Latest / Fintech Conversations with Fexingo: Banking, Payments, and Financial Software Companies / The Hidden Cost of Embedded Insurance Models
Transcript
- Lucas: If these conversations are useful for what you're building or running, I want to say thanks directly. Most listeners support us through loyalty, not ads, so if today felt like a win for your workflow, consider grabbing a coffee via buy me a coffee dot com slash fexingo. It keeps the lights on and the research sharp. Luna: Yeah, that’s fair. And speaking of workflows, have you seen the latest pushback from merchant groups against embedded insurance bundles? It feels like the easy button just got complicated. Lucas: It really has. For years, the pitch from insurtech firms was simple: frictionless coverage. You buy a flight, you get travel insurance. You buy a laptop, you get device protection. The margin for the platform was tiny, almost negligible, because they were just reselling existing products. But as we head into the third quarter of 2026, that model is hitting a wall of regulatory scrutiny and actual loss ratios. Luna: Loss ratios? You mean the payouts are eating into the margins more than expected? Lucas: Exactly. And it’s not just about higher payouts. It’s about who holds the bag when things go wrong. Traditionally, the carrier took the underwriting risk. The platform took the distribution fee. Now, some of the newer fintech players are trying to retain the underwriting risk too, using algorithmic models to price policies in milliseconds based on user behavior data. The problem is those models are still relatively untested over long economic cycles. Luna: So you’re saying they’re pricing risk faster than they can verify it? Lucas: Precisely. Let’s look at a specific example involving a major global payment processor that launched a bundled warranty product last year. They promised zero friction claims. Within six months, they saw a forty percent spike in fraudulent claims originating from high-volume small business sellers. Because the policy was embedded, the verification steps were minimal. The platform had to step in to cover the difference between the collected premiums and the actual losses. Luna: That sounds like a classic moral hazard problem. If the barrier to entry is low, bad actors flood in. Lucas: Right. And here is the kicker. The payment processor didn’t have an insurance license. They were operating in a gray zone of financial services regulation. By mid-summer, regulators in two major jurisdictions started asking hard questions about capital reserves. Essentially, they wanted to know: if this platform collapses, who pays back the policyholders? It shifted the narrative from innovation to systemic risk overnight. Luna: I hadn’t thought about the capital reserve angle. Isn’t that usually a bank thing? Lucas: It is. But when you embed finance deeply enough, the lines blur. This isn’t just about selling insurance anymore. It’s about balancing sheet management. The payment processor had to suddenly set aside millions in liquidity to prove solvency. That capital is now sitting idle instead of being deployed into growth initiatives or product development. It changes the entire unit economics of the service. Luna: So the cost of doing business goes up, which means the consumer probably pays more, despite the promise of seamless integration. Lucas: Almost certainly. In the short term, prices might stay flat to maintain adoption, but the structural cost is there. We’re seeing a trend where platforms are beginning to unbundled these services again. Not all of them, but the ones that tried to own the full stack are retreating to pure distribution roles. They’re partnering with licensed carriers again, accepting lower margins in exchange for offloading the regulatory burden. Luna: Is it really a retreat though, or just a course correction? The demand for embedded solutions hasn’t gone away. Lucas: The demand is definitely still there. Consumers love convenience. The issue is the opacity of the product. When you click one box to insure your purchase, do you actually read the terms? Do you know if the payout comes from the merchant, the platform, or a third-party insurer? In the recent dispute cases, many consumers didn’t know who to contact when a claim was denied. That lack of clarity is what triggered the regulatory heat. Luna: Transparency seems like the missing link here. If the risk transfer isn’t clear, the trust evaporates. Lucas: Agreed. And let’s talk about the data side. These algorithms rely on historical data to predict fraud. But if the platform is constantly changing its user base and transaction types, the data becomes noisy. A model trained on retail electronics sales doesn’t necessarily apply well to digital goods or subscription services. We’re seeing mispricing errors across several sectors right now. Luna: Mispricing in what direction? Are they charging too little or too much? Lucas: Both, depending on the segment. For low-risk users, they might be charging too much, which drives churn. For high-risk segments, they’re charging too little, leading to unsustainable loss ratios. It’s a balancing act that requires constant recalibration. Most platforms aren’t equipped with actuarial teams to do that work in real time. They’re relying on black-box AI tools that optimize for engagement, not accuracy. Luna: Optimizing for engagement over accuracy is a recipe for disaster in financial services. You don’t want a user experience bug to cause a solvency crisis. Lucas: Exactly. And this brings us back to the broader market environment in September 2026. Investors are getting pickier about fintech valuations. They’re looking for sustainable unit economics, not just gross merchandise volume growth. Companies that try to wear too many hats—processor, insurer, regulator—are facing valuation compressions. The market is rewarding specialization. Luna: Specialization makes sense. But doesn’t fragmentation hurt the consumer experience? If I have to jump between three different apps to manage my protections, isn’t that annoying? Lucas: It can be. But there’s a middle ground. We’re seeing the rise of aggregator models. Instead of embedding insurance directly into the checkout flow, platforms are offering a dashboard where users can manage all their policies from different providers. It’s slightly less seamless at the moment of purchase, but it provides much greater transparency and control later on. Users can see exactly who their insurer is, compare premiums, and file claims through a unified interface. Luna: A dashboard approach. That sounds like a compromise between convenience and accountability. Who are the early adopters of that model? Lucas: Several neobanks have started piloting this. One European player launched a financial health hub last month that includes insurance, savings, and investment products from various partners. They take a referral fee rather than underwriting the risk. It’s a cleaner balance sheet and clearer customer communication. Early metrics show higher retention rates compared to their old embedded models. Luna: Higher retention because customers feel safer? Or because they have better options? Lucas: Both. But primarily safety. When you own the risk, you own the reputation hit. When you refer the risk, you’re just a guide. Guides are easier to forgive than owners when things go wrong. It’s a subtle psychological shift, but it matters a lot in customer loyalty. Luna: I guess it also reduces the incentive to push bad products. If you’re just referring, you have to ensure the product is good to keep the customer coming back. Lucas: That’s a key point. In the embedded model, the incentive was often to maximize conversion, even if the product wasn’t the best fit. In the aggregator model, the incentive is to match the right product to the right user. It aligns the interests of the platform, the provider, and the consumer much better. Luna: So the era of blind trust in embedded finance might be ending. We’re moving toward a more transparent, albeit complex, landscape. Lucas: Complexity isn’t always bad. Sometimes complexity is just honesty. Financial products are inherently complex. Hiding that complexity behind a single checkbox creates false simplicity. Unbundling it reveals the true mechanics, which allows for better decision-making. Luna: Do you think traditional insurers will eventually dominate this space again? Or will fintechs find a way to hold the underwriting risk safely? Lucas: I think we’ll see hybrid models. Fintechs will bring the data and the distribution, but traditional insurers will provide the capital and the regulatory compliance. It’s not about one replacing the other. It’s about finding the right division of labor. The companies that figure out how to integrate seamlessly without taking on undue risk will win. Luna: Integration without risk exposure. That’s the holy grail for platform builders right now. Lucas: It is. And it requires a level of sophistication that most startups simply don’t have yet. They’re learning that financial services aren’t just software. They’re regulated assets with real-world consequences. Luna: A valuable lesson. Especially when the stakes involve people’s money and peace of mind. Lucas: Absolutely. As we wrap up, I’m curious where you see the next wave of innovation. Is it in new types of coverage, or in better ways to display existing risks? Luna: Probably the latter. Visualizing risk clearly before the purchase happens could change everything. Lucas: Visualization. Yes. If we can make the invisible visible, we might finally get the trust we’ve been chasing for the last decade.