There is something quietly surprising about discovering one of Kenya’s most consequential technology upgrades not through a press release, a push notification, or a breathless WhatsApp forward from a well-connected contact, but through the decidedly mundane act of reaching for your phone to pay a restaurant bill. That is precisely how I encountered Safaricom’s new M-Pesa focussed ‘My OneApp’ this morning. 

One moment I was opening what I expected to be the familiar green interface of the M-PESA App I had used thousands of times before, and the next I was staring at a new icon, a new splash screen, and an entirely new onboarding sequence that made it immediately and rather dramatically clear that Safaricom had, overnight, fundamentally changed the way of millions of Kenyans interact with their most important financial services app. 

Having been more or less completely blindsided, what followed over the next hour or so was a lesson in how a strategically brilliant idea can be severely undermined by operationally poor execution — and why, given everything that M-PESA represents for Kenya and for Safaricom itself, that gap matters enormously.

M-PESA: Nineteen Years From ‘Send Money Home’ To Kenya’s Digital Financial Services Heartbeat

Before dissecting My OneApp (which I will refer to as OneApp from here on as that just sounds simpler and better) and what it got right and wrong, it is important to hold the full weight of what M-PESA actually is. Most Kenyans have grown so accustomed to its presence in their daily lives that the scale of the thing can become easy to take for granted. 

The story begins not in a Safaricom boardroom but in a modest development project funded by the UK’s Department for International Development, when Vodafone and Safaricom ran a pilot in Thika in the mid-2000s and discovered something that no product manager had anticipated: users were ignoring the intended purpose and instead using the system to send money to each other. 

That single behavioural insight — that Kenyans needed a fast, affordable, trustworthy way to move money between themselves in a country where sending cash to a rural relative meant entrusting it to a bus driver or making a long journey — gave birth, in March 2007, to one of the most consequential financial innovations of the past two decades.

The initial business plan projected 350,000 customers in year one. M-PESA hit 1.2 million. The simple genius of the model — basic SMS and USSD technology that worked on any phone, layered onto Safaricom’s existing network of trusted airtime resellers turned M-PESA agents, operating under a forward-thinking Central Bank regulatory framework — created a network effect that has compounded relentlessly ever since. 

Before M-PESA launched, just 19 per cent of Kenyan adults had access to formal financial services. By 2024, that figure had reached 84.8 per cent. It is not hyperbolic to say that M-PESA changed the social and economic architecture of Kenya more profoundly than almost any other technology intervention in the country’s history.

Fast-forward to April 2026 and the platform that started as “Send Money Home” now processes over 100 million transactions daily in Kenya, serves 40 million monthly active customers — a figure Safaricom announced formally last month, marking 19 years since launch — and handles KES 38.29 trillion in annual transaction value. 

It has evolved from a peer-to-peer transfer tool into a layered financial ecosystem encompassing overdraft credit through Fuliza, savings and loans via M-Shwari, micro-investment through the Ziidi Money Market Fund, stock trading via Ziidi Trader, merchant payments through Lipa na M-PESA and Pochi la Biashara, and a growing infrastructure of developer APIs through the Daraja platform. 

M-PESA now contributes over eight per cent to Kenya’s GDP and sustains more than a million jobs through its agent, merchant and developer ecosystem. When Safaricom’s HY26 investor report states that M-PESA generated KShs 88.1 billion in revenue in a single half-year period, growing 14 per cent year-on-year and representing 45.4 per cent of Kenya service revenue, it is confirming something that anyone who has watched this platform evolve has known for years: M-PESA is not a product that Safaricom sells. It is the very reason Safaricom exists in its current form.

The underlying infrastructure that makes all of this possible was itself rebuilt from the ground up in September 2025, when Safaricom executed its “Fintech 2.0” upgrade — a complete architectural overhaul that moved M-PESA to a cloud-native, microservices foundation and expanded transaction capacity from what was approaching 4,400 transactions per second at peak to a new baseline of 6,000 TPS, with a theoretical ceiling of 12,000 TPS as demand scales. 

The active-active architecture that underpins the new system — two engines running simultaneously rather than one primary and one cold backup — transformed M-PESA from a highly available service into a genuine always-on utility. I wrote extensively about why that September 2025 upgrade was a really big deal at the time, and I stand by every word of that analysis. 

The point is that when Safaricom launched the OneApp today, it was building the consumer-facing superstructure on top of a backend that had just been comprehensively modernized. The technical foundation, in other words, was the best it has ever been. The question was whether the product built on top of it would match that standard.

Born At Decode 4.0: The Summit Where Safaricom Declared Its AI-First Future

I should be candid about something at this point. I was not in the room last week when Safaricom officially unveiled OneApp. The announcement was made on day two of Safaricom’s Decode 4.0 summit in Nairobi — the company’s flagship developer and innovation conference, which this year placed AI and digital innovation squarely at the centre of its agenda. 

It was precisely the kind of event I would ordinarily have prioritized, and the kind of platform that Safaricom consistently uses to signal its most important strategic directions to developers, partners, investors, and the broader technology community. 

Last week, however, competing demands kept me away, and I did not have the opportunity to watch the livestream either. The result was that I walked into this morning entirely unaware that one of the most significant product launches in Safaricom’s recent history had gone live overnight.

That context matters for the review that follows. Decode 4.0 was not an arbitrary venue for this announcement. By choosing its developer summit — a gathering explicitly convened around AI and digital innovation — to unveil the OneApp, Safaricom was making a deliberate statement about what kind of product this is meant to be. 

This is not simply a repackaged M-PESA interface. It is Safaricom’s most visible declaration that it is serious about becoming an AI-first technology company, with the OneApp as the consumer-facing proof point of that ambition. The Decode 4.0 audience — developers who build on the Daraja API, partners who integrate with M-PESA, entrepreneurs who are constructing businesses on top of Safaricom’s platform rails — was exactly the right constituency to hear that message first. They are the people who will ultimately determine whether the OneApp’s mini-apps marketplace becomes a thriving ecosystem or a largely ignored grid of icons.

Understanding this launch in the context of Decode 4.0 also helps explain the gap I encountered this morning between the product’s aspirations and its current reality. Summit announcements, almost by definition, present products at their most polished and most conceptually compelling. The live demo, the keynote narrative, the developer excitement in the room — these are the conditions under which a product like OneApp looks genuinely transformative. 

The conditions under which I encountered it — a restaurant table, a bill to pay, no Safaricom data on my phone, and zero prior warning that anything had changed — are the conditions that reveal what the product actually is in the hands of an ordinary customer on an ordinary morning.

The Strategic Logic Of OneApp Is Beyond Dispute

For years, Safaricom operated two separate consumer-facing applications that served overlapping but distinct customer bases. The M-PESA App catered to 6.7 million active users conducting financial transactions. The MySafaricom App served 2.8 million users managing their connectivity needs — airtime, data bundles, customer service. 

The fact that these two products lived in separate silos was always an architectural anomaly for a company that has spent the better part of five years positioning itself not as a telecommunications operator but as Africa’s leading purpose-led technology company. Every time I had to switch between apps to check my data bundle and then make an M-PESA payment, it was a small but persistent reminder that Safaricom’s digital experience had not caught up with its strategic narrative.

OneApp is the attempt to close that gap decisively. The vision is genuinely compelling. A single, AI-powered platform that consolidates M-PESA balance, airtime balance, bills, bundles and financial services in one interface, with a marketplace of 80-plus integrated mini-applications covering everything from ShopZetu and Twiva for commerce, to Little for transport, Betika for sports betting, Baze TV for entertainment, and government services including the Hustler Fund. 

The integration of Zuri, Safaricom’s AI-powered chatbot, adds a conversational layer that can surface information and complete tasks without navigating through menus. Biometric login — fingerprint and face scan — is built in. An offline mode that allows essential transactions to proceed via SMS when no data connection is available is a genuinely thoughtful addition for a market where data access is not always guaranteed. 

A persistent QR code scan-to-pay button addresses a real friction point in retail environments. Spending tracking, in-app statement downloads, and bundle management built directly within the financial interface give the platform genuine depth.

The global template for what Safaricom is attempting here is well-established. WeChat in China started as a messaging app and methodically expanded until it became the digital operating system for hundreds of millions of users — hailing taxis, paying utility bills, ordering food, investing in funds, booking medical appointments, all without leaving a single interface. 

Grab and Gojek in Southeast Asia followed similar playbooks, expanding from ride-hailing into financial services, food delivery, and everyday commerce. The key principle that makes these platforms work is seamlessness: the user should feel they are in one coherent environment, not navigating between loosely connected mini-sites. Data flows across services to personalise the experience. 

Safaricom’s twist — and it is a meaningful one — is that it is building on top of an existing financial juggernaut with 40 million monthly active customers and a payments infrastructure embedded in the daily economic life of nearly every adult Kenyan. No other company on the continent starts a super-app from that position.

The declared target of reaching approximately 10 million OneApp users — consolidating the combined 9.5 million between both legacy apps — is, in that context, a conservative ambition. If Safaricom executes the proposition well, the ceiling is vastly higher.

What Actually Happened When I Tried To Use It

I want to describe this morning’s experience with clarity, because the details are not trivial. They speak directly to the gap between OneApp’s stated ambitions and its present reality.

Upon opening the app for the first time, I was presented not with a login screen or a simple account upgrade flow, but with a full re-registration sequence requiring me to verify my identity from scratch. That is the wrong first moment for any returning user. The implicit contract of an upgrade is continuity — you expect your history, your preferences, your saved data to migrate with you. This app asked me to start from zero without warning.

It then revealed something that made the re-registration considerably worse. The process would only work when connected to Safaricom’s own mobile data network. Wi-Fi was rejected. My Airtel data connection — on which I have run almost my entire mobile data consumption for well over a year, having switched to Airtel’s Smarta Bundles and genuinely never looked back — was also rejected. 

I was being required to activate a Safaricom app exclusively through Safaricom’s own data pipe. I had to purchase a Safaricom data bundle I had not budgeted for simply to gain access to a service I had been using for years. As someone who has publicly written about switching to Airtel for data and the cost implications of Safaricom’s pricing, there is a particular irony in being forced back to Safaricom data as the price of admission to an app nominally designed to serve me better.

Once inside the app, the next discovery was the most operationally damaging aspect of the entire launch. Every single saved favourite or frequents — till numbers for Lipa na M-PESA, paybill numbers for utilities, regularly used contacts for person-to-person transfers, carefully curated over years of active M-PESA usage — had been completely wiped. Not migrated to a temporary holding area. Not flagged as pending restoration. Simply gone. 

For a customer base of 6.7 million people who had been using the previous M-PESA App, the collective effort required to rebuild those saved favourites or frequents represents an enormous and entirely avoidable imposition on the people who have been M-PESA’s most digitally engaged users. Ironically too, this was delivered, it must be said again, without any advance communication from Safaricom. No email. No SMS alert. No push notification. No social media announcement that reached me in the days before the launch to indicate that a fundamental change to my most-used financial services app was imminent.

The practical consequence at the restaurant table was that I could not quickly locate the till number I needed through the OneApp. I fell back on the MySafaricom App, which was still functioning in its pre-upgrade form, and completed the payment through that. But the implications of that safety net deserve our full attention: if the MySafaricom App had already been decommissioned — as will eventually be the case — I would have been navigating the M-PESA SIM Toolkit to pay a restaurant bill in 2026. For a business whose latest investor disclosures describe the M-PESA Consumer App growing transaction volumes 56.1 per cent year-on-year to 624.2 million, that scenario would have been a brand moment of the worst kind.

As I explored further, the M-PESA Statements screen — one of the most practically important features for any active user — loaded for over a minute and then returned a “Connection timeout, try again later” message. The app crashed on me multiple times. Screen transitions exhibited significant latency, particularly in the minutes immediately following login. For a platform whose Fintech 2.0 infrastructure was specifically designed to be faster, more resilient, and more reliable than its predecessor, these are not minor launch-day hiccups. They are contradictions of the platform’s own stated promise.

The Deeper Problem Is Design Philosophy, Not Just Bugs

Crashes can be patched. Favourites or frequents could in theory be migrated in a subsequent update. Latency can be optimized. These are engineering problems with engineering solutions, and Safaricom has the talent to address them. The more fundamental concern that OneApp’s launch reveals is about design philosophy rather than implementation quality.

The app does not feel as though it was built through a genuine design thinking process — one that placed real customers and their real usage contexts at the centre of every decision, then iterated against feedback before public release. It feels as though it was architected from the inside out, from the perspective of Safaricom’s internal product, technical and commercial requirements, and then presented to customers as a finished product.

The forced re-registration without data migration, the Safaricom-data-only onboarding constraint, the total absence of advance communication — these are not signatures of a team that systematically stress-tested the user journey against the scenarios that actual customers live in every day. The irony is that Decode 4.0, the very summit at which OneApp was unveiled, is precisely the kind of forum where customer co-creation conversations should be happening. 

Developer summits exist not just to announce products but to gather the community’s most critical intelligence about what real users need. If that intelligence informed this launch, it was not reflected in the experience that arrived on customer phones the following week.

I have observed this pattern repeatedly in digital financial services propositions in Kenya over the years, and it stands in meaningful contrast to what the global benchmark apps consistently deliver. When a world-class fintech platform ships a major upgrade, users carry their data forward, receive advance notice of what is changing and why, and encounter a product that has been tested under realistic load conditions. 

The delta between that standard and what Safaricom delivered on launch day is not a technological gap — it is a customer empathy gap. For a company whose stated purpose is transforming the lives of every Kenyan, and whose brand has been built on the trust that 40 million people place in it with their money every single month, that gap is particularly painful. 

There is also the interface itself to consider. The design is functional and clean, using cards and grids in a sober Safaricom aesthetic. But it does not feel exceptional or world-class. The mini-apps marketplace creates genuine discovery overload — dozens of icons, promotional banners, and category tiles competing for attention in ways that overwhelm rather than guide. 

What makes WeChat or Grab genuinely compelling is not the breadth of what they offer but the intelligence with which they surface the right thing at the right moment. One pp has Favourites and Frequents sections that suggest the same ambition — but since the migration wiped all favourites or frequents clean, those sections are blank slates for every returning user on day one.

The Competitive Stakes Have Never Been Higher

In 2019, a rocky M-PESA app launch would have been a brand inconvenience. M-PESA held over 97 per cent of Kenya’s mobile money market and the competitive threat was largely theoretical. The data I analyzed in my recent telecoms sector statistics this past weekend tells a fundamentally different story for 2026. 

M-PESA’s mobile money market share has declined from 97.1 per cent in December 2023 to 89.0 per cent in December 2025, with Airtel Money climbing from 2.9 per cent to 11.0 per cent across the same period. That is an 8.1 percentage point swing involving millions of accounts — not a rounding error but a structural competitive shift, and one that is accelerating.

This context matters for the OneApp discussion in a specific way. Safaricom’s relationship with its customers has always carried a particular tension. The network is genuinely excellent. The quality of service, in most dimensions, is superior to alternatives. But Safaricom is also widely perceived as expensive to the point of being punitive. The popular Kenyan quip that “Safaricom eats your data bundles” — almost certainly a perception amplified by the speed of the network consuming data faster than slower alternatives, more than any literal over-consumption — has taken on a cultural life of its own. 

The result is a love-hate dynamic in which tens of millions of Kenyans remain with Safaricom primarily because M-PESA is structurally non-negotiable in their financial and social lives, not purely because they feel Safaricom has consistently earned their loyalty through competitive value delivery. When you hold your customers primarily through lock-in rather than love, the tolerance for product failures is substantially lower than you might assume.

This is why forcing users to purchase Safaricom data simply to onboard to OneApp is not merely a technical oversight. It is tone-deaf in a specific way that Safaricom’s own customer intelligence should have caught. Similarly, the absence of any advance communication about the migration — in a country where M-PESA is critical financial infrastructure for daily survival, not a social media app — reads as an assumption that customer patience is unlimited. 

In a market where Airtel Money is actively competing on price and steadily expanding its share, Safaricom cannot afford to test the limits of the lock-in thesis. A failed payment at a restaurant, a timed-out statements screen, a fresh onboarding sequence that begins by erasing years of user data — these are exactly the moments at which a customer who has been passively considering switching begins to actively do so.

What Safaricom Must Do, & What OneApp Can Become

The immediate priorities are clear, and Safaricom needs to move on them with the urgency it typically reserves for network incidents rather than software releases. The Safaricom-data-only onboarding requirement must be eliminated — there is almost certainly a technical rationale involving SIM authentication, but the customer experience it produces is unjustifiable at scale. 

A migration path for saved favourites or frequents must be built and deployed urgently, because asking 6.7 million users to manually reconstruct years of carefully accumulated payee data is an imposition that will generate sustained and entirely justified frustration. The performance issues — crashes, latency, the statements timeout — need emergency attention, not a scheduled sprint. 

Safaricom must also develop a communications culture around major product changes that treats customers as participants in the process rather than passive recipients of its outcomes. The Decode 4.0 summit was a remarkable platform for announcing OneApp to the developer and partner community. Therefore, an equally deliberate communication effort aimed at the millions of ordinary customers whose daily financial lives were about to change would have cost comparatively little and prevented an enormous amount of the friction that greeted this launch.

Beyond the immediate remediation, the longer-term trajectory of OneApp genuinely excites me — and I say that not as a concession to corporate politeness but as an analytical position grounded in what this platform is sitting on top of. The Fintech 2.0 infrastructure rebuilt in September 2025, the Daraja API ecosystem supporting over 100,000 developers, the 40 million monthly active M-PESA customer base, and the strategic clarity of the Vision 2030 roadmap give Safaricom assets that no competitor on the continent can replicate. 

The mini-apps marketplace, if curated with intelligence rather than just breadth, could become a meaningful commercial ecosystem in its own right. The offline mode, if it works as described, is an innovation that most global super-apps have not attempted and that speaks directly to Kenyan usage realities. The AI personalisation, if it develops beyond the current static home screen into genuinely contextual recommendations, could transform the app from a transaction tool into something closer to a financial companion. 

Also, the Decode 4.0 community — the developers and partners who were in that room last week — are the people best positioned to build the mini-app experiences that will determine whether OneApp becomes Kenya’s answer to WeChat or simply a more cluttered version of what came before.

The Intent From Safaricom Was 100% Right!

One unified, intelligent platform for Kenya’s most important financial service was always the correct destination. Safaricom has the technical infrastructure, the customer base, the developer ecosystem, and the financial resources to build something world-class. However, today’s launch demonstrated is that good infrastructure and good intentions are not sufficient on their own. 

They need to be matched by a product culture that treats the customer’s lived experience as the primary design constraint, not an afterthought — and a communications discipline that extends the energy of a Decode 4.0 keynote all the way to the forty million people who will never attend a developer summit but who trust Safaricom with their money every single day.

CEO Peter Ndegwa said something at M-PESA’s 40 million customer milestone that I have been thinking about today: “To us, every M-PESA transaction tells a story of someone building their future.” OneApp is supposed to give those stories more tools, more power, and more possibility. 

Right now, on day one, it is still writing its own opening chapter — one that contains genuine promise but also entirely too much friction for a platform that millions of people trust with the money that runs their lives. The next few weeks will reveal whether Safaricom treats this launch as the beginning of a serious iteration cycle or as a finished product. The distinction matters more than almost any other product decision this company will make in 2026.