NAIROBI, Kenya, Aug 25 — Artificial intelligence is moving deeper into Kenyan businesses, but the next phase of adoption is likely to be defined less by how many companies are using the technology and more by whether they can turn it into measurable returns.
The question is no longer simply whether Kenyan businesses are using AI. It is whether the technology is changing how they operate — cutting costs, improving productivity, increasing sales or creating new sources of revenue.
Recent research suggests the country has moved rapidly into AI adoption, with businesses deploying the technology in customer service, software development, marketing, credit scoring, fraud detection and other functions.
But the distance between deploying AI and scaling it across an organisation remains significant.
A 2025 survey of Kenyan business professionals found that 96 per cent of organisations had begun their AI journey, while 35.2 per cent had achieved widespread or advanced implementation. Lack of technical expertise was identified as a major barrier by 48.8 per cent of respondents.
Customer service was the leading area of AI investment, cited by 54.8 per cent of respondents, followed by software development at 51.2 per cent and marketing optimisation at 36.2 per cent.
The figures point to a market that has largely moved beyond asking whether AI has a place in business. The harder question is whether companies can integrate it deeply enough to produce measurable business value.
From experimentation to business value
For smaller businesses, AI is lowering the cost of accessing capabilities that previously required specialist teams or significant investment in technology.
Abraham Mbuthia, CEO and co-founder of digital business solutions provider UzaPoint, said AI was compressing the time required to perform tasks that previously demanded substantial human effort.
“Some of the tasks that used to take maybe months, years to be accomplished are now getting done in a very short period of time.”
“With the AI tools, you can actually gather information and simply get a website in a day or two days.”
Mbuthia said the technology was also narrowing the gap between smaller businesses and larger firms with deeper technical and analytical resources.
“With AI now, as small businesses, we’re giving you the same capacity for larger businesses,” he said.
For smaller firms, that access can change the economics of technology adoption.
Instead of building expensive infrastructure or maintaining large specialist teams, businesses can increasingly access AI capabilities through existing software, cloud platforms or external vendors.
The 2025 survey found that only 11.3 per cent of Kenyan organisations were pursuing in-house AI development, with others relying on custom vendor solutions, AI embedded in enterprise applications or hybrid approaches.
But lower barriers to entry do not automatically produce a return on investment.
The commercial test is whether AI changes the economics of an operation — whether a process becomes faster, cheaper, more accurate or capable of generating additional revenue.
Banking takes AI into high-stakes decisions
The financial sector offers one of the clearest examples of AI moving beyond experimentation into core business functions.
A March 2025 survey by the Central Bank of Kenya found that financial institutions were using AI and machine learning for credit-risk assessment, cybersecurity, customer service, electronic know-your-customer processes and fraud-risk management.
CBK identified credit scoring, fraud detection and personalised customer engagement as leading areas of sophisticated AI integration, while some institutions were already using generative AI to improve operational efficiency.
The regulator, however, highlighted significant constraints.
“Critical gaps remain in data quality, governance, and management.”
CBK also identified dependence on third-party providers, shortages of skilled personnel, the resource intensity of AI and compliance requirements as challenges to wider adoption.
That makes the business case for AI in financial services more complex than simply reducing transaction costs.
An algorithm used to assess creditworthiness or detect fraud has to be fast, but it must also be accurate, explainable and subject to appropriate human oversight.
For lenders, the potential upside is significant. Faster credit decisions, improved fraud detection and more personalised customer engagement can affect both costs and revenue.
But errors can carry equally significant financial, legal and reputational consequences.
The banking experience therefore illustrates a broader lesson for Kenyan businesses: AI becomes more valuable as it moves closer to core operations, but the risks also increase when technology begins influencing high-stakes decisions.
Manufacturing faces a cost equation
Manufacturing presents a different test.
Tobias Alando, Chief Executive Officer of the Kenya Association of Manufacturers, said local companies were increasingly investing in software technology supporting manufacturing operations, although direct AI-led production remained more difficult because of the cost of machinery and technology.
“The trend that we have typically noticed is investment coming in into software technology that supports production in the manufacturing sector.”
“The challenge is the costs of this machinery or with the AI technology. It is still very expensive for developing countries like Kenya to adopt it fully.”
The distinction matters.
Software-based AI can increasingly be accessed through relatively affordable cloud services and enterprise applications. AI-enabled industrial automation, by contrast, can require investment in machinery, sensors, connectivity, data infrastructure and specialised skills.
For manufacturers operating under tight margins, the capital requirement can make full-scale adoption difficult.
Yet technology is increasingly becoming a competitiveness issue.
Kenyan producers are competing against manufacturers in highly automated markets where technology can lower production costs, improve quality control, reduce downtime and optimise energy use.
Alando said manufacturers would increasingly have little choice but to adopt emerging technologies.
“It will not be optional,” he said. “It will be mandatory to have this technology.”
That is likely to encourage selective adoption, with businesses first targeting processes where the return can be demonstrated.
Safaricom shows the revenue opportunity
AI’s commercial value, however, is not limited to cutting costs.
Safaricom has cited AI-driven customer segmentation as one factor behind the growth of its voice business, which increased by 50 per cent over five years despite expectations that the market was mature.
During the same period, mobile-data usage tripled.
The company said AI and machine learning helped it understand customer behaviour and develop personalised voice packages, demonstrating how AI can be deployed not only to make an existing operation more efficient but also to identify opportunities for additional revenue.
The example points to a potentially more valuable phase of AI adoption.
A technology that saves a company money once is useful. A system that continuously helps a company understand customers, develop products and increase sales can have a much larger effect on growth.
That distinction is becoming increasingly important as Kenyan businesses move from experimenting with individual AI tools towards integrating the technology into commercial decision-making.
The pilot problem
PwC’s 2026 AI Performance research provides a warning for companies rushing into AI.
The study found that 82 per cent of African organisations were running AI pilots, but few had scaled them across the enterprise.
African organisations were investing an average of 2 per cent of revenue in AI, compared with 5 per cent among global leaders, while only 32 per cent said their AI investment was sufficient.
PwC Africa Chief Executive Officer Dion Shango said the challenge was not simply to increase the number of experiments.
“The organisations that will win are not those running the most pilots, but those that scale the right AI to transform how they create value.”
The global findings underscore the divide between experimentation and performance.
PwC found that the top 20 per cent of companies captured 74 per cent of AI-driven returns, while the strongest AI performers generated 7.2 times greater AI-driven performance than their peers.
The difference is not necessarily the number of AI tools deployed.
Leading companies are more likely to redesign workflows around AI, pursue new revenue opportunities and use the technology to rethink business models rather than simply automate existing tasks.
For Kenyan businesses, that distinction could determine who benefits from the technology and who merely adds another layer of technology spending.
Skills could determine who captures the gains
The technology challenge is also a people challenge.
PwC found that 64 per cent of African workers were already using AI in their roles, suggesting employees may in some cases be adopting the technology faster than organisations are developing formal strategies around it.
For companies, the issue is therefore not simply whether workers can operate AI tools.
It is whether employees understand where AI should be deployed, how its output should be assessed and which decisions should remain under human control.
This becomes particularly important in areas involving finance, human resources, healthcare and customer data, where poor-quality AI outputs can create financial, legal or reputational risks.
The transition will therefore require more than buying software.
Businesses will need to train employees, redesign workflows and establish clear rules around accountability and human oversight.
The organisations that manage that transition effectively could have an advantage over competitors that treat AI as an isolated technology project.
From hype to measurable productivity
Kenya’s AI transition is therefore entering a more demanding phase.
For businesses, the ultimate test is increasingly straightforward: does AI improve the bottom line?
PwC’s research provides early evidence that commercial gains are already emerging. Globally, 23 per cent of CEOs whose companies invested in AI reported revenue increases, while 25 per cent reported cost reductions over the previous year.
But those gains are unevenly distributed.
The concentration of returns among leading companies suggests that simply adopting AI is unlikely to be enough. Businesses that achieve the greatest value are those that connect the technology to specific commercial problems and redesign operations around the opportunities it creates.
For Kenyan companies, the questions are becoming increasingly practical.
Does an AI customer-service system reduce response times and operating costs?
Does fraud detection reduce losses?
Does predictive technology reduce downtime?
Does route optimisation cut fuel consumption?
Does personalised marketing increase sales?
Does generative AI allow employees to spend more time on higher-value work?
The answers will increasingly determine whether AI remains another technology investment or becomes a genuine source of competitive advantage.
Kenyan businesses have already demonstrated a willingness to experiment. The next challenge is to ensure those experiments translate into measurable improvements in performance.
The AI race is no longer simply about being first to adopt the technology.
It is about turning adoption into productivity, productivity into value and value into measurable returns.
For Kenyan firms, the companies most likely to benefit may not be those with the most AI tools.
They will be those able to answer the harder question: What did AI actually change?
