Josephine Christopher is a senior business journalist for The Citizen and Mwananchi newspapers
Mwananchi Communications Limitted
Dar es Salaam. Tanzanian businesses are being urged to resist the temptation to let artificial intelligence make critical decisions about their customers, as the technology becomes increasingly capable of building products, analysing data and automating routine business tasks.
While AI can dramatically reduce the time and cost of developing digital products, businesses still need people who understand their customers well enough to determine what they actually need, according to Black Swan Chief Technology Officer Mr Harvey Kadyanji.
Speaking during Innovation Week Tanzania 2026, Mr Kadyanji said the rapid development of AI-assisted software was changing the economics of starting and running a business.
Entrepreneurs can now move from an idea to a functioning digital product in a fraction of the time previously required, lowering one of the traditional barriers to innovation.
But the ability to build something quickly does not necessarily mean there is a market for it.
“The problem is that, if you don't have the capital knowledge, AI is very confident,” he said.
That confidence, he suggested, could become a problem for businesses that lack the expertise to question an AI-generated recommendation.
Mr Kadyanji identified four areas where AI still faces limitations in business which include taste, which is the human ability to judge whether something feels right for a particular customer or market.
“You tell it to create something and it clearly just rushes into creating something that it thought matters to you,” he said.
Other Mr Kadyanji described as tribal knowledge which is the practical understanding of a particular industry that comes from experience.
AI can process large amounts of information, but it does not automatically possess the institutional knowledge of how a particular market operates.
“For a Tanzanian business, that knowledge may include how customers negotiate prices, how informal distribution networks operate, seasonal purchasing patterns or the practical challenges of dealing with suppliers and regulators,” he said.
The third issue is knowing what to build, as AI is making product development increasingly accessible, allowing entrepreneurs to create prototypes, websites, applications and analytical tools without large technical teams.
Understanding what customers are willing to pay for requires more than generating a product.
Knowing when the machine is wrong he said can be the biggest risk comes when businesses accept AI-generated information without sufficient scrutiny.
AI systems can produce answers with considerable confidence even when those answers are inaccurate or unsuitable for a particular situation.
The debate comes as Tanzania seeks to strengthen its innovation ecosystem and turn more ideas emerging from universities and young entrepreneurs into businesses capable of creating jobs and contributing to economic growth.
Policy and Partnership Manager at the FCDO East Africa Research & Innovation Hub Ms Njeri Gichuhi said collaboration between universities, government institutions and entrepreneurs was critical to building Tanzania's innovation ecosystem.
She said universities could increasingly serve as innovation hubs while support should extend beyond major urban centres to reach young people with ideas and potential.
StartHub Africa Co-Founder and Managing Director Shakila Mshana said Tanzania had no shortage of ideas but needed more entrepreneurs capable of executing them effectively.
The same issue was highlighted by Westerwelle Startup Haus Arusha Country Director Isaack Shayo, who said entrepreneurial opportunities were available but success depended on structured planning and strategic execution.
The message is significant in an environment where AI is making the early stages of entrepreneurship easier.
A founder can increasingly use AI to generate a business plan, analyse information, write software, produce marketing material and develop a prototype.
But none of those capabilities guarantees that customers will buy the product.
