Safaricom is connecting enterprise AI adoption with employee skills and local developers


For a telecommunications company serving millions of customers, artificial intelligence has applications across a wide range of operations. It can support customer service, help identify fraud, inform credit decisions and improve how employees handle routine tasks. Safaricom’s AI strategy is being developed across these areas, with the company also investing in the skills and external technology ecosystem needed to support its ambitions.

Its HY27 results describe AI as part of the company’s wider technology strategy. Safaricom says it is applying the technology to personalise customer propositions, simplify customer journeys, strengthen fraud prevention and credit decisioning, and improve operational efficiency. The company is also positioning AI as a foundation for its ambition to become an Africa-focused technology business.

The strategy has several distinct components. ZURI is the customer-facing generative AI assistant, PACE is intended to develop AI capabilities among employees, while Decode and Spark extend Safaricom’s involvement into developer skills and startup support. Their roles are different, but together they indicate that the company’s AI plans extend beyond introducing automated responses to customer enquiries.

AI is being integrated into more than one business function

Customer service is one of the most visible applications. Safaricom’s ZURI virtual assistant uses generative AI to extend self-service across multiple channels, according to the company’s HY27 results booklet. The aim is to give customers another way to interact with Safaricom and access assistance without relying entirely on conventional support channels.

Automated support can handle some routine interactions, allowing human agents to concentrate on cases that require additional attention. But the usefulness of a generative AI assistant depends on the accuracy of its responses, its ability to understand customer requests and the ease with which a user can reach a human representative when necessary.

Safaricom’s AI applications also extend to fraud prevention and credit decisioning. The company says it is using AI to strengthen these functions, while behavioural scoring and multiple funding partners form part of its approach to credit assessment and portfolio risk management. These applications involve different operational requirements from a customer-service chatbot, particularly where automated decisions can affect access to financial products.

The company has also described AI-supported customer verification in Safaricom Ethiopia, where the technology is being used to improve the speed and accuracy of onboarding and reduce the resources required for routine verification. The example shows that the company’s AI applications are not limited to customer support in Kenya, although the operating context and implementation differ across markets.

Employee capability is part of the technology investment

Deploying AI across business functions requires employees who understand how to use the technology and recognise where its limitations lie. Safaricom’s PACE programme is intended to build AI fluency across the organisation, forming part of its plan to develop an AI-focused workforce.

The need for internal capability goes beyond training employees to use new software. Staff involved in customer service, product development, risk management and operations need to understand how AI-generated outputs are produced, when they require review and how sensitive customer information should be handled.

The programme also reflects a wider challenge facing organisations adopting AI: technology procurement and deployment can move faster than the development of internal expertise. Without sufficient technical and operational understanding, companies risk relying on systems that employees cannot adequately assess or supervise.

Safaricom’s HY27 materials identify privacy, security, fairness, transparency and human oversight as safeguards for its AI applications. These principles are particularly relevant where systems interact with customer information or contribute to financial decisions.

PACE is therefore positioned as part of the company’s internal capacity-building effort. The available HY27 disclosures do not provide enrolment, completion or post-training performance figures, so the scale of the programme’s impact cannot yet be measured from the reported results.

Decode and Spark extend the strategy beyond Safaricom

Safaricom’s technology programmes also reach into Kenya’s developer and startup ecosystem. Decode is intended to provide students, developers and startups with practical skills and access to Safaricom platforms and expertise. The programme gives the company a route to engage with people developing products and services outside its own workforce.

Spark Fund addresses a different part of the ecosystem. Safaricom describes the initiative as supporting technology startups through product collaboration, access to investors and pathways to scale. The programme connects the company with businesses that may develop solutions relevant to its markets or wider technology ambitions.

These initiatives are not interchangeable. Decode focuses on skills, developer access and practical engagement, while Spark supports startups through collaboration and potential routes to investment and growth. Their inclusion in Safaricom’s technology strategy indicates an interest in building relationships with external innovators as well as developing internal systems.

The company’s position as a telecommunications and mobile money operator gives it a substantial technology platform, customer base and distribution network. For developers and startups, access to those resources can be useful when building products that need infrastructure, commercial partners or routes to customers. However, the available HY27 materials do not quantify the number of developers supported, startups funded, products launched or businesses that have reached commercial scale through these programmes.

Responsible AI will determine how far the strategy can go

Safaricom’s stated safeguards place governance alongside its operational ambitions. Privacy and security are central where AI systems process customer information, while fairness and transparency become important when automated tools influence decisions about credit, service access or customer treatment. Human oversight provides a mechanism for reviewing decisions and handling situations where automated systems may produce incomplete or inappropriate results.

For a company operating across telecommunications and financial services, these considerations apply to several different types of AI deployment. A virtual assistant responding to a billing query presents a different risk from a system contributing to credit assessment or fraud detection. Each requires controls suited to the consequences of an incorrect output.

Safaricom’s HY27 investor presentation describes AI as a foundation for its technology company ambitions, with customer experience, customer value and productivity among the areas expected to benefit. The company has also linked its strategy to a longer-term ambition of becoming Africa’s leading purpose-led technology company by 2030.

That ambition depends on whether AI can deliver measurable improvements across the business. For customer support, relevant measures include successful resolution, response quality and customer satisfaction. For internal adoption, employee capability and operational efficiency provide a more useful assessment than training participation alone. For Decode and Spark, developer outcomes, startup growth and commercially deployed solutions would show whether ecosystem support is producing lasting value.

Safaricom has established programmes across these areas, but they are at different points in the technology value chain. ZURI brings AI into customer interactions, PACE develops internal capability, and Decode and Spark connect the company to external talent and businesses. The commercial test is whether those separate efforts produce reliable services, better operating performance and technology products that can be sustained beyond the programmes themselves.

Go to TECHTRENDSKE.co.ke for more tech and business news from the African continent and across the world.

Follow us on WhatsApp, Telegram, Twitter, and Facebook, or subscribe to our weekly newsletter to ensure you don’t miss out on any future updates. Send tips to info@techtrendsmedia.co.ke

Facebook Comments

By George Kamau

I brunch on consumer tech. Send scoops to george@techtrendsmedia.co.ke
Back to top button
×