AI: Why most Nigerian SMEs try AI and quit

The numbers look impressive on the surface. A Google and Ipsos survey published earlier this year found that 88 per cent of Nigerian adults have now used an AI chatbot, an 18-point surge from 2024. Another study by Zoho found that 93 per cent of Nigerian organisations have already begun their AI journey. Read those figures quickly and Nigeria sounds like a country charging headlong into the future. Read them more slowly, and a more complicated picture emerges.

Because beginning a journey is not the same as completing one. Trying a chatbot once or twice is not the same as weaving it into the daily running of a business. And for the overwhelming majority of Nigeria’s small and medium enterprises, which account for 96.9 per cent of businesses and contribute 46.32 per cent of national GDP according to the SMEDAN and PwC MSME Survey 2024, artificial intelligence remains something they have heard about, briefly experimented with, and quietly set aside.

The gap between awareness and sustained use is the real story. While adoption soars in headline terms, only 27 per cent of Nigerians say they know “a lot” about AI. Enthusiasm is running well ahead of understanding, and in business that tends to end in disappointment. Entrepreneurs sign up for tools they cannot fully operate, produce results they cannot use, and conclude that the technology simply does not work for them. Most of the time, the technology is not the problem.

The first and most stubborn obstacle is infrastructure. Nigeria faces infrastructure barriers such as limited broadband access, inconsistent electricity, and high data costs that constrain deeper AI adoption. These are not minor inconveniences. Running an AI-powered workflow requires a reliable internet connection, a charged device, and enough uninterrupted time to actually learn the tool. For a business operating from a shop where power may come and go unpredictably, none of those conditions can be taken for granted. Nigeria’s AI adoption rate stood at just 10.1 per cent in the first quarter of 2026, below both the global average and the 15.4 per cent average recorded across countries in the Global South. The infrastructure deficit is one of the clearest reasons why.

Cost compounds the problem. The biggest challenge Nigerian SMEs face is the cost of adopting AI. With limited budgets and strict banking rules, most businesses struggle to set aside funds for AI infrastructure. “The cost of AI technologies is a massive barrier,” says Dr Adebayo Okonkwo, a digital economy researcher at the University of Lagos. “For most SMEs, investing in AI feels more like a luxury than a necessity.” The most popular AI tools are priced in dollars at a time when the naira continues to face significant pressure. A monthly subscription that appears modest to a user in London or San Francisco can represent a meaningful share of a small business owner’s weekly income in Nigeria. When the return on that investment is unclear, and when the tool has not yet been mastered, many owners do the rational thing and cancel.

There is also the skills problem, which receives less attention than it deserves. Many Nigerian schools and technical institutions are not producing enough graduates with the expertise needed for AI. Even when businesses can afford the technology, they often lack skilled workers to use it effectively. This matters because AI tools do not run themselves. They require someone who understands how to prompt them correctly, interpret their output critically, and integrate them into existing processes. Without that knowledge, even a well-resourced business will struggle to extract value. Technical challenges such as inadequate infrastructure and outdated technology intersect with social issues including resistance to change and low digital literacy, exacerbating the divide.

The tools themselves are part of the problem too, even if that is rarely acknowledged in the promotional material. The dominant AI platforms were built primarily using data from the English-speaking West. Their defaults, their suggested use cases, and the tone of their outputs reflect that origin. For a Nigerian business owner whose customers speak in a particular idiom, reference local prices, and operate within a specific cultural context, AI-generated content often requires so much editing before it is usable that the efficiency argument simply collapses. While the cognitive alignment with AI’s value exists, the behavioural adoption is lagging, and the mismatch between what these tools produce and what Nigerian businesses actually need is one of the reasons why.

Trust plays a role too, though it is difficult to measure. Nigerian entrepreneurs have watched numerous technology waves arrive with grand promises and deliver inconsistent results. That earned scepticism does not disappear simply because a new category of tool is generating global excitement. Many Nigerian business owners are sceptical about AI, and given the history of overpromised digital solutions in the market, that caution is not unreasonable. It is, in many cases, wisdom accumulated through experience.

What tends to get lost in the broader conversation is an honest accounting of what sustained AI adoption actually demands. Initial investments required for software licences, hardware upgrades, and employee training can be prohibitively high. Many SMEs lack the capital to invest in these technologies if there are no quick returns. Businesses are discouraged from pursuing AI integration due to additional financial burdens caused by ongoing maintenance expenditure. For a business balancing suppliers, staff costs, rent and unreliable utilities, absorbing those costs while waiting for a return that may take months to materialise is a gamble most cannot afford.

The businesses that do manage to make AI work tend to have a few things in common. They are usually run by someone with prior digital experience, operating in a sector where AI output can be translated quickly into billable work, and with enough financial cushion to sustain a learning curve. Respondents who had adopted AI tools reported that the technologies enhanced operational efficiency, automated repetitive tasks, and improved inventory accuracy. Those benefits are real. But they are not uniformly accessible.

The challenge for policymakers, technology companies, and business support organisations is to stop measuring success by how many people have opened an account and start asking how many are still using these tools three months later. To harness the full potential of AI for sustainable development, it is crucial to address challenges through targeted policies that promote digital literacy, investment in infrastructure, and collaboration between governments, educational institutions, and the private sector. Without that, the cycle will continue. Nigerian SMEs will keep trying AI. And most of them will keep quitting.