, the ability to ask the right questions of technology, to judge where AI adds value and where it does not.
This is where targeted, hands-on support is beginning to make a difference. UNCTAD’s Empretec business development centres, operating in several developing countries, work directly with entrepreneurs to identify concrete AI use cases and build practical capabilities. Training programmes, mentoring and bootcamps combine core business planning with the use of accessible AI tools, helping firms move from curiosity to application.
Drawing on global evidence, surveys from the Empretec network and case studies, the report concludes that successful AI adoption depends less on the technology itself than on the ecosystem surrounding it. Skills, data, finance, regulation and trust.
On that basis, UNCTAD has put forward recommendations for an Entrepreneurship Policy Framework for AI adoption. The aim is not to prescribe a single model, but to equip governments with policy options that create enabling settings where AI can flourish as a practical tool for entrepreneurs across developing countries.
A Narrow Window
For Africa’s entrepreneurs, the stakes are high. AI arrives at a moment when the continent’s working-age population is expanding rapidly, while formal job creation struggles to keep pace. Used well, AI can help small firms scale, formalise and compete not only locally, but globally. Used poorly or left to deepen existing divides, it risks reinforcing patterns of exclusion.
The report’s message is ultimately pragmatic rather than ideological. AI will not automatically democratise opportunity. But with the right conditions, connectivity, skills, finance, clear rules and trust it can become a quiet multiplier for the kind of entrepreneurship that Africa already has in abundance.
The machines are learning fast. Whether entrepreneurs are allowed to learn with them remains a policy choice.
