Every Banking App Needs Embedded Insurance
Banks and fintechs already own the two things insurance lacks in Africa: distribution and trust. Why embedded cover is the natural next layer of every app
1 September 2025
In a low-trust market, nothing sells insurance like paying a claim fast. Why claims speed — not marketing spend — is the real growth engine for insurtech
For most of my life as a car owner, I never made a claim. Not because nothing happened — things happened — but because the process was so tedious and so opaque that filing felt like more trouble than the loss itself. I suspect that quiet resignation describes most policyholders on this continent. They bought cover once, tried to use it, and learned not to bother.
That is the real state of insurance in much of Africa: not an absence of policies, but an absence of trust that the policy will pay. And it points to the single most important lever an insurtech has. Not the app. Not the premium. The claim.
Everything before a claim is a promise. The claim is where the promise is tested, and where the customer decides — for good — whether insurance is real to them.
The numbers around this are unforgiving. More than 30% of customers switch their insurer within a year of a poor claims experience.1 That is a staggering rate of churn attached to a single interaction. It means the claims process is not a back-office cost to be minimised. It is the product's most important feature, and the point at which most insurers quietly lose the customers their marketing worked so hard to win.
Flip that around and the opportunity becomes obvious. If a bad claim loses you a customer, a fast one wins you a reputation.
At Octamile, we set out to compress the claims cycle from weeks to hours — and, where the data allows an automated decision, to minutes. The point was never the technical achievement of paying a motor claim in under an hour. The point was what that speed does to a person.
When a payout lands before the customer has finished bracing for a fight, something changes. The abstract promise becomes a concrete fact. They tell people. And in a market where nobody believes insurance pays, one credible story of a fast payout travels further than any campaign. We have processed thousands of claims on this model,2 and the pattern holds: the fastest claims produce the loudest advocates.
This is why I think of speed as a distribution strategy, not an operations metric. A marketing budget rents attention. A fast claim earns a witness. In a low-trust category, the second compounds and the first does not.
Speed only works if the path to it is simple. A claims flow riddled with forms, documents, and waiting periods cannot be fast, because every step is another place to stall and another reason for the customer to give up — exactly as I once did.
So the discipline is to strip the process to its essentials: capture what you genuinely need to make a decision, automate the assessment wherever the risk allows, and pay. Fraud control matters, and technology lets you detect bad claims without punishing the honest majority with friction. But the default posture has to be pay quickly, not delay until disproven. Every day of unnecessary delay is a day you are teaching the customer that their earlier cynicism was correct.
Speed and simplicity are not two goals. They are the same goal seen from two sides.
It is tempting, as a technical founder, to fall in love with the machinery — the model, the pipeline, the dashboard. But the customer does not want any of that. They want the single emergency that could wipe out their savings to not wipe out their savings. That is the real-world problem, and the technology is only worth what it delivers against it.
This is the test I keep coming back to. Are we building something that makes a hard day survivable for someone who cannot absorb the loss? If the answer is yes, the speed and the simplicity follow naturally, because you are optimising for the person rather than the process. If the answer is no, no amount of engineering will fix it.
If you are building insurance in a market that does not trust it, the sequence is clearer than it looks. Win the claim, and you win the customer. Win enough claims fast enough, and the customers do your distribution for you.
It is slower to build than a marketing funnel and far harder to fake. But it is also the only version of insurance that earns the right to serve people who have been burned before — which, on this continent, is nearly everyone. The rest of the playbook, from embedding cover into everyday transactions to competing with the deeply held belief that faith is enough, only works once the claim is fast enough to be believed.
Speed is not a feature you add to the product. In this market, it is the product.
Banks and fintechs already own the two things insurance lacks in Africa: distribution and trust. Why embedded cover is the natural next layer of every app
1 September 2025
Insurtech in Africa isn't fighting user inertia — it's competing with faith. Why the real rival is the church, and what that means for how you sell cover
9 November 2025
On AI, insurance, building in Africa, and what I'm learning. No cadence promises. Only when I have something worth saying.