Latest / Fintech Conversations with Fexingo: Banking, Payments, and Financial Software Companies / How Fintechs Use Real-Time Data to Personalize Insurance Premiums
Transcript
- Lucas: Luna, I want to talk about a shift that is quietly reshaping one of the biggest industries in the world — insurance. But not the kind of shift you might expect. We're not talking about a new app that lets you file a claim faster. We're talking about how fintechs are using real-time data to personalize your insurance premium down to how you drive this morning. Luna: You mean like pay as you drive insurance? That's been around for a while though. Progressive's Snapshot program started over a decade ago. Lucas: Right, but that was the early version. It used a plug-in device that monitored your driving for a few months and then gave you a discount. What's happening now is far more granular. We're talking about premiums that adjust in real time, sometimes by the mile or even by the minute, using data from your phone, your car's sensors, or even your smart home devices. Lucas: I came across a pilot from a German fintech called wefox — they ran a test in 2025 with a usage-based auto insurance product that cut premiums by an average of 30 percent for low-risk drivers. But here's the kicker: the data was pulled directly from the car's onboard diagnostics port, not a separate dongle. No extra hardware. Luna: Huh. So they're piggybacking on data the car already collects — speed, braking, time of day, mileage. That feels both brilliant and a little creepy. Lucas: Exactly. And that tension — between personalization and privacy — is really the core of this episode. Because the more data you feed into the model, the more accurate the risk assessment gets, but the more surveillance-like it becomes. And it's not just auto insurance. We're seeing the same logic applied to health insurance, home insurance, even life insurance. Lucas: If today's topic is useful for something you're building or running, I'll just mention that the way these conversations stay ad-free and independent is listener support. It's a buy me a coffee model — buy me a coffee dot com slash fexingo. That's what keeps the lights on, no sponsorships. Appreciate anyone who finds value in that. Luna: Yeah, and honestly, we get to pick topics like this because of that. No advertiser telling us what to cover. Lucas: Alright — back to insurance. Let's talk about how the underwriting model actually changes when you move from historical data to real-time data. Traditionally, insurers set rates based on broad categories: age, gender, zip code, credit score, past claims. That's all backward-looking. With real-time data, you can assess risk as it happens. Luna: So the classic example is a driver who only takes the car out at 3 AM on a Sunday for a quick grocery run. Under a traditional model, they might get lumped in with other drivers in their age bracket. But with telematics, they'd get a very low premium because their actual driving behavior is low-risk. Lucas: Precisely. And that's good for the consumer, but it also means the insurer can price more competitively for the segments they want. The flip side is that high-risk behavior — like hard braking at 2 AM — could spike your premium instantly. Some policies now have a 'dynamic pricing' clause that allows the rate to change every month based on the previous month's data. Luna: Does that create a disincentive to drive safely? Or maybe it encourages it? I mean, if you know you're being monitored, you might drive more carefully. Lucas: That's the argument the industry makes — it's a feature, not a bug. Lemonade, for example, uses behavioral nudges in their pet insurance. They've got a device called Lemonade Pet that tracks your dog's activity and sleep patterns. If your dog is more active, they might suggest a higher activity level leads to lower health risk. It's a form of gamification. Lucas: But critics say this can lead to discrimination. If your phone's GPS shows you spend a lot of time in a high-crime neighborhood, could that affect your home insurance rate? Even if you've never filed a claim? That's the kind of question regulators are starting to ask. Luna: Right, because it's not just about what you do — it's about where you are, when you're there. That's a lot of data points. And the algorithms are black boxes in many cases. Lucas: Yeah, transparency is a huge issue. The wefox pilot I mentioned — they actually published their algorithm's decision rules. That's rare. Most fintechs treat their risk models as proprietary secrets. So the consumer has no way to know why their premium went up or down. Lucas: Let's talk numbers. The global usage-based insurance market was estimated at about $32 billion in 2025, and it's projected to grow at a compound annual growth rate of around 18 percent through 2030. That's faster than traditional insurance growth, which is typically single digits. Luna: And who are the big players beyond wefox and Lemonade? I know Progressive still dominates in the US with Snapshot, but are there any notable startups? Lucas: Sure. In the UK, there's a company called Zego that offers pay per mile insurance for gig economy drivers — delivery riders, Uber drivers. They use telematics from the driver's phone. In Asia, there's Singapore-based Igloo, which offers micro-insurance for things like smartphone screen damage using real-time device data. And in Latin America, there's a startup called 360insure that uses satellite imagery to adjust crop insurance based on weather data. Luna: So it's not just auto. It's really every line of insurance is being touched by this trend. Lucas: Exactly. And the technology enabler is often the same: cheap sensors, ubiquitous connectivity, and machine learning models that can process streaming data. The fintech angle is that many of these startups are using the same infrastructure that powers payments — APIs, cloud data lakes, real-time analytics — to build insurance products. Lucas: But there's a big regulatory hurdle. In the EU, GDPR places strict limits on how personal data can be used for automated decision-making. In the US, it's more fragmented, but some states like California have their own privacy laws. Insurers have to be careful that their real-time pricing doesn't run afoul of unfair discrimination statutes. Luna: I've read that some insurers are starting to offer 'data dividends' — they share a portion of the savings with customers who opt in to data sharing. That seems like a way to build trust. Lucas: Yeah, it's a clever model. Lemonade does something similar with their Giveback program — unclaimed premiums are donated to charities chosen by policyholders. That's not exactly a data dividend, but it's in the same spirit of aligning incentives. The question is whether consumers will actually feel they're getting a fair deal. Lucas: I want to bring it back to a specific case that I think illustrates the potential and the pitfalls. There's a fintech in France called Alan — they started as a health insurance challenger for small businesses. Recently, they launched a feature where your premium can decrease if you use their wellness app to track steps, sleep, and nutrition. Luna: And I bet the app is designed to be sticky — you log your data, you get a dopamine hit from the discount, and Alan gets a wealth of health data to refine their risk models. Lucas: Exactly. But here's the rub: the discount is capped at 10 percent. And the data is anonymized and aggregated, they say. But privacy advocates argue that you're effectively paying with your data. The discount is the price of giving up privacy. And that's a trade-off not everyone can afford to refuse, because if you don't opt in, you pay a higher premium. Luna: That's the 'privacy poverty' trap. Lower-income individuals may feel forced to share data to get affordable coverage, while wealthier individuals can choose to pay more to keep their data private. Lucas: Right. And that's where regulators are starting to focus. In New York, the Department of Financial Services issued guidance in late 2025 saying that usage-based insurance must offer a non-telematics alternative at a comparable rate. They want to prevent a two-tier system. Lucas: So the landscape is evolving. On one hand, the technology is incredibly powerful. On the other, the ethical and regulatory frameworks are still catching up. For fintechs building in this space, the winning approach might be the one that balances personalization with transparency and fairness. Luna: What do you see as the next frontier? I mean, we're already looking at real-time auto and health. What about life insurance? Could your premium change based on your daily exercise? Lucas: It's already happening. John Hancock in the US offers a program where you wear a fitness tracker, and if you meet activity goals, you get up to 15 percent off your premiums. But that's still periodic — they check your data monthly. The next step is continuous underwriting. Imagine your life insurance premium adjusting weekly based on your Apple Watch data. That's a few years out, but technically feasible. Lucas: And then there's the possibility of using alternative data altogether. Some startups are looking at social media activity or public records to assess risk. That's more controversial. But the direction is clear: the insurance industry is moving from static risk pools to dynamic, individualized pricing. Luna: It feels like the fintechs that win in this space will be the ones that earn consumer trust through transparency and real value. Not just lower premiums, but clear explanations of how data is used and how pricing works. Lucas: Absolutely. And I think that's the key takeaway for today. The technology is here. The data is flowing. But the human element — trust, fairness, choice — that's what will determine whether usage-based insurance becomes a win-win or a privacy nightmare. Luna: Well, it's a fascinating space. I'll be watching the regulatory developments closely. Lucas: Me too. And we'll keep an eye on the startups that are doing it right. Thanks for the conversation, Luna.