Kenyan startup Fastagger is adapting AI to inexpensive phones and unreliable connections through Auni, a tool that reads mobile-money statements. New reporting describes how merchants use it to identify busy hours and returning customers. Some processing works offline, while detailed questions require connectivity; roughly 6,000 businesses have signed up.
Fastagger, a Kenyan software startup, develops Auni, a tool that turns mobile-money statements into information for small businesses. New interviews with founders and customers describe roughly 6,000 business signups. Launched in 2024, Auni now sits inside the merchant application offered by Kenya’s mobile-payment service M-Pesa. Its users include salons, restaurants and small manufacturers.[1]
For financial statements, the founders modified existing open-source text-recognition software to handle unusual layouts, stamps and errors. A small neural network helps identify document structure. Extracted information can be organised into tables on the phone, supporting inexpensive devices and unreliable connections. Reading statements and some analysis and visualisation work offline; more detailed analysis and questions require internet access.[1]
Peter Chege runs a salon in Nairobi and used to write down sales by hand. Auni shows him the busiest times and helps him spot formerly loyal customers who no longer come in. A fish restaurant uses the tool to identify loyal customers instead of reading payment messages individually. The founders observed merchants at work while developing the product and are now working on agents intended to carry out marketing actions.[1]
