Sending money across Kenya used to mean navigating real physical risk. People relied on post office money orders, cross country buses, or cash stuffed into envelopes and jacket pockets, all of it exposed to loss or theft along the way.

The idea that would eventually replace that system didn't start in Nairobi. It began in Johannesburg in 2003, when Vodafone's Nick Hughes, drawn to the question of how technology could widen access to finance, secured backing from the UK's Financial Deepening Challenge Fund.

His proposal was for a mobile platform that let customers receive and repay loans through their handsets, built around Safaricom's existing airtime reseller network as its distribution channel.

Given how scarce smartphones were in Kenya at the time, the team settled on SMS and the SIM toolkit already sitting on every handset as the simplest way in.

That early framework got tested through a partnership with the microfinance lender Faulu Kenya, whose borrowers in Thika and Mathare began repaying loans through the new system.

Safaricom PLC Safaricom PLC

Pauline Vaughan, who went on to head M-PESA, recalled how quickly pilot users started bending the platform to their own purposes.

Traders would deposit cash before heading to Nairobi to buy stock, then withdraw it on arrival, using the system less as a repayment tool and more as a safe way to carry money.

Turning that insight into something nationwide meant Safaricom had to sink $10 million into adapting the platform for the broader Kenyan market.

The bigger obstacle wasn't technical, though. It was regulatory. A service that moved money, rather than airtime, inevitably pulled the Central Bank of Kenya into a process usually left to telecom regulators alone.

Resistance inside the CBK ran deep. One senior executive reportedly warned the platform could trigger a market collapse, a fear sharpened by the pyramid schemes then wiping out savings across the country.

Multinational banks operating in Kenya quietly campaigned against a telecom firm wading into territory they considered their own.

It was Governor Njuguna Ndung'u's decision to bring Safaricom and the banks into one room that changed the trajectory. He reframed M-PESA not as competition for the banks but as a technological fix for Kenya's financial inclusion problem.

"I think we have found a technological tool to solve our financial inclusion problem," said Ndung'u, then Governor of the Central Bank of Kenya.

That framing won buy in, but on one condition: customer funds would have to sit in a licensed bank, even as Safaricom's own platform handled the crediting and debiting of e-money.

Commercial Bank of Africa turned out to be the only institution willing to build the reconciliation systems the arrangement demanded. Because internet speeds in Kenya were still slow at the time, the M-PESA platform itself ended up hosted in Germany, while CBA's Nairobi data centre handled the daily settlement work.

When the service finally launched in March 2007, under Information Minister Mutahi Kagwe and Finance Minister Amos Kimunya, uptake outran even Safaricom's own expectations.

Active users climbed from around 19,700 a month after launch to more than 1,041,000 by November, after Michael Joseph scrapped the original target of 300,000 customers in favor of something far bigger.

None of that growth would have been possible without the dealer network Safaricom built almost by accident. The company initially approached Caltex assuming it owned its branded fuel stations outright, only to learn the stations were franchised to independent operators.

That discovery produced the aggregator model still running M-PESA's agent network today, where a few thousand head offices now oversee more than 250,000 agents spread across the country.