Introduction

Artificial intelligence is often portrayed as a technology that requires high-speed internet, expensive hardware, and massive data centers. In Kenya, one startup is proving that AI can also thrive on basic mobile phones and spotty connections. Fastagger, founded by Mutembei Kariuki and two colleagues, has built a business assistant that works offline and draws insights from mobile-money records. By meeting small merchants where they already are—on their phones—the company is reshaping how AI serves entrepreneurs in emerging markets.

What Happened

Kariuki first considered the idea in 2019 while attending an artificial intelligence conference in Tokyo. He noticed that discussions about AI's future largely ignored Africa. Returning to Kenya, he teamed up with Jude Mwenda and Stephanie Njerenga to launch Fastagger. Their flagship product, Auni, debuted in 2024 and quickly gained traction after being integrated into M-Pesa, the country's dominant mobile-money platform. The platform has since attracted roughly 6,000 business users, ranging from barbershops and restaurants to small manufacturers.

The tool helps owners make sense of sales data they previously tracked by hand. Peter Chege, who runs a salon and barbershop in Nairobi, used to record every transaction on paper. With Auni, he can see which days and hours bring the most customers, spot loyal patrons, and send personalized messages to people who have stopped visiting. A fish restaurant in Nairobi uses the same system to identify regulars and offer automatic discounts, replacing a slow process of reviewing SMS receipts one by one.

Fastagger's engineers adapted an open-source optical character recognition model to read the often-messy PDF statements generated by mobile-money apps. Because many users have unreliable internet, critical steps such as extracting text and organizing it into tables happen directly on the device. More advanced analysis and question-answering still require a connection, but the core functions can run without one.

Why This Matters

Most AI products are designed for users with smartphones, laptops, and dependable broadband. Fastagger took the opposite approach, building for inexpensive handsets and intermittent networks. More than 80 percent of Kenyan adults use mobile money, making those apps a rich source of financial data. By turning that data into simple charts and recommendations, Auni gives small-business owners visibility into their operations without requiring them to become data analysts.

The model also raises broader questions about how AI should be distributed globally. Daniel Björkegren, an economist at Columbia University who studies technology in developing economies, warns against creating a rigid divide between lightweight AI for poorer regions and powerful AI for wealthier ones. He argues that users everywhere should have access to a wide range of tools, and that constraints in places like Kenya can produce efficiencies that benefit all markets.

Key Takeaways

  • Fastagger's Auni assistant runs offline on low-cost phones and integrates with M-Pesa.
  • About 6,000 Kenyan small businesses use the tool to analyze sales and customer behavior.
  • Features include identifying peak hours, tracking loyal customers, and automating marketing messages.
  • The startup adapted document-reading AI to handle low-quality mobile-money PDFs.
  • Fastagger remains a three-person team backed by accelerator programs from Google, Microsoft, and Nvidia.
  • The company also built an SMS-based workforce tool for security firms operating in remote areas.

Conclusion

Fastagger shows that AI does not have to follow the Silicon Valley playbook to be useful. By designing for the realities of basic phones, patchy internet, and paper-based record-keeping, the startup is giving Kenyan merchants practical tools to grow their businesses. As the team works on autonomous marketing agents and expands into workforce management, its approach may offer a blueprint for inclusive AI development far beyond Africa.