Robert Marshall is the founder and CEO of Whisker Labs. Whisker Labs, a spinout of Earth Networks, delivers next-generation home energy intelligence technology to realize the full potential of the connected home. In 1992, Marshall co-founded AWS Convergence Technologies, the company that would become Earth Networks, by pioneering the networking of weather sensors and cameras using the internet. By developing groundbreaking technology to find "signals" — valuable, meaningful intelligence — in big-data "noise," Marshall improves people's lives and protects their livelihoods. He has appeared on CNN, BBC World News and ABC Nightly News and has been quoted in major news outlets that include the New York Times, the Washington Post, Nature and Scientific American.

Paul Carroll

One of my goals for the Predict & Prevent movement is that it will be able to lay out a clear economic argument, showing that the savings are greater than the cost of the investment in prevention. You and the Insurance Information Institute, the Triple-I, recently reported on a study that found significant savings from installing your Ting devices in homes. Would you start us off by telling us what you found?

Bob Marshall

We document that Ting prevents 0.39 electrical fire claims per 1,000 home-years. If you multiply that by the severity, which has gone up considerably over recent years, then you get to $81 per year per home in savings from Ting. 

That's obviously greater than the cost of a Ting, and that's why insurers love the idea. Not only does it protect their customers and create a great experience and good engagement, but it delivers a clear ROI, paying for itself and beyond.

Paul Carroll

And the benefits are actually greater than the cost savings on fire damage, right? Preventing a fire keeps a family out of danger and saves them from a huge amount of hassle and dislocation.

Bob Marshall

A fire is often devastating for the family. You could lose pets, you could lose lives, the whole thing.

The savings on the insurance side are higher than what's calculated there, too. There is also the cost to the agents, who often have to work with families every week for a year or more to try to itemize all the losses and damage from a catastrophic fire and help them recover. 

The best claim is one that never happens. To the extent we can prevent fires, it's good for everybody.

Paul Carroll

You’ve said that people who install a Ting may become more open to other Predict & Prevent initiatives. I'll share a Triple-I blog on the topic, but would you briefly explain how that works?

Bob Marshall

Homeowners have an innate fear of fire, so when a carrier partner offers them Ting, they're very motivated to say, “Yes. I want that.”

We've worked really hard to deliver a simple and seamless experience for the homeowner. You just plug the Ting into the wall. Setup takes two minutes. Then we deliver valuable information every week with summary reports, power outage notifications, and other beneficial insights.

If you lead with Ting and the homeowner opts in and has a great experience, then when you follow with, say, water, they're much more likely to say, "Hey, I like this fire thing the carrier offered me. I think I'll do the water thing, as well."

Paul Carroll

What’s the latest on the number of homes you’re in?

Bob Marshall

We currently have over 1 million active homes in our network. We're consistently adding 40,000 to 50,000 homes per month, so we're growing very rapidly.

The ROI report was super important for us. Gathering enough data to document results is never easy when you're dealing with low-frequency perils such as fire and even water damage. You have to have a lot of data to properly document the loss prevention, but we have that now. We overcame a number of obstacles with that research and paper to make the results really clearly documented, which is awesome.

Paul Carroll

If you do the math, based on the current number of homes you serve and the prevention of .39 fire claims per 1,000 homes, you’re preventing some 400 fires a year. And the number will only grow as you expand your reach.

The last time we talked, a few months ago, 30 carriers were working with you to provide Tings to their customers. Where do you stand now?

Bob Marshall

I think we're at 34 now, and obviously going up. At this point, it's pretty clear most every carrier is going to work with us because Ting is proven to work. 

We're trying to make the experience more seamless and easier for the carrier, because partnering and distributing loss-prevention devices isn’t something they naturally do. And I think we're pretty much there. 

Paul Carroll

I assume it’s important for insurers that you automatically verify that a Ting is plugged into a wall socket and active, not just sitting in a box, unopened. I know home insurers struggle to not just know that an owner has a security system but that it’s activated.

Bob Marshall

Yes, absolutely. The way we structure our partnerships with carriers, Whisker Labs doesn't get paid if the Ting is not installed and active. We're structured in a way where we're 100% aligned.

Paul Carroll

What progress have you made in your international expansion efforts, and what challenges are you encountering given the different electrical standards globally?

Bob Marshall

We are working on opportunities to expand outside North America, though I can't talk about it too much. I think I'll have more to say on that in the coming months.

The electrical problems and fires are worse in many parts of the world. The electric codes are not as rigid. The buildings are older. The homes are older. The wiring is older. The voltage is higher, which creates more potential for the arcing that can cause fires.

The opportunity for us to prevent fires is even higher outside North America than it is here.

Paul Carroll

How does your technology help monitor electricity quality, particularly for data centers and other situations where reliable power is critical? I’ve read that increased demand is degrading quality.

Bob Marshall

We are doing a ton of work in that regard. Bloomberg actually did a comprehensive analysis a few months ago using our Ting data along with a database of data centers. What's clear is that the power quality for homes in the vicinity of data centers is materially worse.

With bad power quality, your large appliances like air conditioners, water heaters, refrigerators—anything with a motor—their energy efficiency is materially reduced. Air conditioners are half of the energy used in a home. If you reduce their energy efficiency by 15% or 20%, that's a material cost to the homeowner that is hidden. We also see that other power-quality problems—outages, power surges, brownouts—happen much more often where the grid is stressed in the vicinity of data centers. Our preliminary analysis suggests that costs to homeowners from poor power quality can be up to $1,000 per year. 

It's not exclusively near data centers. In general, with the grid becoming more stressed because of the demands and complexity, we're seeing a decrease in the power quality that is very clear and unambiguous.

Paul Carroll

Your network of sensors is proving to be useful in pinpointing grid problems that could lead to wildfires, such as the Lahaina and Eaton Fire disasters. What progress have you made in delivering this critical information to utilities ahead of time rather than retroactively?

Bob Marshall

We are working extraordinarily hard on solving the problem, and we are making some progress.

One key issue is trying to pinpoint the exact source of any given fault that could cause a wildfire. We can do that reasonably well, though we still have work to do. 

When you look at cases like Lahaina and Eaton, our data shows that the entire grid was under incredible stress and was experiencing a high frequency of faults for many hours in advance of the wildfire ignitions. Faults occur when tree limbs touch a wire or wires touch each other, and each incident can produce a spark that ignites a wildfire. Most don't, or we'd have wildfires everywhere.

What our data could help utilities with very quickly is seeing when their grid is stressed and making better decisions about shutting off the power. If you shut the power off, there's no energy to create the spark that causes the fire.

For some of these devastating wildfires, the only solution is to prevent the spark, because when you have 70 mile-an-hour winds and dry brush, there's no way to stop a fire once it starts. There's no amount of water or firefighters that can contain it. But that's a tough decision to turn off power to any community, and utilities have for decades focused on keeping power on essentially at all costs.

Paul Carroll

Any closing thoughts on the industry’s move toward a Predict & Prevent model?

Bob Marshall

We're excited, and we really appreciate that The Institutes, Triple-I, and the insurance sector are embracing the Predict & Prevent future.

I think that vision is so key, and the direction that you all have helped establish is truly taking hold. We're pleased to be able to make our contribution to it and hopefully help drive it forward.

Paul Carroll

Thanks, Bob. I always feel more encouraged after we talk. 

Almost exactly one year ago, we published our thought leadership article "Embedded Insurance: Major Disruptor Can Bridge Huge Coverage Gap."

We pointed out that embedded insurance isn't new and that the wide-ranging, "at point of sale" opportunity is significant. Purchasing life insurance at the airport before flight departure was a perfect example of "version 1.0" of embedded insurance. We also shared a Forrester forecast that "the embedded insurance market is expected to grow from $156.06 billion of gross written premiums in 2024 to more than $700 billion by 2029, a CAGR of 35%."

It now appears that its growth is even greater than expected. 

The embedded insurance market is now forecast to reach $1.1 trillion in global gross written premiums (GWP) by 2033, representing about 15% of the total GWP.

"Embedded insurance is becoming a growth engine for global financial services, a trajectory which reflects a customer- and technology-driven reshaping of how protection is bought, sold, and experienced," according to the latest Open and Embedded Insurance Observatory Report.

The report says the opportunity does not belong to banks and fintechs alone. Competition from adjacent industries is intensifying. Regulatory attention is beginning to focus on these models, which will require flexibility. Legacy IT systems are still a limitation for many incumbents, while data privacy and trust remain mission-critical.

A continuously evolving segment is auto insurance, added at point-of-sale together with a car purchase or lease. Although not a new concept, real-time insurance quotes are new and ease of adding or switching easily fits alongside financing and other add-on offers — all right at the dealership before driving off the lot. Numerous and growing insurer/car brand alliances, whether with or without driving data sharing, have popped up throughout the automotive industry. Such household purchases are major life moments, with an opportunity to switch insurers, hence the constant attention.

Embedded auto insurance

Of the many categories of embedded insurance, auto insurance represents an ideal opportunity and the most effective and frictionless delivery model. Consumers seek simplicity and one-stop purchase experiences, and their loyalty to auto insurers is eroding quickly in the face of continuing premium increases. Auto insurance is now a commodity, and switching is more frequent than ever.

According to Polly's Q2 2025 Quarterly Embedded Auto Insurance Report, the connection between insurance engagement and dealership profitability grew even stronger. Dealers who introduced insurance quotes into the sales process saw an average 20% lift in finance and insurance (F&I) gross profit — an extra $313 per deal.

When customers went a step further and purchased a policy, the effect was even greater. Those deals delivered a 31% lift in F&I gross, or $501 more per transaction.

The takeaway is clear: Whether a customer simply views quotes or binds coverage, insurance engagement is one of the most reliable levers for increasing dealership profit. It creates trust, keeps deals moving forward, and consistently raises the ceiling on F&I performance.

Technology-enabling embedded insurance

Insurtechs and integrations are the primary enablers and drivers of embedded insurance. This partnership benefits both customers and businesses by offering convenience, new revenue streams, and personalized coverage options powered by technology.

Selected insurtech companies specialize in enabling embedded insurance solutions for various industries:

• Cover Genius: Designs embedded insurance platforms for large brands like eBay, offering diverse coverage from shipping protection to rental car insurance.
• Clearcover: Has an embedded insurance strategy that includes partnerships with companies like Experian to offer bindable quotes to consumers when they are shopping for auto insurance.
• Roamly: Offers software tools and a platform that allows non-traditional insurers and other businesses, like car dealerships and marketplaces, to embed insurance into their workflows using APIs.
• Extend: Focuses on modernizing warranties and protection plans for e-commerce retailers.
• Wakim: Provides white-label, usage-based liability coverage for the gig economy and equipment rental.
• Zego: Uses application programming interface (API) technology to offer flexible commercial insurance to platforms such as Uber and Deliveroo, providing "pay-as-you-go" coverage for drivers.
• Bolttech: Provides a platform to embed tailored insurance products directly into existing customer journeys, from car dealerships to fintech apps.
• Matic: Offers an embedded insurance platform for financial institutions, allowing partners to offer competitive auto insurance options at the point of sale, particularly through partnerships with mortgage lenders.
• Tint Embedded Insurance: Helps brands embed insurance directly into their platforms, aiming to increase conversion rates and profitability by making insurance a feature, not a standalone product.
• Openkoda: Provides an open-source framework for building and deploying custom insurance applications, including embedded forms for quoting and policy sign-ups, with a focus on speed and control.

Auto insurance focus

Polly enables embedded auto insurance by integrating its digital insurance marketplace into the car-buying process at dealerships, allowing customers to compare quotes from multiple insurance carriers and purchase coverage at the point of sale. This seamless integration uses technology to connect the dealership's existing software with the insurance marketplace, so customers can get instant quotes and choose the best policy without leaving the dealership or going through a separate, time-consuming process. 

Embedded auto insurance partnerships

Carvana and Root

Carvana and Root have a partnership where Carvana sells auto insurance, underwritten by Root, to its customers during the online car purchase process. Customers can get an insurance quote and bind a policy from Root directly through the Carvana checkout, streamlining the process of getting their new car covered. While Carvana is the seller, the actual insurance policy is with Root Insurance.

Stellantis and bolt

Stellantis has partnered with bolt, an insurtech company, to provide embedded auto insurance for its Chrysler, Dodge, Jeep, Ram, Fiat, and Alfa Romeo customers in North America. The partnership aims to simplify and personalize the insurance purchasing process by allowing customers to buy insurance directly through Stellantis brand websites and apps, with future plans for usage-based options using telematics data.

OEM role in embedded auto insurance

While auto manufacturers (OEMs) do not directly sell auto insurance at their dealerships, many major insurance companies partner with dealerships to offer insurance options on-site, and some financial services arms of OEMs offer insurance-related products.

OEMs like Tesla and Volvo are changing the game, making insurance part of the car ownership experience itself.

Insurance companies that partner with dealerships include Travelers, Zurich, and Ally, with some having a strong history in the auto industry. Dealerships often facilitate insurance by having agents or brokers available to help customers with insurance needs at the point of sale.

• Partnerships with insurance companies: Dealerships frequently partner with major insurance providers like Travelers, Zurich, and others to make insurance purchasing convenient for buyers.
• OEM financial services: The financial arms of some manufacturers, like Ally, have established insurance divisions specifically for the automotive sector, including dealerships.
• On-site agents: Dealerships often have insurance agents or brokers on-site to help customers who don't have current insurance or are unhappy with their existing provider.

Other noteworthy embedded models/partnerships

• Liberty Mutual partners with Jaguar Land Rover North America to provide tailored auto insurance solutions for Jaguar vehicle owners in the U.S. during the car buying process
• Tesla comes with built-in insurance features
• Toyota Auto Insurance is underwritten by Toggle, a digital and embedded insurance company that is part of Farmers Insurance
• INSHUR formed a partnership with ride-sharing service Uber in 2018 to embed insurance directly into Uber's platform, providing on-demand drivers with streamlined, personalized insurance coverage that adapts to driving schedules
• Turo, a peer-to-peer car-sharing platform, collaborates with Liberty Mutual to offer embedded insurance for its users
• Chubb just announced the debut of a new AI-powered optimization engine within Chubb Studio, the company's global technology platform for embedded insurance distribution partnerships. Sean Ringsted, chief digital business officer at Chubb, said the new tool lets digital distribution partners enhance engagement, improve conversion, and support financial resilience with relevant insurance protection. 

Looking ahead

Implementing embedded insurance distribution channels is not a trivial undertaking, and there will be several technical, regulatory, business, and cultural obstacles, so you need to get started.

Whether you are an insurer, insurtech, agent, broker, MGA, retailer, wholesaler, or anywhere else in the insurance ecosystem and supply chain, you must invest now in learning how your business can participate in the embedded economy of the future.

This week's newsletter is really just an excuse for me to share a cool story about sensors so tiny that they're being used to track the migration of hundreds of individual Monarch butterflies as they travel from Canada to winter in Mexico.

I've been banging the drum about the importance of ever-shrinking sensors since at least 2013, when Chunka Mui and I published "The New Killer Apps" and listed ubiquitous sensors as one of our six technology megatrends to exploit. I've been fascinated by Monarch butterflies since coming upon a traffic jam on a country road in Mexico in the '90s, stepping out of my car and realizing that the "leaves" on trees up the hill were actually millions of Monarch butterflies. So I just couldn't pass up this week's story of Monarchs being outfitted with sensors that include a solar panel, a battery, a radio, and an antenna--while weighing six-hundredths of a gram.

Oh, and there are plenty of implications for insurers, where ever cheaper, ever more powerful, ever smaller sensors are already enabling the move to a Predict & Prevent model and where, as the butterfly story shows, there is still loads of room for progress. 

An article in the New York Times says about 400 butterflies have been fitted with the sensors and tracked, via a phone app, as they made their way south. Some were tracked for as long as nine weeks as they headed south to the winter colonies where they and their ancestors were born. One was tracked as it blew out to sea from Cape May, NJ, to the Bahamas and then flew west to Florida. As you might imagine, only about one in four survives the arduous journey. 

Even at 60 milligrams, the sensors add 12-15% to a Monarch's body weight, and they aren't cheap; they cost $200 apiece. But Moore's Law has been taking care of size and cost issues for electronics for some 60 years now and let the inventors get the sensors to the point where they're practical. The inventors also took advantage of the billions of Bluetooth devices that are already out there: If a Monarch flew within 300 feet of a Bluetooth-enabled device, the device would pick up the butterfly's radio signal and share its location with the tracking app.

Moore's Law and the spread of "mesh" networks like the one Bluetooth allows the butterfly sensors to access will continue to benefit the Monarch trackers--and insurers that choose to take advantage.

Telematics in auto insurance shows what can happen as technology moves down the size and cost curves. When Progressive pioneered its Snapshot program in 2008, the company quickly gained market share, but success was limited by the fact that Progressive had to pay for dongles and that drivers would then have to figure out how to insert them under their dashboards. When motion sensors became cheap enough that they were routinely embedded in smartphones, Progressive rolled out an app that not only had almost zero marginal cost but that was super-convenient for drivers. Its market share soared from fourth among U.S. auto insurers in 2015 to second this year. Its combined ratio in 2024 was more than six percentage points below the industry average. 

While other uses in insurance haven't had the same sort of dramatic success, some are getting there and enabling the move to Predict & Prevent. 

Whisker Labs's Ting device, which plugs into a wall socket, has now demonstrated that it prevents so many home fires that more than 30 insurers are giving the device to customers for free. Water leak sensors keep shrinking in size and increasing in capability, to the point that some insurers are at least experimenting with giving them away to policyholders. Nauto's windshield cameras -- one pointed at the road, one at the driver, with AI monitoring and warning the driver of impending danger -- is reducing accidents by 60-70%-plus in truck fleets. Roost sells batteries for smoke detectors that contain sensors and communication capabilities so they can send an alert to your phone and let you know of a problem when you aren't home. Home security systems now let you just affix inexpensive sensors to windows and doors that can communicate wirelessly to you or a monitoring company, without all the wiring that used to be required.  

FitBit, Oura and other fitness trackers are riding the sensor cost/size curves to keep adding capabilities. My first FitBit, which I bought maybe 10 or 12 years ago, just tracked my heart rate and my time sleeping. My Oura ring now tells me about my heart rate, my heart rate variability (which I didn't even know was a thing until Oura told me about it), my blood oxygen level, body temperature and more. Separate devices can track blood sugar, blood pressure, etc., and many of those sensors will find their way into the devices we wear on our fingers or wrists, much as motion sensors and so many other capabilities have been absorbed into our smart phones. That's just how technology works: Everything gets cheaper and gets absorbed into a dominant platform.

Insurers will also be able to benefit from the sort of "mesh" approach that the Monarch butterfly trackers use. The basic idea is that a device doesn't need to communicate directly with its host. It can just "mesh" with another device, which can then connect with the host or can even just keep passing along information to other devices (in this case, using Bluetooth) before reaching one that can connect with the host. 

Bluetooth is available to insurers that want to collect a signal from a sensor in a home, in an office, in a factory, in a car, on a person, or whatever. Amazon also offers a mesh network called Sidewalk, based on Echo and Ring devices. If you have enough power to get a signal to one of the hundreds of millions of those devices, you can collect that information. There are surely other mesh networks available, too, if not on the Amazon or Bluetooth scale.

Cost and size will still be an issue for some potential uses of sensors by insurers, but today's issues won't be tomorrow's. Moore's Law will keep shrinking devices and slashing costs, so if you can see a plausible case for use of a sensor, you need to be thinking about what the capabilities and costs will be like a few years from now and, perhaps, start experimenting today.

The real issue is just one of creativity for the insurance industry: What information can we imagine gathering via sensor that will let us prevent or at least minimize a loss, so we can protect people and limit claims?

If we can track a single butterfly from New Jersey to the Bahamas to Florida, what can't we do?

Cheers,

Paul 

In an in-depth study, Weighing the Evidence, RGA has quantified the expected mortality and morbidity impacts of incretin-based drugs, including GLP-1s, approved as anti-obesity medications (AOMs) and diabetes treatments in the U.S., U.K., Canada, and Hong Kong. Results include the following key findings:

Population mortality

• By 2045, incretin-based drugs such as GLP-1s could reduce mortality in the U.S. by 3.5% in a central scenario, 8.8% in an optimistic scenario, and 1.0% in a pessimistic scenario.
• Under the same central scenario, mortality could decrease by 2.0% in the U.K., 2.6% in Canada, and 1.4% in Hong Kong.
• Mortality improvements vary by age, with ages 45-59 seeing the biggest reduction and age 85+ the lowest reduction.

Population morbidity

• Populations could see smaller but still positive reductions in the incidence of cancers over the same period.

Insured mortality and morbidity

• Insured groups and annuitants are likely to see somewhat lower mortality and morbidity reductions than the general population.

The report also explores how the impact of these treatments could exceed those projected in RGA's central scenario and the potential for further reductions in mortality and morbidity as the therapeutic landscape for incretin-based drugs expands beyond obesity and diabetes. The authors further outline considerations for insurers on how to incorporate the report's insights into future improvement bases and highlight implications for underwriting, claims, and policyholder behavior.

Key Takeaways

1. At the general population level, AOMs will have a meaningful impact on mortality. 

This will differ by geography, largely reflecting the obesity profiles of different markets, and by age, sex, and access to the drugs.

RGA models the impact of AOMs over the next 20 years to 2045 with reference to three key groups of assumptions: effectiveness, uptake, and relative risk of mortality and morbidity. The report calculates optimistic and pessimistic scenarios by flexing these key assumptions to plausible higher and lower values. The chart below shows the expected mortality impact of these scenarios at the population level.

Table 1: Cumulative population mortality improvements over 20 years to 2045 due to AOMs under three scenarios

It is important to recognize that the impact of AOMs will vary by age, reflecting differences in obesity levels, the mortality risk associated with obesity, and differences in uptake. The table below shows the expected mortality impact in the central scenario by market and by age.

Table 2: Cumulative population mortality improvement impacts over 20 years to 2045 due to AOMs by market and age

2. AOMs will likely have a smaller impact on general population morbidity.

The report defines morbidity as the incidence of claims under a typical critical illness product. Cancer is the largest single cause of morbidity incidence in critical illness products. While reducing body mass index (BMI) does reduce the risk of cancer incidence, it is not to the extent that lowering BMI reduces mortality risk. Therefore, RGA expects morbidity impacts to be generally smaller than the corresponding mortality impacts.

Table 3: Cumulative population morbidity improvements over 20 years to 2045 due to AOMs under three scenarios

3. Insured groups are likely to see somewhat lower mortality and morbidity impacts than the general population. 

Lower average BMI means less scope for improvements, even though insureds have greater access to the drugs.

Insured lives and annuitants typically come from a higher average socioeconomic group than the general population and generally are expected to have a lower average BMI. Insured lives are also typically underwritten and have a different mix of causes of mortality and morbidity than the general population. The RGA model projects that the lower average BMI for insured groups has more impact than the increased ability to access the drugs, and so the overall mortality and morbidity impact is typically lower than for the general population. The actual impact AOMs will have will reflect the characteristics of a life and health insurer's insured portfolio.

4. An insurer's current mortality trend assumptions likely include anticipated improvements from drivers such as medical advances. 

It may be too early to make material adjustments to those assumptions for AOMs, but they increase confidence in future mortality and morbidity improvements.

RGA's study results described so far reflect changes to current mortality and morbidity rates. When translating to impacts on assumptions, it is important to consider that (re)insurers already assume positive improvements in the future. Anti-obesity medications are a tangible advance contributing to these future improvements.

5. This is a fast-moving space with significant uncertainty. 

Model assumptions will need refining as new evidence emerges and as new indications for the drugs are approved. 

The upside potential of these drugs is exciting, but challenges linked to safety, side effects, access, and adherence need to be overcome to achieve the full impacts anticipated in this report. While cost is currently a barrier to uptake, growing competition and the arrival of generic and oral formulations are expected to lower costs significantly. The next wave of incretin-based therapies is poised to offer significant advantages over the current generation for treating diabetics and for weight loss in those living with obesity. RGA's model has already anticipated some of these developments, but this is a fast-moving space that requires continual review of model assumptions. 

Incretin-based therapies are under active investigation to treat a growing spectrum of medical conditions ranging from neurodegenerative disorders to substance abuse. As approved indications continue to broaden, and adoption scales in those with established disease, the cumulative impact on public health could be profound.

There is growing interest in the potential of incretin-based drugs as preventive medicines. Their systemic anti-inflammatory effects, metabolic regulation, and influence on satiety and insulin sensitivity suggest they could help prevent the onset of multiple chronic conditions. If these benefits extend to individuals without established disease, there could be a significant reduction in morbidity and mortality across the general population.

6. Insurers should consider the impact on business, including pricing and reserving assumptions, new policyholder behaviors, underwriting, and claims.

RGA's companion paper, "Evaluating Biometric Trend Drivers: How to reflect medical breakthroughs and other drivers in forward-looking assumptions," explores the practicalities for insurers in maintaining an up-to-date view on emerging biometric trend drivers such as AOMs and provides a framework for incorporating the insured mortality impacts into insured improvement bases. 

The use of AOMs introduces the risk of anti-selective policyholder behavior, as individuals who have lost considerable weight may lapse their rated policies and re-enter with better terms. As such, insurers may not capture the full economic benefit of improved mortality and morbidity. 

The increasing use of AOMs will significantly influence underwriting risk assessment. As evidence accumulates, underwriting approaches must evolve to recognize improvements while maintaining vigilance to validate these therapies' potential. Accurate disclosures at the underwriting stage may need to be validated at the claims stage, and claims assessors will require a deep understanding of the use of AOMs to ensure accurate interpretation of disclosures made as part of the insurance application. 

Conclusion

AOMs have the potential to significantly improve population mortality and disease incidence rates. 

The impact on insured groups is likely to be somewhat lower, and it may be too early to make material adjustments to insured trend assumptions, but the efficacy of AOMs to date increases confidence in future mortality and morbidity improvements. 

This is a fast-moving space with significant uncertainties, so monitoring developments closely will be vital to responsible and successful insurance practices. Model assumptions will need continual refining as new evidence emerges and as new indications for incretin and hormone-based medications are approved. 

To learn more, view the full study: "Weighing the Evidence: Quantifying the mortality and morbidity impacts of GLP-1 and other incretin-based drugs in the US, UK, Canada, and Hong Kong populations." 

Additional Contributors: Dr. John J. Lefebre, Vice President and Senior Technical Global Medical Director, Global Medical; Matt Battersby, Senior Vice President, Head of Global R&D; Marilda Kotze, Vice President, Head of Global Underwriting & Claims

Guide | Beyond Legacy Tech: A Modernization Guide for the AI Era

The insurance industry is at a crossroads. While many carriers are exploring AI, few have achieved true transformation. Nearly two-thirds remain stuck in pilot projects, held back by outdated, siloed systems that fragment data and slow innovation.

In this new guide from Origami Risk, discover why modernization—not experimentation—is the foundation for AI success. Learn how modern, cloud-based SaaS platforms enable insurers to move faster, scale smarter, and compete in an era defined by intelligence.

Download the guide to uncover:

1. Why AI adoption has stalled, and how to break free from legacy drag
2. How modern, multi-tenant SaaS platforms accelerate AI deployment
3. A side-by-side look at build-versus-buy modernization paths
4. Tested frameworks to align technology, finance, and operations stakeholders
5. Strategies to turn modernization into a growth engine for underwriting, claims, and customer experience

AI is no longer a pet project—it’s the next stage of insurance evolution. But only those who modernize their core systems will harness its full potential.

Don’t let legacy tech hold you back.

Download the Guide Now  

Sponsored by: Origami Risk

According to Experian's State of Claims 2025 report, more than half of insurers (54%) say claim errors are increasing, and nearly seven in ten (68%) find submitting clean claims more challenging than they did a year ago. As costs rise and customer expectations grow, insurers are realizing that traditional automation – while once sufficient – can no longer keep pace with the complexity of modern claims.

We are now entering the era of agentic AI: intelligent, goal-driven systems that work collaboratively to interpret data, make decisions, and adapt in changing conditions. When combined with intelligent orchestration, agentic AI allows insurers to accelerate claims resolution, improve accuracy, and deliver transparency at every step of the process.

Digital Insurance Workflows: Accelerating Resolution

Claims handling has long been one of the most resource-intensive areas of insurance operations. Agentic AI changes that dynamic. Instead of relying on static workflows or human-led review queues, autonomous agents can continuously gather and assess data from multiple sources—emails, repair estimates, photos, and videos—and validate coverage in real time.

An AI agent can instantly classify the complexity and exposure of a claim, cross-reference it with policy details, and even calculate settlement recommendations based on prior decisions. Straightforward claims can be resolved automatically within hours, while complex or high-risk cases are routed directly to human adjusters with a full, auditable summary of every AI-driven action.

This creates a system where human expertise is amplified, not replaced—allowing adjusters to focus their time on empathy, judgment, and nuanced decision-making.

The Power of Automating Insurance Claims

The claims process is pivotal to the customer journey. Bottlenecks lead to frustration, slow resolutions, and lost trust. Insurers that leverage agentic AI demonstrate operational excellence while simplifying tracking, providing visibility, and improving efficiency.

Contrary to concerns about AI replacing jobs, this is a collaborative model: AI handles repetitive or time-sensitive tasks, and humans focus on strategic decisions and customer engagement. This collaboration benefits both employees and clients.

Accuracy Isn't Optional

Accuracy is mandatory in insurance. Errors or inconsistencies undermine both compliance and customer confidence. Agentic AI enforces consistent, compliant processing while minimizing human error and leakage.

Fraud detection is another area where AI excels. By cross-referencing claims with historical data, public records, and behavioral patterns, agentic AI identifies anomalies before payouts are made. These systems don't just detect fraud—they prevent it, learning from emerging patterns and adapting continuously.

Keeping Teams Involved

Human adjusters remain central to automated workflows. Human-in-the-loop strategies empower staff to validate complex claims, oversee governance, and ensure customer satisfaction.

Autonomous agents automatically route large losses, liability disputes, or high-risk flags to the human team with a complete, aggregated summary of every AI-driven action. Every step is logged and auditable, giving supervisors and regulators full transparency while ensuring confidence in autonomous workflows.

Being Present And Forward-Thinking

When adopted strategically, agentic AI transforms claims from a cost center into a competitive differentiator. Insurers gain speed, scalability, and compliance—while delivering higher customer satisfaction.

The next step for leaders is identifying high-impact, low-complexity workflows to pilot and refine. The insurers who take that step now will define what "intelligent insurance" means for the industry over the next decade.

The insurance industry is facing pressure from all sides. Customer expectations are rising, as they expect the same sort of nearly instant responses they get when shopping or viewing bank balances online--creating problems for carriers that still rely on manual, paper-driven workflows. Yet regulators are requiring fairness and transparency even as the pace quickens. And newcomers with modern tech stacks are capturing market share.

This is the reality of today's insurance industry. Research published in Frontiers in Artificial Intelligence shows that AI is already transforming underwriting, fraud detection, and claims management. At the same time, customer churn is accelerating due to yearly rising rates, claim severity, and cost increases.

The stakes for success could not be higher. The technology to achieve success already exists, and its adoption is accelerating. Forward-leaning carriers are already leveraging it, gaining efficiency and customer loyalty. If organizations choose not to act, they risk being left behind by both the industry and, most importantly, the policyholders.

Reimagining Insurance in the AI Age

AI brings with it the unbound potential to elevate the way insurers operate. At the core is the ability to extract valuable insights from years of unstructured data. Take underwriting. Traditionally, it has been a data-heavy process; however, when an underwriter is augmented with AI, they can make faster, more accurate risk assessments. This accelerated process is enabled in part by the insurer's expertise and AI's ability to analyze vast datasets and identify emerging trends and patterns that may have otherwise remained buried in their records.

Claims processing is another area that can experience an efficiency boost by leveraging AI. Automated systems powered by AI can analyze claims data more efficiently, speeding up decision-making and reducing the time it takes to settle claims. In an industry where customer experience is paramount, a quicker claims process is one of the keys to both maintaining and improving satisfaction and loyalty.

Some major insurance carriers who have adopted this approach include:

• Lemonade: Using its claims handling AI agent, aptly named AI Jim, Lemonade can resolve simple property claims in seconds or minutes. While there are still claims that will need human intervention, Lemonade makes funds available one to two days after claim approval.
• Allstate: The claims process is often complex and riddled with insurance jargon. To avoid these complexities, Allstate automated communications with AI to improve efficiency, customer experience, and bring empathy back to the industry.

Another financial impact of AI-enhanced systems is their ability to help identify fraud by flagging suspicious patterns and anomalies. This detection ability is critical in protecting insurers from hemorrhaging profits due to fraudulent losses.

Where to Unlock Business Value With AI

For many insurers, daily operations are dominated by routine. This means agents and adjusters spend valuable hours navigating paperwork across siloed and legacy systems, leaving less time for meaningful customer interactions. By embedding AI into core processes, carriers have an opportunity to shift their workforce's focus toward higher-value activities that directly affect satisfaction, loyalty, and retention.

To ensure that this integration is done correctly, carriers should look for areas where AI-driven solutions will provide real returns on investments, including:

• Repetitive tasks: Today, adjusters spend hours on data entry and form processing. AI can automate intake, document classification, and updates, freeing time for customer service.
• Real-time decision support: Adjusters must piece together information from multiple sources while customers wait. AI can surface policy details, claim history, and regulatory guidance instantly during conversations.
• Cross-system orchestration: Staff often toggle between claims platforms, CRMs, and document repositories. AI can connect these systems and present a unified view, speeding up responses.
• Complex pattern recognition: Fraud detection still depends on manual review of anomalies. AI can flag suspicious patterns early, guiding investigators to high-risk cases more efficiently.
• Where humans & AI can collaborate: Employees are stretched thin between admin work and customer care. By handling repetitive tasks, AI empowers adjusters to focus on empathy, trust-building, and problem-solving.

By shifting the balance from routine to meaningful work, AI enables insurers to improve both operational efficiency and customer loyalty.

Meeting Today's Expectations

Customer expectations have reached a breaking point. Policyholders demand instant support, yet too often encounter rigid scripts instead of real solutions. A 2025 study by J.D. Power found that 57% of customers are shopping for insurance year over year, a staggering jump from 49% in 2024.

It's not a complex equation: Insurers with superior customer experience see higher returns by unleashing human expertise for relationships, judgment, and empathy—not by eliminating jobs. To build on positive customer experiences, insurers should leverage AI to instantly handle claims while meeting rising standards for compliance, privacy, and fairness. The path forward is clear: innovation depends on strategic human-AI orchestration.

Artificial intelligence has entered a new era. It's no longer just a statistical predictor crunching historical data. It's now a creator, planner, and autonomous actor capable of generating content, making decisions, and executing multi-step tasks. This leap from traditional AI to generative and now agentic AI has fundamentally changed the risk landscape. These new AI systems therefore demand a rethink of how we measure, manage, and insure the risk.

Traditional insurance frameworks, predominantly built on backward-looking data and well-understood failure modes, are not suited for systems that learn, adapt, and change behavior in real time. As AI becomes more deeply woven into business, infrastructure, and daily life, the question is no longer if it will fail but how and who bears the cost when it does.

To unlock the full potential of AI safely and at scale, the insurance industry must innovate. This is not just about transferring financial risk, but also about creating market incentives for trustworthy AI adoption. Insurers and risk managers will need to deploy new tools to quantify, price, and monitor AI exposure, ensuring that innovation and safety evolve together. This is urgent, as many AI risks are sitting silently inside existing policies, often unpriced, unmanaged, and waiting to materialize. The systemic risk posed by this silent coverage represents a significant, largely unmodeled aggregation exposure for carriers and creates uncertainty for the insured.

In the sections that follow, we explore how the AI risk profile is evolving and why a new generation of assurance and insurance mechanisms will be critical to building confidence in the intelligent systems that will increasingly shape our world.

The Evolving AI Risk Profile

AI systems have gone through three major generations, each more capable and complex than the last. With every step, the risk profile has expanded.

1. Traditional AI: Early AI systems were essentially statistical predictors. They learned patterns from structured data to forecast outcomes -- for example, credit scores, demand forecasts, and spam detection. Their risks were relatively stable and easy to quantify, mostly limited to data quality problems or model misspecification.
2. Generative AI: Generative AI (e.g. large language or diffusion models) doesn't just analyze data; it creates content. This creative power comes with new risk: producing plausible but false outputs (hallucinations), reusing copyrighted material from training data, or shifting behavior as APIs or retrievers change over time. Because these systems are composable (built from multiple moving parts) and dynamic (updated frequently), they can change behavior without warning.
3. Agentic AI: The newest wave, agentic AI, adds autonomy, reasoning, and tool use. Autonomy brings systemic risk: small local errors can cascade across an entire chain of actions, a phenomenon known as compounding uncertainty. When such systems fail, tracing the root cause or conducting causation analysis becomes extremely difficult due to opaque failure modes and information asymmetry.

The critical challenge is that AI can now fail while doing exactly what it was designed to do. Unlike software bugs or cyberattacks, these failures emerge from within due to biased training data, drifting knowledge, or complex feedback loops. Managing such behavior requires continuous, evidence-based oversight rather than static, one-off testing.

From Checklists to Continuous Monitoring

For AI systems to be insurable or trusted in safety-critical domains, they must undergo rigorous, transparent, and repeatable AI risk management. That means moving from checklist validation to continuous monitoring, where systems are tested and challenged throughout their lifecycle. This risk management framework provides the necessary evidence and controls that underwriters will demand to price the exposure.

Best practice frameworks point to four foundations, which should be viewed as future underwriting criteria:

• Governance and Tiering: Treat the whole workflow from data pipelines to prompts and APIs as the governed unit. Tier systems not just by impact but also by autonomy (how much they act without human approval) and volatility (how often components change). Every modification should trigger a change-impact review.
• Design Standards: Start from intent: what is "failure" in business or operational terms? Translate that into measurable technical metrics, justify every heuristic (prompt templates, data filters, reward models), and document assumptions and known residual risks. Build guardrails and fallback plans from day one.
• Validation Uplift: Move beyond static benchmarks. Combine domain-grounded tests with adversarial evaluation and scenario stress-testing. Measure calibration and selective prediction; use red teaming to expose hidden vulnerabilities. Where LLMs are used as judges, demand statistical checks for bias and consistency.
• Monitoring: Deploy continuous monitoring across inputs, outputs, and dependencies. Track drift, fragility, and anomalous behavior. Establish clear service-level objectives for safety and accuracy. Keep humans in the loop for escalation and design rapid rollback and patching playbooks.

In this new landscape, model probe systems for blind spots, test procedural reliability, and pressure-test entire pipelines. The goal isn't just compliance, it's resilience: building AI systems that remain safe, and trustworthy even as they evolve. Experience in managing cyber risk means insurers can build on existing practices, but tools and methods will need to be adapted to AI systems.

The Case for AI Insurance: Turning Risk into Resilience

As AI systems become more autonomous and unpredictable, they test the limits of traditional insurance models. Losses caused by AI errors often don't fit neatly into existing policy lines like cyber, product liability, or professional indemnity, therefore creating coverage uncertainty. This often results in "silent coverage," which creates hidden liabilities, unpriced exposures, and uncertainty for both insurers and insured. This unreserved, unmodeled exposure threatens aggregation events and solvency for carriers.

From our perspective, it matters less whether AI risks eventually sit within existing policy lines, emerge as embedded features, or evolve into a new, standalone class of AI insurance. What matters is that AI risks are material and growing, creating significant exposure to portfolios and businesses alike. As such, they must be rigorously understood, quantified, and managed. Businesses adopting AI will need confidence that, when failures occur, clearly defined insurance coverage stands behind the technology if they decide to transfer the risk into the market.

To make AI risk insurable, the market will need innovative tools and pricing mechanisms that reflect how AI operates:

• Performance-Based Guarantees: Policies could trigger payouts if the AI underperforms (e.g., if its accuracy or reliability drops below a defined threshold). This mechanism could be structured as an endorsement on Product Liability or a custom Financial Loss policy.
• Usage-Based Insurance: Premiums can scale with AI activity (e.g., per API call, per decision), creating dynamic, real-time pricing that mirrors exposure levels.
• Premium Differentiation (Bonus–Malus): Safer systems should cost less to insure. Firms that can demonstrate robust governance, transparent validation, and effective monitoring would pay lower premiums. In contrast, opaque or unaudited systems would be priced prohibitively high or deemed uninsurable.

This market mechanism does something regulation alone cannot: it aligns financial incentives with technical rigor. Underwriters will demand strong assurance, continuous monitoring, and clear audit trails to minimize both frequency and severity. Post-incident protocols will help to contain financial losses. Like cyber, insurers and brokers will shape the standards for testing, validation, and operational oversight. By linking AI assurance to premium levels, insurance can become a catalyst for safer, more trustworthy AI adoption, rewarding those who invest in resilience and transparency while discouraging reckless deployment.

This article first appeared on Instech.

In the U.S. financial services sector, the rising costs of traditional B2C customer acquisition are self-evident. A 2023 report indicates that the B2C customer acquisition cost (CAC) has surged by 60% during the past few years.

Turning to a B2B2C growth model through partnerships, also known as alternative distribution channels, provides an avenue for organic growth, aligning with modern consumer preferences for simplicity, digital accessibility, and trust through brand loyalty. Partnerships not only offer startups an economical path to scale but also enable legacy incumbents to enter adjacent markets quickly, as trying to do it all in a vertically integrated model is becoming increasingly challenging.

Without executing a robust partnership strategy, startups like Trust & Will may not be able to amass 400,000 users, and Oscar Health might have been nipped in the bud. Likewise, legacy firms such as State Farm and USAA could have faced much more challenges in surpassing their competition, had they opted to go solo in tackling the market.

Of the various partnership models, the distribution partnership stands out due to its complexity and the absence of a definitive playbook. It is a collaboration where one sells another's financial products, with a primary focus on expanding sales reach rather than boosting operational or technological capabilities.

The advantages of a successful distribution partnership are manifold. Such partnerships not only ensure a cost-effective go-to-market approach but also create a competitive moat as a successful partnership typically requires three to 12 months to procure, unlike B2C counterparts, which can be established much more rapidly. There is no shortage of examples of a three-person marketing team putting out more quality content daily than $100MM companies put out in a month. A competitive moat, established through distribution partnerships, is critical for ensuring longevity in a highly competitive market, especially one where startups can iterate quickly.

However, this landscape has been changing. In recent years, the introduction of new technology and data has transformed the distribution partnership model within the financial services industry. Yet, the outcomes have been mixed at best.

While distribution partnerships are common in the tech and legacy financial service spaces, emerging financial service companies have grappled with leveraging the potential of B2B2C channels. Many high-profile partnerships have fizzled out prematurely, while others remain lackluster. A significant proportion of fintech or insurtech firms that embraced the B2B2C strategy have either floundered or pivoted.

Partnerships undoubtedly present numerous advantages, such as reduced customer acquisition costs, improved customer experiences, and mutual revenue opportunities. These advantages are clear in theory, but what are the underlying causes of the notable failures in practice? Let's explore some of the frequently overlooked challenges.

Challenge 1. Confusion Over Sales Strategies

One of the primary challenges that partnership leaders face in the financial services sector is the ambiguity in distinguishing among various distribution partnership strategies.

In the scope of distribution partnerships, terms like co-selling, cross-selling, and reselling often lead to confusion due to their subtly different interpretations arising from a lack of consensus on their exact definitions.

Consider co-selling, a strategy that has recently garnered attention for its advantages for complex products or services. In essence, co-selling unlocks value via synergistic collaboration between partners, working together to meet a common customer need throughout the sales process, a challenge beyond the capabilities of a single partner alone.

Reselling is generally traditional, standard, and straightforward, mirroring a vendor-buyer relationship. Co-selling, in contrast, involves a deeper, more integrated partnership that extends beyond conventional frameworks.

Cross-selling is another common form of distribution partnership, where complementary products are offered to an existing customer base. An example of this is the bundling of travel insurance with travel bookings, a notable success in cross-selling. However, most successful cross-selling today is confined to simple and commoditized products that have short sales cycles with low contract value, which yield lower margins for both parties involved.

Additionally, it's critical to differentiate co-selling from cross-selling. Cross-selling primarily involves offering supplementary products through the existing delivery mechanism, while co-selling is about two companies working together to cater to their shared customer base through innovative, previously non-existent delivery mechanisms.

The partnership between Lemonade and SoFi also serves as an instructive example. Despite their seemingly complementary products – mortgage and homeowner's insurance – the anticipated cross-selling actually requires a co-selling approach due to the significant gaps in education, experience, and customer expectations to purchase insurance versus mortgage. Both products are complex and require a longer sales cycle. A simple embedded insurance solution could not fill such a gap. This became evident when the partnership unwound in 2023 after two years of underperformance.

Understanding these nuances is fundamental to nurturing successful partnerships in the financial services industry.

Challenge 2: Absence of a Robust B2B2C Partnership Framework

The success of a B2B2C partnership strategy is often jeopardized from the onset by misaligned expectations. Partnership leaders regularly confront frustration when stakeholders perceive the process as a simple "plug and play," rather than recognizing it as the complex collaboration process it is. Success thus hinges on developing a shared understanding and strategic alignment among all stakeholders, such as partners' leadership team, operation team, and internal teams, to ensure commitments are made and resources are effectively optimized from the get-go.

To navigate these challenges effectively, a comprehensive framework is required, one that not only addresses the overarching issues but also tackles specific problems such as:

Partnering with companies that have little customer profile or needs overlap, leading to little gains beyond initial excitement.

Engaging with channels lacking authentic sales or marketing incentives, which proves futile, despite surface-level compatibility.

Over-reliance on revenue sharing without adapting it to distinct distribution partnership models might not adequately motivate partners seeking more support for their unique challenges. Moreover, those without experience in selling specific products may not see the projected benefits as tangible. Generally, revenue sharing tends to be more effective in reselling scenarios than in other contexts.

Lack of incentive alignment, risking half-hearted commitment internally and externally, causing project delays or directional chaos.

An aligned strategic partnership framework is essential. Without it, pinpointing the root causes of B2B2C challenges becomes nearly impossible. Teams may struggle to determine if issues arise from inadequate marketing, product issues, business development missteps, or fundamentally unviable channels. Blame games may start, often leading to the partnership's dissolution. 

Take the SoFi and Lemonade partnership as an example: Despite their status as fintech and insurtech leaders in the B2C space with exceptional growth at the time, the absence of a solid partnership framework left the insurtech's business model misaligned and interests unmet, rendering the partnership a victory only on paper.

Challenge 3: Not Establishing Clear Expectations and Full Commitment

An internally aligned partnership framework and a solid business case are essential, but setting clear expectations and ensuring full commitment from motivated partners are equally important. As the saying goes, "it takes two to tango." A partnership, much like a marriage, aims to collaboratively create innovative solutions with lasting motivation. The distribution of contributions and benefits in these agreements is rarely quantitatively symmetrical, and results typically emerge over time.

Commitment and trust are the cornerstones of lasting partnerships. It's worth mentioning that partnerships differ fundamentally from typical sales interactions; there isn't a pre-packaged product awaiting a purchaser. Instead, the partnership involves a process of continuing, collaborative execution. Clear expectation setting, coupled with fostering deep commitment not just from executives but also from those stakeholders on the ground, is essential for effective execution and aligning both long-term goals and short-term resource investments. 

"Reflecting on our journey, I've seen a recurring challenge," stated Ara Agopian, CEO of SolarInsure. "While our channel partners' executives often share our excitement about the partnership, that enthusiasm doesn't always reach the teams on the ground. This mismatch is typically rooted in external market pressures and a shortfall in product knowledge and training. Historically, this misalignment of commitment has led to several failed partnerships, stemming from our own miscalculated expectations and a lack of engagement from our partners. We've since refined our partnership strategy, now ensuring the commitment requirements are clear from the outset in our contracts."

Partnerships marked by transparency, collaboration, effective communication, and robust commitment stand a much greater chance of success. Conversely, concerns over intellectual property, methods or similar trust-related issues can almost completely undermine such endeavors. Effectively, setting expectations is a critical process for both assessing and motivating partners, a step without which true commitment is rarely attainable.

The collaboration between Allstate and Nationwide is an excellent example of this. The misalignment of expectations among top executives and the underestimation of the technological integration effort contributed to its failure, highlighting how such discrepancies can threaten the success of even the most established brands.

Challenge 4: Misalignment in the Commoditization of Customer Relationships

Even in the presence of a robust partnership framework, a compelling business proposal, and clear expectations and commitment, the financial service industry faces a significant challenge: aligning the commoditization potential of various customer relationship models. The concept of "commoditization potential of customer relationships" – essentially, the ability to monetize customer relationships – is influenced by various factors. These include the length of the sales cycle, the degree of product commoditization, the expected customer lifetime, the average value of contracts, and the complexity of the product. Failure to effectively navigate these aspects can often lead to partnerships that unfortunately do not yield significant outcomes.

Strategic planning and execution are critical in partnerships, especially when there are differing perspectives and strategies on commoditizing customer interactions, even for the same client base.

The effectiveness of a partnership hinges on aligning these models. Some companies focus on long-term cycles with infrequent but high-value interactions, while others prioritize more regular engagement. Additionally, approaches to client relationship management can range from tightly controlled to more relaxed. These disparities can create substantial obstacles to successful collaboration.

For example, life insurers have persistently sought to collaborate with fiduciary RIAs to cross-sell life insurance products targeting affluent clientele. Despite numerous attempts over the years, only a few have achieved significant scale. Even with products targeting the same clientele as RIAs with concrete use cases, many insurers struggle to appropriately "commoditize" advisor and client relationships and offer the support required, which fundamentally differ from life agent-customer relationships. Such misalignment can render the partnerships ineffective.

Challenge 5: Divergent Business Models

Expanding on the previous point, a sufficiently distinct customer relationship model often implies a different business model which necessitates a deliberate effort to bridge the gap. For example, a common challenge in partnerships arises when SaaS solution providers collaborate with financial product or service vendors. Theoretically, the broad customer base of SaaS companies appears to complement perfectly with the high-value contracts typical of financial products, suggesting an ideal match. However, in reality, this combination often results in complexity rather than simplicity.

Several factors contribute to these challenges. For instance:

• Many SaaS providers lack specialized expertise or an appropriate marketplace for effectively cross-selling financial products, hindering seamless integration.
• The introduction of third-party financial products involves risk, with potential liabilities that SaaS providers are often reluctant to take on.
• The differing business models of these entities imply different operational cadence: SaaS providers usually operate rapidly, while financial product or service providers need more deliberation and understanding.
• SaaS entities generally function in a less stringent regulatory environment and may be resistant to additional legal constraints.
• Additionally, SaaS companies might not fully understand the financial nuances and the commoditization possibilities inherent in their non-SaaS counterparts, leading to misaligned expectations and strategies.

Understanding and overcoming these complexities demands a nuanced and tailored approach to partnership strategies between varying business models. When executed with precision, these strategies can lead to substantial and meaningful success.

For example: Trust & Will, a SaaS solution in digital estate planning, exemplified a B2B2C partnership strategy by partnering with digital term life insurance distributors. In a highly competitive space with low entry barriers, Trust & Will distinguished itself by embracing a B2B2C approach since 2021, one of the first among its competitors. The team developed a robust partnership framework that gained internal support. This strategy was well executed in the subsequent years. By leveraging their digital capabilities and understanding of customer relationships nuance, they crafted an incentive structure perfectly aligned with the needs of term life distributors seeking differentiation. This strategic partnership with term life distributors contributed to Trust & Will's remarkable growth, further establishing its leadership in the digital estate planning domain.

The promise of B2B2C in financial services remains strong, but success requires more than ambition and alignment on paper. It demands surgical precision in partner selection, a shared commitment to execution, and a deep understanding of the underlying dynamics, from cost recovery models to business model compatibility. As the industry continues to evolve, those who approach B2B2C partnerships with discipline, clarity, and a framework rooted in operational realism will be the ones to convert potential into durable, scalable growth.

Note:

Other success cases: SasID and NAR (National Association of Realtors), IHC specialty benefits with USAA; New York Life and AVMA, Allstate's historical expansion due to partnering with Sears Roebuck & Company, Petco and Nationwide 23.

Underwriting has always been the heart of insurance. But that heart is now beating faster. For decades, underwriters have assessed risk one policy at a time. Today, they're under mounting pressure to process increasing volumes, respond to brokers faster, and still maintain profitability.

At Send's INFUSE webinar, my fellow panelists Tom Nasso, CUO, from Falvey Insurance Group, and Dan Walsh, CUO, from Equinox, shared compelling perspectives on the drivers behind this transition. Their insights crystallized that the future of underwriting lies not in individual policies but in portfolio-level intelligence powered by data and technology, for the majority of risks.

Why Shift from Policy to Portfolio Underwriting?

Tom articulated this evolution precisely: "Portfolio underwriting and the use of data to drive underwriting is the result of the market's need to address speed, efficiency, and decision-making."

This observation captures the fundamental challenge facing underwriters today. The modern underwriting environment demands both speed and precision - traditionally competing priorities. Portfolio underwriting provides the framework to achieve both. Rather than navigating submissions in isolation, underwriters can identify trends across their entire book, monitor performance metrics in real time, and detect emerging risks before they materialize into losses.

At Aurora, we've advanced this concept by embedding algorithmic intelligence directly into the underwriting workflow. Modern technology can automate the manual steps of case underwriting - risk assessment, rules application, pricing calculations, and quote generation - enabling underwriters to focus on strategic portfolio management, broker relationships, and value-creating decisions rather than repetitive administrative tasks.

How Algorithmic Underwriting Works

During the webinar, I talked through this process in action. When a broker submission is sent in, systems can automatically extract critical details and enrich them with supplementary data points: location-specific information, exposure analysis, and peril modeling. This enriched dataset flows through rules and pricing engines, which generate consistent, auditable assessments and quotations within seconds.

This represents algorithmic underwriting - leveraging automation to handle the computational heavy lifting for complex underwriting, while empowering underwriters to excel at what truly differentiates them: cultivating relationships, optimizing portfolios, and executing informed strategic decisions that drive sustainable growth.

Insurers Are Already Seeing Significant Cost Reductions

The question of ROI naturally arises with any transformative technology. The empirical results provide a compelling answer. Carriers implementing algorithmic underwriting solutions are realizing significant expense reductions in trading operations.

That level of efficiency doesn't just cut costs - it fundamentally unlocks organizational capacity. You no longer need hundreds of underwriters to manage growth; you need skilled professionals focused on high-value decisions, supported by technology that enables scalable growth.

But efficiency is only part of the story. The real value lies in improving performance. When underwriters maintain real-time, granular visibility into portfolio performance, they can identify emerging patterns and intervene before trends deteriorate into losses. This operational agility transforms automation from a back-office efficiency tool into a genuine, profitable growth catalyst.

Legacy Systems: Still the Barrier to 'Smart Underwriting'

Transformation and change, however, is rarely frictionless. Both Tom and Dan highlighted persistent challenges: entrenched legacy systems, inconsistent data, and the cultural resistance that accompanies new technology.

At Aurora, we encounter these barriers regularly. Many carriers want to move forward but hesitate, concerned that their data isn't ready. My counsel is simple: start anyway. Begin laying the foundation. Use technology and AI to clean, structure, and organize your existing data assets. Simultaneously, establish robust processes to capture new data from inception - ensuring it's clean, structured, vast, and granular. Even if it isn't perfect, the progress you make today will determine how effectively you can use automation tomorrow.

And then there's culture. As Dan said during the session, this isn't just about tools - it's fundamentally about trust and organizational alignment. Underwriters and brokers need to experience automation as an enabler rather than threat. Success requires cultivating confidence that technology augments human judgment rather than replacing it.

Looking Ahead

Underwriting has always existed at the intersection of art and science. What's changing now is the scale and speed at which the science amplifies the art.

Automation, data, and algorithmic systems are enabling underwriters to make better decisions, faster, while freeing them to focus on the human aspects that have always mattered most: relationships, intuitive judgement, and experience.

The next chapter of underwriting isn't characterized by machines taking over. It's defined by human-technology collaboration that makes underwriting demonstrably smarter, faster, and more efficient - for underwriters, brokers, and clients alike.