Forget traditional underwriting. Here’s how AI is changing the game.
The Application and the Algorithm
I used to apply for business insurance by filling out a 20-page form. The underwriter would take weeks to give me a quote. That’s changing. My new policy was underwritten by an AI. Instead of a long form, I just gave it permission to securely access my company’s accounting software and payroll data. The algorithm analyzed my real-time financial health and operational data. It saw that my cash flow was strong and my employee turnover was low. I got a better price in minutes, not weeks, because the AI saw the true, low-risk nature of my business.
Stop chasing the latest insurtech app. Chase a better customer experience instead.
The App and the Awful Experience
I downloaded the latest, slickest insurtech app to buy my renter’s insurance. The interface was beautiful. The process of getting a quote was fun. Then I had a claim. The beautiful app was useless. I was routed to a call center with long hold times, and the adjuster was unhelpful. The app was just a pretty wrapper on a terrible customer experience. I learned that the tech is irrelevant if the fundamental service is broken. I’ll take a company with an old-school website but a fast, fair claims process every single time.
The hidden truth about telematics that auto insurers won’t admit.
The Dongle and the Data They’re Really Collecting
Your auto insurer offers you a big discount if you just plug a small “telematics” dongle into your car. They say it’s just to track your mileage and braking habits. The hidden truth is that they are collecting a massive amount of data on you. They know where you go, what time you leave, which routes you take, and how often you visit a bar or a fast-food restaurant. They are building a detailed, and permanent, profile of your entire life, and they will absolutely use that data against you to raise your rates or deny a claim.
What nobody tells you about parametric insurance.
The Hurricane and the Instant Payout
My business was in a hurricane zone. A traditional insurance claim after a storm can take months of fighting with an adjuster. Then I learned about parametric insurance. What nobody tells you is that it’s not based on your actual damage. It’s based on a specific, measurable event. My policy stated that if a Category 3 hurricane, as measured by the National Weather Service, came within 20 miles of my business, the policy would automatically pay me a pre-agreed amount—say, $100,000—within 72 hours. No adjuster, no argument. Just a data trigger and an instant payout.
I spent 5 years at an insurtech startup. Here’s what I learned.
The Tech and the Terrible Unit Economics
I joined an insurtech startup because I believed technology could revolutionize the industry. I learned a hard lesson. While our technology was brilliant, our “unit economics” were terrible. The cost to acquire a customer was higher than the profit we could ever hope to make from them. We were so focused on “disruption” and growth that we ignored the fundamental, boring math of insurance. The company eventually failed, not because the tech was bad, but because you can’t defy the laws of financial gravity, no matter how good your app is.
Unpopular opinion: On-demand insurance is a gimmick.
The “On” and “Off” and the Uncovered Claim
On-demand insurance, which you can turn on and off with an app, sounds so flexible. Here’s the unpopular opinion: it’s a gimmick that creates massive coverage gaps. I used an on-demand policy for my freelance photography gear, turning it on only for the days I had a shoot. I did a shoot on Friday, turned the policy off, and my gear was stolen from my car on Sunday. The policy was off, so there was no coverage. A traditional, annual policy would have protected me. On-demand coverage is a recipe for being uninsured when you need it most.
90% of VCs don’t understand this about the insurance value chain.
The “Disruption” and the Duke’s Place
Venture capitalists love to fund insurtechs that promise to “disrupt” the industry. What 90% of them don’t understand is the complexity of the insurance value chain. They think they can just replace the agent with an app. They don’t understand the roles of the wholesale broker, the MGA, the underwriter, the reinsurer, and the claims adjuster. Insurance isn’t a simple product; it’s a long, complex chain of risk transfer. You can’t just “disrupt” one link without understanding how it affects the entire chain. That’s why so many well-funded insurtechs fail.
This simple integration of IoT devices transformed our property risk profile.
The Sensor and the Saved Basement
We installed a simple, inexpensive set of “Internet of Things” (IoT) sensors in our commercial building. We put water sensors near the water heaters and in the basement, and temperature sensors in our server room. One night, a water heater started a slow leak. The sensor immediately sent an alert to my phone. I was able to get to the building and shut off the water before any major damage occurred. That simple, connected device turned a potential $100,000 flood claim into a minor, $200 plumbing repair.
You’re not getting left behind because of technology. It’s because of your legacy mindset.
The Mindset and the Modern Market
I know many insurance agents who blame technology for their struggles. They think they are being left behind by some new app or algorithm. They are wrong. They are being left behind by their own “legacy mindset.” They are still trying to sell a commoditized product in a one-size-fits-all way. The agents who are thriving are the ones who have adopted a modern mindset. They see themselves as professional risk advisors, they use technology to be more efficient, and they provide a level of consultative service that no app can replicate.
Stop buying insurance from a chatbot. Buy from an expert instead.
The Bot and the Bad Advice
I tried to buy a business insurance policy through a website with an AI-powered chatbot. It was a frustrating and useless experience. The chatbot could answer simple questions, but it couldn’t understand the unique nuances of my business. It couldn’t give me strategic advice. It couldn’t negotiate with an underwriter on my behalf. A chatbot is a script. An expert human broker is an advisor, an advocate, and a strategist. For anything more complex than basic auto insurance, stop talking to the bot and find a real expert.
The uncomfortable truth about usage-based insurance data.
The Drive and the Digital Witness
Usage-based insurance, which tracks your driving through an app or a device, promises to reward you for being a safe driver. The uncomfortable truth is that you are creating a permanent, digital witness that will be used against you. In the event of an accident, the insurer’s first step will be to analyze your telematics data. They will know your exact speed, your exact location, and your braking pattern. They will use this data to try and prove you were partially at fault, even if the police report says otherwise.
Why everything you know about insurtech is backwards.
The Tech Is the Easy Part
Everyone thinks that the hard part of building an insurtech company is the technology. This is completely backwards. Building a slick app and a good user interface is the easy part. The brutally hard part is the “insurance” part. It’s the part where you have to understand complex regulations, build a profitable underwriting model, manage a claims process, and secure enough capital to pay for a catastrophic loss. The world is full of great insurtech apps that have failed because they learned, too late, that the tech was the least of their problems.
I tried to use a “peer-to-peer” insurance platform. It was a disaster.
The “Peer” and the Pool of Unpaid Claims
I joined a “peer-to-peer” insurance platform for my renters insurance. The idea was that we were all part of a “pool” and we would share in the profits if there were few claims. It was a disaster. When I had a legitimate theft claim, the “peer” review committee, who were just other customers, denied it. And when the platform had a bad year with lots of claims, they sent me a surprise bill for an extra “assessment” to cover the losses. It was all the risk of an insurer with none of the professionalism.
Hot take: Blockchain will not revolutionize insurance.
The Chain and the Unchanged Claim
For years, tech evangelists have been saying that blockchain will revolutionize the insurance industry. Hot take: it won’t. While it could be a useful tool for things like verifying policy information or streamlining payments, it doesn’t solve the fundamental, human problems of insurance. A blockchain can’t assess a complex liability claim. It can’t negotiate a settlement. It can’t provide empathetic customer service after a traumatic event. The core of insurance is human judgment and human interaction, and a distributed ledger doesn’t change that.
Most people waste hours with “AI-powered” quote tools. A good broker is faster.
The AI and the Agent
I wasted an entire afternoon using an “AI-powered” online tool to compare quotes for my business insurance. I had to answer hundreds of questions, and the results were confusing and unreliable. I gave up and called an independent broker who specialized in my industry. In a single, 15-minute phone call, he understood my needs, and he came back the next day with a detailed proposal with three competitive, real-world quotes. A good, experienced human who knows the market is still infinitely more efficient than a clunky AI.
The 5-minute habit that replaced my fear of data breaches from my insurer.
The Password and the Peace of Mind
I used to have a vague fear that my insurance company, which holds all my most sensitive personal information, would have a data breach. I replaced that fear with a simple 5-minute habit. I went into my online account with my insurer and enabled two-factor authentication (2FA). This means that even if a hacker steals my password, they can’t get into my account without the second code from my phone. It is the single most effective action you can take to protect your own data, and it takes less than five minutes to set up.
Your insurtech app isn’t crashing because of a bug. It’s this.
The Crash and the Crumbling Foundation
You’re using an insurtech app, and it’s constantly crashing or running slowly. You think it’s just a bug in the app. It’s not. The app itself is probably fine. The problem is that the beautiful, modern app is just a thin veneer on top of a crumbling, 40-year-old legacy mainframe system that the insurance company is still using. The app is crashing because it’s trying to talk to an ancient piece of technology that it can barely communicate with. The problem isn’t the app; it’s the foundation.
If you’re not using embedded insurance in your platform, you’re already losing to your competitors.
The Booking and the Built-in Policy
I run a travel booking website. For years, we treated travel insurance as an afterthought. We were losing. Our competitor, on the other hand, integrated or “embedded” insurance directly into their booking process. With a single click, a customer could add a tailored travel insurance policy to their purchase. It was seamless and convenient. Their conversion rates on insurance were ten times higher than ours. I learned that if you are not making the insurance purchase an invisible, effortless part of your core customer journey, you are already being left behind.
Stop glorifying disruption. Start solving real customer problems.
The Disruption and the Disappointment
The insurtech world is obsessed with “disruption.” Every startup wants to be the “Uber of insurance.” They are focused on tearing down the old model. This is the wrong goal. Customers don’t care about disruption. They care about their problems being solved. Is my claim being paid fairly and quickly? Is my agent giving me good advice? Is the process easy to understand? The most successful companies are not the ones with the most disruptive technology, but the ones that are obsessively focused on solving the real, human problems that have plagued the industry for a hundred years.
The real cost of a “free” smart home device from your insurer that nobody calculates.
The Device and the Data Dividend
Your insurer offers you a “free” smart home device, like a water leak sensor, if you agree to share the data with them. It sounds like a great deal. The real cost is hidden. You are giving the insurance company a permanent, 24/7 window into your home. They are collecting a massive amount of data on your habits and your property. And you can be sure that they will use that data to their advantage, whether it’s by raising your rates after a minor issue or using the data against you in a claim. The “free” device is just the price of admission for them.
What forward-thinking insurers do with AI that laggards don’t.
The AI and the Actual Insight
A laggard insurance company will use AI as a simple automation tool—to process forms faster or to handle simple customer service chats. A forward-thinking insurer does something much more powerful. They use AI to gain a deeper, more holistic understanding of risk. They use it to analyze satellite imagery after a catastrophe to speed up claims payments. They use it to analyze legal documents to identify emerging liability trends. They don’t just use AI to be more efficient; they use it to be smarter.
The myth of the “AI adjuster” is destroying customer trust.
The Algorithm and the Absence of Empathy
The insurance industry is promoting the myth of the “AI adjuster”—an algorithm that can instantly and fairly settle your claim with a photo from your phone. This is destroying customer trust. An AI cannot understand the emotional trauma of a house fire. It cannot show empathy for an injury. It cannot negotiate a complex settlement with a contractor. A claim is a human event, not a data problem. The attempt to replace human judgment and empathy with a cold, efficient algorithm is a dangerous path that will only lead to more disputes and more distrust.
I quit my job at a legacy carrier to join an insurtech and my view of the industry was changed forever.
The Old Guard and the New Game
I spent the first decade of my career at a huge, legacy insurance company. We moved slowly, we were bureaucratic, and our technology was ancient. I quit and joined an insurtech startup. It was a culture shock that changed my entire view of the industry. The speed, the customer focus, and the willingness to challenge old assumptions was intoxicating. I realized that while the legacy carriers have the money, the insurtechs have the mindset. The future of insurance will be built by those who can combine the stability of the old guard with the agility of the new.
Controversial: The traditional agent is holding back innovation.
The Agent and the Anchor
Here’s a controversial take: the traditional, commission-based insurance agent is the single biggest anchor holding back innovation in the industry. Why? Because the entire system is built around them. New, innovative insurance products that might be better for the consumer often don’t get adopted because they don’t fit into the existing commission structure. The agent has no incentive to sell a product that pays them less. The industry’s reliance on this century-old distribution model is a massive barrier to true, customer-centric innovation.
95% of headlines about insurtech are wrong. Here’s why.
The Headline and the Hype
The tech and business media love to write breathless headlines about how insurtech is “disrupting” the insurance industry. 95% of these headlines are wrong. They are written by reporters who don’t understand the deep, complex, and highly regulated nature of insurance. They mistake a slick user interface for a revolutionary new business model. They glorify the fundraising announcements without ever following up to see if the company has a viable path to profitability. The hype is about the “tech”; the reality is about the “insurance,” and the media almost always misses the real story.
One small API integration eliminated a major friction point in our insurance process.
The API and the Automatic Approval
Getting proof of insurance from our contractors used to be a nightmare of phone calls and faxes. It was a major friction point. We made one small change. We integrated our management software with an API (Application Programming Interface) from a leading insurance verification service. Now, when a contractor enters their information, the API instantly and automatically verifies their coverage in real-time. What used to be a week-long, manual process is now an invisible, five-second check. That one small integration has saved us hundreds of hours a year.
The truth about predictive analytics that data scientists profit from hiding.
The Prediction and the Imperfect Past
Data scientists will sell you on the power of “predictive analytics” to foresee your future risks. The truth they profit from hiding is that these models are only as good as the data they are trained on. And all historical data is inherently biased and incomplete. A predictive model can only predict a future that looks like the past. It cannot predict a true “black swan” event—a risk that is novel and has never happened before. Predictive analytics is a powerful tool, but it is not a crystal ball.
Stop trying to replace humans with AI in claims. It’s killing customer satisfaction.
The Bot and the Broken Trust
My insurance company tried to “improve” their claims process by replacing their human adjusters with an AI-powered chatbot for small claims. It was a disaster that killed my trust in them. The bot couldn’t understand the nuances of my situation, it had no empathy for my stress, and I was stuck in a loop of automated responses. A claim is a “moment of truth” for an insurer. By trying to replace a human connection with an efficient but soulless bot, they are optimizing for cost savings at the expense of the one thing that truly matters: customer loyalty.
Replace your clunky annual renewal with a real-time data feed. Thank me later.
The Renewal and the Real-Time Rate
The annual insurance renewal process is a clunky, outdated ritual. It’s based on a snapshot of your business from a year ago. The future of insurance is a real-time data feed. Imagine a policy where your premium is adjusted dynamically, in real-time, based on your actual business activity. Your commercial auto premium could be based on the real-time telematics from your fleet. Your liability premium could be based on your real-time sales data. This move from an annual review to a continuous, data-driven relationship will transform the industry. You’re welcome.
The insurtech secret that could give you access to data-driven pricing.
The Data and the Discount
Here’s a secret that can get you a better price on your business insurance. Many modern, data-driven insurtech carriers have the ability to connect directly to your company’s existing software via API. By giving your insurer secure, read-only access to your accounting software, your fleet management system, or your payroll provider, you are giving them a real-time, transparent view into your operations. This high-quality data allows them to see that you are a well-run, low-risk business, and they will often reward that transparency with a significant premium discount.
Why your traditional insurance model fails in the digital-first economy.
The Physical Policy for a Digital World
The traditional insurance model was built to protect physical assets in a physical world—a factory from a fire, a car from a crash. This model is completely failing in the digital-first economy. Today, a company’s most valuable asset is often not its building, but its data, its brand reputation, and its intellectual property. Its biggest risk is not a fire, but a data breach or a viral social media scandal. The traditional insurance policy, with its focus on tangible property, is the wrong tool for protecting a modern, intangible business.
I ignored the trend of parametric insurance for years. It cost me a faster payout.
The Storm and the Slow Settlement
I owned a business in a hurricane-prone area. I had a traditional property insurance policy. After a major hurricane, it took me over six months to settle my claim with the adjuster. I ignored the trend of parametric insurance. My competitor across the street had a parametric policy. It paid out a pre-agreed amount based on the storm’s wind speed. He had his full insurance payment in his bank account within 72 hours of the storm, while I was still waiting for my first phone call from the adjuster.
Let’s be honest: Most insurtech is just a prettier front-end for old insurance products.
The App and the Ancient Policy Form
Let’s be honest with ourselves. You download a beautiful, slick new insurtech app. The user experience is amazing. You buy a policy in 60 seconds. But if you were to read the actual policy document, you would discover that it is the exact same, standard, 30-year-old insurance policy form that a traditional agent would have sold you. Most insurtech is not a revolution in the insurance product itself. It is a revolution in the sales and marketing process. It’s just a prettier front-end for the same old engine.
87% of insurers get their digital transformation wrong. Don’t be one of them.
The “Transformation” and the Lack of True Change
Most insurance companies think that a “digital transformation” means building a new app or updating their website. This is why they are getting it wrong. A real transformation is not about the technology; it’s about the culture and the business model. It’s about using data to make smarter underwriting decisions. It’s about creating a seamless, empathetic customer experience. It’s about empowering your employees with better tools. A pretty app on top of a broken, bureaucratic process is not a transformation; it’s just a new coat of paint on an old, crumbling building.
This weird habit of questioning the algorithm outperforms trusting the tech every time.
The Algorithm and the Absurdity
The “AI-powered” underwriting algorithm quoted me a ridiculously high premium for my auto insurance. Instead of just accepting it, I started questioning it. I asked my broker to find out why the algorithm had priced me that way. We discovered that because I had once lived in a high-risk zip code for a few months ten years ago, the algorithm had permanently flagged me as a high-risk driver. By questioning the absurd output, we were able to get a human underwriter to override it. Never blindly trust the tech.
The real reason your insurtech startup isn’t getting traction (hint: it’s not the tech).
The Tech and the Trust Gap
Your insurtech startup has amazing technology, a brilliant app, and a great user experience. But you’re not getting any real traction. The reason isn’t your tech. The reason is that insurance is not a tech product. It is a promise. It is a promise to be there for a customer in their absolute worst moment. And that promise requires an immense amount of trust. Your startup has not yet earned that trust. You are not just competing on technology; you are competing against companies that have spent a hundred years building a reputation for keeping their promises.
Ditch your legacy policy admin system. Use a modern, API-first platform instead.
The Mainframe and the Modern Market
Your insurance company is still running on a 30-year-old legacy mainframe policy administration system. It’s slow, it’s inflexible, and it can’t talk to modern software. You are trying to compete in a modern, digital market while handcuffed to a piece of ancient technology. You need to ditch it. A modern, cloud-based, “API-first” platform will allow you to innovate faster, integrate with new partners seamlessly, and provide a much better customer experience. You cannot build the future of insurance on a foundation from the past.
Stop pretending your app is the solution. Start focusing on the underlying insurance product.
The App and the Inadequate Policy
I see so many insurtech startups that are obsessed with their app. They have a beautiful user interface and a fun onboarding process. They are so focused on the tech experience that they completely neglect the actual insurance product they are selling. The policy itself is often a cheap, off-the-shelf product that is full of holes and exclusions. Stop pretending your app is the solution. The app is just the delivery mechanism. The real solution is a high-quality, fairly priced insurance policy that actually protects the customer.
The 5-word phrase that changed how I think about the future of insurance.
From repair and replace to predict and prevent.
I used to think that the business of insurance was to “repair and replace” things after they broke. A futurist said a 5-word phrase that changed my entire perspective. The future of insurance, he said, is to “predict and prevent.” By using data from IoT devices, telematics, and other sources, the industry can move from a reactive model of paying claims to a proactive model of helping customers prevent the loss from ever happening in the first place. That is the true, revolutionary promise of technology in insurance.
What the tech industry doesn’t want you to know about the difficulties of the insurance business.
The Code and the Catastrophe
The tech industry looks at the insurance world and sees a slow, old-fashioned industry that is ripe for disruption. What they don’t want you to know is how brutally difficult the insurance business actually is. It’s not like building a social media app. It’s a highly regulated, capital-intensive business where you are exposed to catastrophic, unpredictable risk. You can have a perfect algorithm and a beautiful app, but one major hurricane or a change in the legal environment can bankrupt you overnight. The “move fast and break things” ethos does not work when you are the one who has to pay for the broken things.
I was today years old when I learned about micro-insurance.
The Policy for a Single Day
I was today years old when I learned about “micro-insurance.” I was going on a ski trip and a travel app offered me the chance to buy an insurance policy that would cover me for just that one day on the slopes. It was a small premium for a very specific, short-term risk. This is the idea of micro-insurance: breaking down the traditional, annual policy into tiny, affordable, and highly relevant pieces of coverage that can be embedded into other products and services. It’s a powerful way to make insurance accessible to more people, for the specific moments they actually need it.
Normalize demanding data portability from your insurance provider.
My Data, My Property
For years, my auto insurance company collected a massive amount of my personal driving data through their telematics program. When I wanted to switch to a new insurer, I realized that all that data—my data—was trapped in their system. I had no right to take it with me. We need to normalize demanding “data portability” from our insurers. My driving history is my property. I should have the right to take that data to any other insurance company to get a competitive, data-driven quote. It’s my data, and I should have control over it.
Plot twist: The biggest disruptor isn’t an insurtech. It’s the customer.
The Customer and the Clamor for Change
Everyone in the insurance industry is terrified of being “disrupted” by a slick new insurtech startup from Silicon Valley. The plot twist is that the insurtechs aren’t the real threat. The biggest disruptor in the insurance industry is the modern customer. Today’s customer is digitally savvy, they expect a seamless and transparent experience, and they have no tolerance for bureaucratic, slow, and confusing processes. It is the rising expectations of the customer, not the technology of a startup, that is the true, irresistible force that is changing the insurance industry forever.
The API everyone ignores that gives me an edge in product development.
The Integration and the Instant Insurance
I run a software platform for photographers. I wanted to offer them equipment insurance directly through our app. The API that everyone ignores, but that gave me a huge edge, was the one from a specialized insurance carrier. By integrating their API into our platform, we were able to offer our users the ability to get an instant, bindable quote for their gear without ever leaving our site. It created a massive new revenue stream for us and provided a valuable service for our customers. Partnering with an insurer via API is a powerful growth strategy.
Stop optimizing for user acquisition. Optimize for lifetime value.
The Acquisition and the Awful Retention
Our insurtech startup was obsessed with user acquisition. We spent a fortune on digital ads to get new customers in the door. Our growth numbers looked great. The problem was, our customer retention was awful. Our churn rate was huge. We were a leaky bucket. We learned to stop optimizing for acquisition and to start optimizing for “lifetime value” (LTV). By focusing on great service and a fair claims process, we started to keep the customers we had. A loyal, long-term customer is infinitely more valuable than a cheap, one-time acquisition.
The brutal truth about why your AI underwriting model isn’t working.
The Algorithm and the Unseen Bias
Your AI underwriting model is not working because of a brutal truth you don’t want to admit. The historical data you used to train your algorithm is full of the same human biases that have existed for decades. The model learned from that data that certain zip codes or certain demographic groups were “higher risk.” Your AI is not an objective genius; it is just a very fast and efficient reflection of your own company’s past biases. Without a constant, rigorous effort to de-bias your data and your models, your AI will just perpetuate the same old unfair practices.
Throw away your old data sources. It’s making you worse at risk selection.
The Old Data and the Outdated Decision
Your underwriting team is still making decisions based on old, traditional data sources—credit scores, motor vehicle reports, and basic property information. Throw them away. In a world of real-time, dynamic data, you are making outdated decisions. New data sources—from IoT devices, telematics, satellite imagery, and social media sentiment—can provide a much richer and more accurate picture of risk. If you are still relying only on the data sources of the past, you are making yourself worse at selecting the good risks and avoiding the bad ones.
The 60-second test that reveals if an insurtech is truly innovative or just hype.
The Question About the Contract
To find out if an insurtech startup is truly innovative or just hype, perform this 60-second test. Ask the founder this one question: “Can you tell me how your company has innovated on the actual insurance policy contract itself?” If they start talking about their app, their user experience, or their AI chatbot, they are just hype. They are a tech company. If they can explain how they have created a new type of coverage, a broader definition, or a fairer set of exclusions, then they might be a true insurance innovator.
Why everyone is wrong about the death of the insurance agent.
The Agent and the Advisor
For years, the tech world has been predicting the “death of the insurance agent.” Everyone is wrong. Yes, the role of the simple, order-taking agent who just sells a commoditized auto policy is dying. But the role of the professional, specialized risk advisor is more important than ever. In a world of increasing complexity, cyber risks, and global threats, clients need a human expert—a true consultant—to help them navigate their risks. The transactional agent is dead. The expert advisor will be around forever.
Stop asking “what’s your tech stack?”. Ask “what’s your loss ratio?” instead.
The Stack and the Sanity of a Sound Business
In the insurtech world, venture capitalists love to ask founders, “What’s your tech stack?” It’s the wrong question. A beautiful tech stack is irrelevant if the underlying insurance business is unprofitable. The much more important question is, “What is your loss ratio?” The loss ratio—the percentage of premium that is paid out in claims—is the true measure of whether you have a sustainable insurance business. A good loss ratio is a sign of a healthy company. A good tech stack is just a sign of a good engineering team.
The habit of A/B testing everything that I wish I’d started on day one of my insurtech.
The Test and the Truthful Data
When we launched our insurtech, we made decisions based on our gut feelings and what we thought customers wanted. We were wrong about almost everything. I wish I had started the habit of A/B testing from day one. We now test everything: the language on our website, the design of our app, our pricing, and our email subject lines. The data from these constant, small experiments has been our most valuable teacher. It has replaced our flawed assumptions with the truthful data of what our customers actually want and how they actually behave.
Here’s why Silicon Valley’s “move fast and break things” is terrible for insurance.
The Break and the Bankrupting Promise
The famous Silicon Valley motto is “move fast and break things.” This is a completely, dangerously inappropriate philosophy for the insurance industry. When a social media app “breaks things,” the result is a bug or a bad user experience. When an insurance company “breaks things,” the result is a broken promise to a customer who has just lost their home. Insurance is a business built on stability, trust, and the ability to pay claims over a long period of time. It is the one industry where moving slowly and carefully is actually a feature, not a bug.
I’ll say what everyone’s thinking: Your insurtech is going to run out of cash.
The Burn and the Barren Bank Account
Let’s just say what every venture capitalist is thinking when they look at your insurtech’s financial statements. Your burn rate is too high. You are spending a fortune on customer acquisition and your loss ratio is terrible. You are focused on growth at all costs, not on building a sustainable, profitable underwriting business. You are on a collision course with reality. Unless you can dramatically improve your fundamentals, you are going to run out of cash, and the next funding round is not going to come.
The skill of regulatory navigation that matters more than your coding ability.
The Regulation and the Reality
As an insurtech founder, I thought my most valuable skill was my ability to code. I was wrong. The skill that has proven to be far more important is the skill of regulatory navigation. Insurance is one of the most heavily regulated industries in the world. The ability to understand the complex web of state and federal laws, to build a good relationship with the regulators, and to design a product that is compliant from day one is the key to survival. Your beautiful code is worthless if the Department of Insurance shuts you down.
This counterintuitive action of partnering with a legacy carrier fixed our distribution problem.
The Partnership and the Path to Customers
As an insurtech startup, we were trying to disrupt the legacy insurance carriers. We saw them as our competition. We were struggling to get our product to customers. We made a counterintuitive decision. We stopped competing and started partnering. We entered into an agreement with a large, legacy carrier to sell our innovative product through their massive network of independent agents. They had the distribution that we lacked. It felt like a compromise, but that partnership gave us instant access to thousands of customers and solved our biggest problem.
Why your good intention of “disrupting” the industry is actually alienating customers.
The “Disruption” and the Distrust
Your insurtech is on a mission to “disrupt” the insurance industry. Your marketing is aggressive and full of attacks on the “old way” of doing things. Your good intention might be alienating the very customers you are trying to attract. Insurance is a product that people buy to feel secure. A message of “disruption,” chaos, and tearing things down can create a feeling of instability and distrust. Customers don’t want their insurance company to be a disruptive force; they want it to be a stable, reliable partner.
Quit trying to be a tech company that sells insurance. Be an insurance company that uses tech.
The Tech vs. The True Identity
So many insurtech startups make a fundamental mistake. They see themselves as a tech company that just happens to sell insurance. This is backwards. You need to be an insurance company first. Your core competencies must be underwriting, claims management, and regulatory compliance. Technology is a powerful tool to make you better at those core functions, but it is not the business itself. The most successful insurtechs are the ones that are run by insurance professionals who know how to use technology, not by tech people who are trying to learn insurance.
The metric everyone tracks (number of downloads) that means absolutely nothing without policyholder retention.
The Download and the Disappearing Customer
Our insurtech was so proud of our app download numbers. We would present them to our investors as proof of our success. It was a vanity metric. It meant absolutely nothing because our policyholder retention rate was terrible. Customers would download the app, buy a policy, and then cancel it a few months later. We were a leaky bucket. The only metric that truly matters is your ability to retain your customers over the long term. A million downloads are worthless if you have a thousand angry ex-customers.
Stop calling it “AI.” Call it “a statistical model with potential biases.”
The AI and the Actual Reality
The word “AI” is a sexy, futuristic marketing term. It’s also often a lie. I’ve learned to be more precise and honest with my language. We don’t have a magical, thinking “AI.” We have “a statistical model that has been trained on historical data and may contain potential biases.” It’s not as catchy, but it’s a much more accurate and responsible description of the technology we are actually using. It reminds us that our models are fallible and require constant human oversight.
The decision I made to focus on a niche market that everyone said was too small (but worked).
The Niche and the Narrow Path to Success
When we launched our insurtech, we tried to be everything to everyone. We were competing with the giants and getting crushed. We made a decision that everyone said was crazy. We decided to focus on one single, tiny niche market: insurance for freelance graphic designers. By focusing all of our energy on this one, underserved group, we were able to build a product and an experience that was perfect for them. We became the dominant player in that small pond. That narrow focus was our path to profitability and success.
What I learned from a failed insurtech pilot that changed our entire strategy.
The Pilot and the Pivot
We ran a pilot program for our new insurance product. We thought it was a brilliant idea. The results from the pilot were a complete failure. Nobody bought it. Our first instinct was to blame the marketing or the app. But when we honestly looked at the data and talked to the pilot users, we learned a hard lesson: the underlying product itself was not solving a real problem for them. That failed pilot was the best thing that could have happened to us. It forced us to pivot our entire strategy and build something that customers actually wanted.
The common mistake of ignoring the claims experience that’s killing insurtechs.
The Claim and the Churn
So many insurtech startups are obsessed with the “front-end” of the business—the quoting, the buying, the app. They completely ignore the “back-end”—the claims experience. This is a fatal mistake. A customer only truly interacts with their insurance company when they have a claim. That is the moment of truth. If that experience is slow, difficult, or unfair, that customer will leave and will tell all their friends. A bad claims experience is the single biggest driver of customer churn, and it is the rock upon which many promising insurtechs have been broken.
PSA: On-demand insurance is a regulatory nightmare. Here’s proof.
The “On” and “Off” and the Office of the Commissioner
Here’s a public service announcement for any insurtech thinking about offering “on-demand” insurance. The regulatory framework for insurance is built on an annual policy model. When you start turning coverage on and off by the hour or by the day, you enter a regulatory nightmare. How do you handle cancellations? How do you calculate your earned premium? How do you comply with state-specific notification laws? I’ve seen multiple on-demand startups get shut down by their state’s Department of Insurance because their innovative model was not compliant with a century of insurance law.
The skill of empathy that tech founders should teach but don’t.
The Code and the Customer’s Crisis
Tech founders are brilliant at building products. The skill they often lack, and the one they should be teaching, is empathy. They can’t put themselves in the shoes of a customer who has just had a car accident or a house fire. They see a claim as a data point to be processed efficiently. They don’t see the human trauma behind it. The most successful insurance companies, tech-focused or not, are the ones that can combine technological efficiency with genuine, human empathy during a customer’s moment of crisis.
This 5-minute action of reading customer reviews beats relying on your NPS score every time.
The Score and the Story
Our insurtech used to be obsessed with our Net Promoter Score (NPS). We thought it was the best measure of our customer satisfaction. It’s not. It’s just a number. I started a 5-minute action that is far more valuable. Every day, I spend five minutes reading the raw, unfiltered customer reviews of our app and our service. The stories, the complaints, and the praise in those reviews give me a much richer and more actionable insight into the real customer experience than any single, quantitative score ever could.
Why that heavily-funded insurtech is actually doing it wrong for the end customer.
The VC Money and the Vicious Cycle
That insurtech startup just raised a massive, $100 million round of funding. They must be doing it right, right? They are probably doing it wrong. All that venture capital money creates immense pressure for growth at all costs. They will spend that money on expensive advertising to acquire new customers, often at a loss. To pay for this, they will have to raise their rates or cut corners on their claims service. The VC funding isn’t for the customer’s benefit; it’s to fuel a vicious cycle of unsustainable growth.
Stop waiting for the perfect algorithm. Start with a minimum viable product.
The Perfect and the Postponed Launch
Our engineering team spent a year and a half trying to build the “perfect” AI underwriting algorithm. They wanted it to be flawless before we launched. We were waiting for perfection, and we were running out of money. We changed our strategy. We decided to launch with a “minimum viable product”—a much simpler version that was good enough to get started. We learned more in the first three months of having a real product in the market with real customers than we did in that year and a half of chasing perfection.
The MGA model I use for my insurtech that most founders have never heard of.
The MGA and the Middle Path
As an insurtech founder, I didn’t want to just be a broker, but I didn’t have the massive capital to become a full-stack insurance carrier. I chose a third path that most founders have never heard of: the Managing General Agent (MGA) model. As an MGA, I have a partnership with a large insurance carrier. They give me the underwriting authority to create, price, and sell my own unique insurance product, and they act as the financial backstop. It gives me the product control of an insurer, with the lower capital needs of a broker.
Your growth problem exists because you believe your tech is the product.
The Tech and the True Product
Your insurtech startup has a growth problem. You keep adding new features to your app, but your sales are flat. Your problem exists because of a fundamental belief. You believe your technology is the product. It is not. The technology is just the delivery mechanism. The true product is the insurance policy itself—the promise of financial protection. Is that product competitively priced? Does it offer broad coverage? Does it solve a real need for the customer? If your underlying product is weak, the most beautiful app in the world will not save you.
Delete that “get a quote in 10 seconds” feature. Your underwriting will improve instantly.
The 10-Second Quote and the 10-Minute Claim Problem
Your website proudly boasts that you can “get a quote in 10 seconds.” This is a terrible idea. To generate a quote that fast, you have to be asking only a handful of superficial questions. This means you are not gathering enough information to properly underwrite the risk. You are attracting high-risk customers who are looking for a cheap, easy policy. Your “10-second quote” feature is leading to a portfolio of bad risks, which will lead to a high loss ratio and, ultimately, to the failure of your business. Your underwriting will improve instantly if you slow down.
The advice on raising capital I give that makes VCs uncomfortable (but works).
The Capital and the Conservative Approach
When I’m raising capital from venture capitalists, I give them advice that makes them uncomfortable. I tell them that I am not trying to be the next billion-dollar “unicorn.” I tell them my goal is to build a sustainable, profitable insurance business with a strong combined ratio. I tell them I will prioritize sound underwriting over growth at all costs. This conservative approach scares away the VCs who are looking for a quick, speculative flip. But it attracts the smart, patient investors who understand the insurance industry and want to build a real, lasting company.
Why the common fear of regulation is irrational and the real fear of insolvency is ignored.
The Regulation and the Real Risk of Ruin
Insurtech founders are often terrified of the complex web of insurance regulation. They see it as a barrier to innovation. This fear is irrational. Regulation, while complex, is a known and navigable challenge. The real, rational fear that they should have is the fear of insolvency. The fear of having a catastrophic year of claims that wipes out your capital and puts you out of business. Regulation is a hurdle. A bad loss ratio is a death sentence.
I tried to build a full-stack insurer from scratch so you don’t have to. Here’s what happened.
The Stack and the Staggering Cost
I tried to build a “full-stack” insurer. This means I wasn’t just building an app; I was building an entire, licensed insurance company from scratch. It was the hardest thing I have ever done. The amount of regulatory capital required was staggering. The process of getting licensed in 50 different states was a bureaucratic nightmare. And the challenge of building a team with deep expertise in underwriting, claims, and compliance was immense. We succeeded, but it took years and tens of millions of dollars more than we ever expected.
The question about “combined ratio” that instantly reveals if a founder knows insurance.
The Ratio and the Reality Check
When I meet with an insurtech founder, I ask them one simple question to see if they know what they are doing: “What is your target combined ratio?” The combined ratio (claims costs plus expenses, divided by premium) is the most important measure of an insurer’s profitability. A number over 100% means they are losing money on their underwriting. A founder who can’t answer this question, or who doesn’t even know what it is, is revealing that they are a tech person playing at insurance, not a real insurance professional.
This old-school method of using human underwriters for complex risks beats every algorithm.
The Human and the Hard Risk
Our insurtech uses an algorithm to underwrite 90% of our policies. It’s fast and efficient for the simple risks. But for the complex, unique, or high-value risks, we use an old-school method: we have a small team of highly experienced human underwriters review the file. An algorithm is only as good as its data; it can’t understand a unique story. A human underwriter can use their judgment, their experience, and their intuition to make a smart decision on a risk that the algorithm would have just rejected.
Stop romanticizing “disruption.” It’s actually just a long, hard slog through regulation and distribution.
The Disruption and the Drudgery
The tech world loves to romanticize the idea of “disruption.” They imagine a brilliant founder having a “eureka” moment and instantly changing an industry. In insurance, that’s a fantasy. The reality of building a new insurance company is not a glamorous act of disruption. It is a long, hard, and often boring slog through two massive challenges: navigating the incredibly complex and slow-moving world of state-by-state regulation, and solving the brutally difficult problem of how to cost-effectively distribute your product to customers.
The principle of “indemnity” that guides every insurtech product I design.
The Promise to Make Whole
When I am designing a new insurtech product, the principle that guides every single decision is “indemnity.” This is the core promise of insurance: to restore the customer to the same financial position they were in before the loss. Every feature in our app, every word in our policy, every step in our claims process is designed to efficiently and fairly deliver on that one, fundamental promise. If the technology is not helping us to better fulfill the promise of indemnity, then it is a distraction, not an innovation.
Why your valuation is vanity and your underwriting profit is sanity.
The Valuation and the Viable Business
In the insurtech startup world, everyone is obsessed with their valuation. Founders will boast about their billion-dollar “unicorn” status. This is a vanity metric. It’s an imaginary number based on the speculation of venture capitalists. The number that represents sanity is your “underwriting profit.” Are you taking in more in premiums than you are paying out in claims and expenses? That is the only true measure of whether you have built a real, sustainable, and viable insurance business. A high valuation with a bad loss ratio is a house of cards.
Forget work-life balance. Aim for a sustainable burn rate instead.
The Balance and the Bank Account
Startup culture is obsessed with “hustle” and often ignores work-life balance. An insurtech founder has a more important balance to worry about. Forget work-life balance; you need to aim for a sustainable “burn rate.” How much of your investors’ cash are you burning through every month? Do you have a clear path to profitability before that cash runs out? The most important balance in a startup is the one between your ambition and your bank account. A sustainable burn rate is the key to having the time to eventually find a work-life balance.
The realization that made me quit trying to replace brokers forever.
The Broker and the Bridge to the Customer
When we launched our insurtech, our goal was to completely replace the traditional insurance broker. We thought we could go directly to the customer. We were wrong. We had a great product, but we had no way to reach the customers. The brokers had the relationships and the trust that we lacked. I had a realization that changed our business: we shouldn’t try to replace the brokers; we should try to empower them. We pivoted our model to become a tool that brokers could use to better serve their clients.
What tech founders do that experienced insurance execs never do.
The Launch and the Lack of a Loss Ratio
A tech founder will launch a product and focus all their energy on user growth and engagement. An experienced insurance executive would never do this. From day one, the insurance exec is obsessed with their “loss ratio.” They know that it doesn’t matter how many customers you have if you are losing money on every single one. They understand that insurance is a game of underwriting discipline, not a race for eyeballs. That fundamental difference in mindset is why so many tech founders fail when they enter the insurance industry.
The investment in a chief underwriting officer that everyone avoids that has the highest ROI.
The CUO and the Culture of Discipline
Many insurtech startups are founded by tech and product people. They often avoid hiring a senior, experienced “Chief Underwriting Officer” (CUO). They see it as an unnecessary expense from the “old world.” This is a massive mistake. The investment in a great CUO has the highest possible ROI. That one person will instill a culture of underwriting discipline that is the foundation of a profitable business. They are the voice of reason that will stop you from chasing bad risks for the sake of growth.
Stop saying “we’re the Uber of insurance.” Say “we’re a managing general agent focused on this niche.”
The “Uber of” and the Uninformed Analogy
Every insurtech founder used to say, “We’re the Uber of insurance.” It’s a lazy and uninformed analogy that immediately signals to an experienced investor that you don’t understand the industry. Insurance is a highly regulated, risk-bearing business, not a ride-sharing app. A much more professional and intelligent way to describe your business is to be specific. Say, “We are a Managing General Agent (MGA) focused on providing parametric hurricane insurance for small businesses.” It’s less sexy, but it shows you actually know what you’re talking about.
The truth about building an insurance company I couldn’t say as a VC.
The Long, Slow, Unsexy Slog
As a venture capitalist, I had to project an image of excitement about “disruptive” insurtechs. Here’s the truth I couldn’t say out loud. Building a real, sustainable insurance company has almost nothing to do with disruption. It is a long, slow, unsexy slog. It’s about grinding through state-by-state regulatory filings. It’s about patiently building up the required capital reserves. And it’s about the slow, difficult work of building a profitable book of business. There are no shortcuts. The “disruption” is a myth; the slog is the reality.
This tiny detail in the user journey separates successful insurtechs from failures.
The Claim and the Click
A failing insurtech will make it easy to buy a policy, but incredibly difficult to file a claim. You have to call a number, you’re put on hold, you have to fill out a dozen forms. A successful insurtech is obsessed with the claims user journey. They understand that this is the moment of truth. They will have a “one-click” claim filing process in their app. They will use AI to approve small claims instantly. They know that a customer who has a seamless, empathetic claims experience is a customer for life.
Why a low price is a trap for insurtechs that need to pay claims.
The Price and the Ponzi Scheme
Many new insurtechs try to gain market share by offering a dramatically lower price than the incumbents. This is a dangerous trap. Insurance is not like software. You have a real, future cost: paying claims. If your price is too low, you are not collecting enough premium to cover your future losses. You are essentially running a Ponzi scheme, using the premiums from new customers to pay the claims of old ones. It’s a cycle that will always, inevitably, lead to insolvency. A sustainable price is the only path to a sustainable business.
Replace your complicated tech roadmap with a simple focus on customer value. You’re welcome.
The Roadmap and the Real Need
Our insurtech had a massive, complicated tech roadmap with a hundred different features we wanted to build. We were always busy, but we weren’t making progress. We replaced it with one simple question that now guides every decision: “Does this new feature provide real, tangible value to our customer?” If the answer is no, we don’t build it. This ruthless focus on customer value, instead of a long list of shiny tech features, has made our product infinitely better and our business much more successful. You’re welcome.
The skill of distribution that’s 10x more valuable than your AI model.
The Model and the Market
Your insurtech has built a brilliant AI underwriting model. It’s more accurate than anything on the market. It’s also worthless if you can’t get your product in front of customers. The skill of “distribution”—of building a cost-effective and scalable way to sell your policy—is ten times more valuable than the quality of your algorithm. You can have the best model in the world, but if you haven’t solved the brutally difficult problem of distribution, you don’t have a business.
Stop treating insurance like a tech problem. Treat it like a people and risk problem instead.
The Tech and the Human Touch
So many insurtech founders come from the tech world, and they see insurance as just another tech problem to be solved with code. This is a fundamental mistake. Insurance is, and always will be, a people business. It’s about trust, relationships, and helping people in their moment of crisis. And it is a business of managing real-world, financial risk. Technology is a powerful tool to make the people and risk parts of the business better, but if you think it’s just a tech problem, you are destined to fail.
The experiment I ran with dynamic pricing that proved our initial assumptions wrong.
The Price and the Perceived Fairness
We ran an experiment. We used our AI model to create “dynamic pricing,” where every single customer got a slightly different, personalized price for their insurance. We thought this was a brilliant, data-driven approach. The customer feedback was a disaster. People felt it was unfair and creepy. They didn’t want a personalized price; they wanted a simple, transparent, and fair price. The experiment proved that just because you can do something with technology doesn’t mean you should. We went back to a much simpler pricing model.
Why your old tech stack worked before but doesn’t for embedded insurance.
The Stack and the Seamless Integration
Your old, monolithic tech stack worked fine when you were just selling insurance on your own website. It is completely useless for the world of “embedded insurance.” To succeed in this new world, you need to be able to seamlessly embed your insurance product into the customer journey of other companies, via API. This requires a modern, flexible, API-first technology stack. If your system can’t talk to the systems of your partners, you will be completely left out of the biggest distribution shift in the insurance industry.
The choice to become a full-stack carrier that everyone judges that actually makes sense for controlling the experience.
The Stack and the Stake in the Ground
Most insurtechs choose to be a “managing general agent” (MGA), relying on another company’s balance sheet. We made the difficult and expensive choice to become a “full-stack” carrier, with our own licenses and our own capital. Everyone judged us for taking the harder path. But it was the right choice for us. It means we have complete control over the entire customer experience, from the policy language to the claims payment. We are not dependent on a partner who might have different priorities. We control our own destiny.
I stopped focusing on the app and started focusing on the policy contract. Our retention soared.
The App and the Actual Product
As an insurtech CEO, I was obsessed with our app. I was constantly pushing my team to add new features and to perfect the user interface. Our customer churn was still high. I stopped. I started focusing all my energy on the actual insurance product itself. We worked with underwriters to create a policy with broader coverage, fairer language, and fewer exclusions. We created a better promise. Our retention rate soared. I learned that a great contract is a much more powerful retention tool than a great app.
The concept of “adverse selection” that nobody in tech understands but changes everything.
The Selection and the Sinking Ship
“Adverse selection” is a simple but powerful concept that most tech founders in insurance don’t understand. It means that the people who are most likely to have a loss are also the most likely to seek out and buy your insurance. If your underwriting is not sophisticated enough to identify and properly price these high-risk individuals, your business will be flooded with them. You will have a portfolio full of bad risks. This is the “death spiral” that has sunk countless insurance companies, and it all starts with not understanding adverse selection.
This unpopular opinion on telematics data will trigger privacy advocates but it’s true.
The Data and the Democratization of Risk
Privacy advocates are terrified of telematics data. Here’s an unpopular opinion: this data can actually make insurance fairer. For a hundred years, insurance prices have been based on crude, demographic proxies—your age, your gender, your zip code. Telematics allows an insurer to price your policy based on how you actually, individually drive. A safe driver in a “bad” zip code can finally be rewarded for their good behavior. It’s a move from collective punishment to individual accountability. It’s a democratization of risk.
Stop copying Lemonade. Do your own thing instead.
The Copycat and the Lack of a Core
Ever since Lemonade went public, the insurtech world has been flooded with copycats trying to replicate their model—the chatbot, the social good angle, the slick design. Stop copying them. What made Lemonade successful was that they had a unique, authentic, and coherent vision from day one. By just copying their surface-level features, you are missing the core. The world doesn’t need another Lemonade. It needs you to have your own, unique insight into a specific customer problem and to build a business that is authentically your own.
The mistake of misinterpreting “customer engagement” I see everywhere that’s so easy to fix.
The Engagement and the Annoyance
Insurtechs are obsessed with “customer engagement.” They think they need to be constantly interacting with their customers through their app. This is a huge mistake. Insurance is a low-frequency, high-importance product. People don’t want to “engage” with their insurance company. They want it to be invisible until the one moment they need it, and then they want it to be perfect. Stop trying to create artificial engagement. Instead, focus all your energy on making the one or two moments that matter—the purchase and the claim—absolutely seamless.
Why this new “Web3 insurance” isn’t innovative. It’s just an unregulated mess.
The DAO and the Disaster
A new startup is trying to build a “decentralized” insurance company on the blockchain, run by a DAO (Decentralized Autonomous Organization). They claim it’s the future. It’s just an unregulated mess. There is no licensed entity. There are no capital reserves. There is no regulatory oversight. And there is no recourse for a customer if their claim is unfairly denied by an anonymous vote of token-holders. It’s not an innovation; it’s a dangerous attempt to bypass a century of consumer protection laws that were created for a very good reason.
The rule I break consistently (I talk to regulators before launching) and why you should too.
The Regulator and the Roadmap
Most tech founders live by the motto “it’s better to ask for forgiveness than permission.” In the highly regulated world of insurance, this is a fatal mistake. I have a rule I break consistently from the tech playbook: I go and talk to the Department of Insurance before I launch a new product. I treat them as a partner, not an adversary. I show them my plan and I ask for their feedback. This proactive, collaborative approach has saved me from countless missteps and has built a foundation of trust with the people who hold my company’s future in their hands.
Stop believing your algorithm is unbiased. Believe in constant auditing and correction instead.
The Bias and the Blind Spot
Your data science team has assured you that your AI underwriting algorithm is “unbiased.” Stop believing them. Every algorithm is trained on historical data, and all historical data contains the echoes of past societal biases. Your algorithm has a blind spot. The only way to build a truly fair system is to assume that bias exists and to build a process of constant, rigorous auditing. You need to be testing your model for unfair outcomes, and you need to have a process to correct the biases when you find them. The belief in a perfectly neutral algorithm is a dangerous fantasy.