Global Enterprise AI expenditure is projected to reach approximately $644 billion in 2025, according to Gartner. But the independent research conducted on the outcomes of AI projects suggests that a majority of that investment will not be delivering the intended business value for which it was proposed and approved.
There's also a recurring story that plays out in boardrooms across Nairobi, Johannesburg, Accra, and Lagos with uncomfortable regularity. The CEO approves an AI initiative, and the technology partner is selected. The pilot runs for three to six months, and the prototype results look promising.
Then, when scaling to large-scale production, they encounter multiple errors, and the promised result isn't delivered. The process ends, and the vendor is ousted. The organisation returns to using spreadsheets, manual approvals and WhatsApp threads. The budget for AI is simply a line item that nobody is interested in discussing.
What the research data actually shows
The rate of failure for AI projects in 2026 isn't a secret, though the exact value in circulation needs more scrutiny than they usually receive. What surprises us is that this rate has remained high and stubborn over the years despite the clear evolution of knowledge, access to the best tools, and working with intelligent engineers.
Subscribe for free to continue reading this article
Subscribe SubscribeAlready Have an Account? Log In
