At the recent International Insurance Society annual meeting, panelists laid out some intriguing thoughts on insurtech, saying that the first wave is ending and describing the outlines of the second wave that has already begun to form.
The key takeaways: 1) The industry is focusing too much on what panelists called “the shiny objects,” like artificial intelligence, and not enough on the issues that really matter to customers and to corporate bottom lines; and 2) incumbents will bear much of the responsibility for whatever success the second wave has, depending on how well they retool their mindsets and their processes to absorb the steady stream of innovations that smaller players will provide.
The best image: the mullet. That’s how Jamie Yoder, president of Snapsheet, characterized innovation in insurance to date. “There has been so much attention paid to the front end [basically, sales],” he said, “but there’s a huge opportunity to clean up the back.”
In addition to Yoder, whose young company digitizes and automates the claims process in personal lines, the panel consisted of: 1) Chris Cheatham, who founded a startup, RiskGenius, that uses AI to facilitate review of policies and that was recently acquired by a bigger startup, Bold Penguin; and 2) Jeff Berezny, SVP and general manager of enterprise products at Trov, which provides a digital technology platform for incumbents. The panel was moderated by James Maudslay, global head of insurance at Equinix, which operates a huge network of data centers. (You can watch the entire session here.)
Yoder argued forcefully that much of the focus for innovation is on the wrong issues. He said that the sexy part of what Snapsheet does is the computer vision that can analyze photos of damage in a car crash and use artificial intelligence to provide an estimate of the cost of repairs. But he said computer vision saves only maybe 15 minutes on estimating — in a claims process that can languish for days. He said he can get from a first notice of loss to a settlement in 30 minutes, but the computer vision only works in 1% of the cases. Why not, Yoder asked, focus on the big delays that happen as the insurers for those involved in a crash gather and exchange data, sort out responsibility and get money to clients so they can get their cars repaired?
He said the secret to innovation isn’t to have AI do the estimate. “Al will do,” Yoder said. [Lest his joke doesn’t come across in print: He’s saying that Al, as in someone perhaps named Albert, could do the work that everyone seems to want to assign to artificial intelligence and still allow for considerable innovation.]
“The real issue is all the handoffs,” Yoder said. The solution is to “have the claim be the captain of the process. It’s determining, what can I do myself? It’s assigning tasks to vendors, updating the customer, getting an issue to a person when it needs a person.” At the moment, he said, claims processes are managed “based on massive diaries and to-do lists, which don’t really foster efficient throughput of work. And this is in a piece of the business that is 70% of the cost.”
He said it’s also crucial to focus on all parts of the process, because so many can yield improvement. “We’d do all this work with carriers to speed their claims, then find they were still cutting paper checks,” Yoder said. “It’s like you decide to drink a diet Coke but then order a Big Mac to go with it.”
Berezny agreed that the beginning of the claims process “is just phone tag. Call me back, call me back, call me back….” So, having AI that manages the process, to reduce the administrative load on people, and that speeds the handoffs would matter much more than AI that prepares an estimate.
He said the second wave of insurtech that is taking shape now will be more about “selling picks and shovels” and not about the “gold rush” that is being capped off now with the huge valuations that Lemonade and Root achieved following their initial public offerings.
But he said the incumbents have a lot of work to do to prepare. While being cloud-based and digital are table stakes for being able to innovate these days, Berezny said, “Insurance just isn’t there yet.”
Yoder said incumbents need to develop modular processes, based on application programming interfaces (APIs), so companies can have “plug-and-play innovation.” Are information systems designed so “you can just plug your innovation in, or are you stuck with a new wave of IT integration projects every time” you see an insurtech whose software could improve your business? “Innovation isn’t about a point in time. This is a journey.”
He said the need to be able integrate innovation quickly is all the more important because so much is becoming available. Yoder, who until two years ago ran the insurance practice at PwC, said clients used to need to invent a host of capabilities to be able to pursue a digital strategy. “Now,” he said, “you can just pull together a list of insurtechs with those capabilities, and there you go. You should get there a lot faster and for a lot less money.”
Berezny said he sees the new wave of insurtechs leading to “not just direct but embedded” insurance. As a result, he said Trov is working with mortgage groups to embed home insurance in their sales processes, with new digital banks to offer insurance as part of their portfolios and with affinity groups to include insurance in their dealings with members.
“Insurance needs to meet people where they are,” Berezny said. “You can’t just assume they’ll come to your great, shiny website.”
Yoder agreed that embedded insurance has great promise. “People sometimes look for innovations that are way beyond what anyone ever thought of,” he said, “but sometimes the opportunities for innovation are staring you in the face.”
Cheatham said that, with “the first wave of insurtech coming to an end,” he hears a lot about how AI will supposedly eliminate human jobs. But he said the second wave will be about “augmenting humans, not replacing them.” In commercial insurance, where RiskGenius and Bold Penguin focus, “There are too many nuances for us to think that we’re going to get rid of humans in any part of the process.”
He said that innovation in commercial lines moves more slowly than in personal lines, where Trov and Snapsheet live. But he said, “We’ve shrunk processes from weeks down to days, and we can go so much further. I see a big acceleration going forward.
“We’ll catch up to you other guys soon enough.”
P.S. Here are the six articles I’d like to highlight from the past week:
The plague that is coronavirus continues to threaten our way of life and our very existence.
And while we wait for our best scientists and Big Pharma to develop and deliver the vaccine we so desperately need to begin the return to some semblance of our prior lives, it is notable that finance and insurance – the nation’s third-largest industry – is coping better than most would have ever imagined, thanks to the technologies that have been gradually but surely emerging for at least the past two decades.
Now, with few if any acceptable choices, these technologies and the benefits they offer participants in the insurance ecosystem are being more eagerly embraced and adopted by insurers in the battle they are waging for their survival. Artificial intelligence, in all its manifestations, drives many solutions. It helps us transform the digital rivers of data flowing between us and our many connected devices and facilitates effective real-time decision making. As noted on April 30 by Microsoft CEO Satya Nadella, “We saw 2 years of digital transformation in two months.”
WFH (work from home) is a perfect example and, defined more broadly, encompasses video conferencing, distance learning, educational and training sessions and internal business communications platforms. Platforms such as Zoom and Slack are well-known solutions, but dozens of similar services have proliferated. Artificial intelligence powers all of them and allows for total individualization of communication preferences. Insurers have been able to leverage them in operations, and in a matter of just weeks have redeployed thousands of inside employees from corporate buildings to their own homes, hardly missing a beat. In many cases, customer experience has actually improved as distributed workflow algorithms ensure better matching of customer profiles with staff expertise and skill sets, while also reducing the associated potential risk of litigation.
The claims process has become another major beneficiary of artificial intelligence and other new technologies, including computer vision, machine learning, digital customer self-service and electronic claim payments. Contactless (or touchless) claims is now a reality, with some large insurers reporting up to 50% use of this method across all claims.
In a world where claims are submitted and paid without physical inspection and validation, fraudulent claims detection becomes an even higher priority, and here again solutions incorporating artificial intelligence, predictive analytics and computer vision are being deployed effectively. The recent partnership announced between Shift Technologies and Snapsheet is a perfect example, and there are many more in the works and underway.
Analytics also plays a critical role in streamlining claims settlement, which is recognized as a primary driver of customer loyalty and retention. This includes the current rapid adoption of digital claims payments, which speeds up the process while removing significant processing costs for carriers and getting money owed more quickly to customers; for many, cash flow has become more important than ever. The speedy settlement of claims has previously meant higher costs and a risk of overpayment, especially in high-volume catastrophes but by employing AI technologies these risks are now being mitigated.
Here is a partial list of the many additional ways in which AI and technology will play an increasingly critical role;
a solid working knowledge of AI and technology will become a prerequisite for even entry-level employees and most certainly for organizational advancement
the nature of risk has changed, and AI allows carriers to respond in real time (e.g., several large life insurers ceased issuing new policies on those over 75 almost immediately after COVID-19 altered mortality rate tables
data allows management to navigate change, and AI is an important tool in generating operational road maps
remote work forces present new management challenges in maintaining morale, motivation and corporate culture; AI enables automated monitoring and appropriate development and testing at the individual level
in a world of limited resources, partnerships will become more critical, and, for those that incorporate AI, relevant corporate expertise will be mandatory
continuous learning and technology development will help meet carriers’ long-term needs and goals
In an impressive example of the power and potential for technology in the insurance industry, conference production companies have quickly pivoted scheduled live events to virtual alternatives. (Insurance Nexus by Reuters Events has reimagined its “Insurance AI and Innovative Tech USA” conference scheduled for Chicago and is presenting it as a free, conference and meeting platform on May 27-28. The conference features over 30 speakers, 20 case studies, thought leadership discussions and fireside chats with key insurance decision makers, interactive sessions with live panels, Q&A and polls. A digital exhibition will feature over 20 solution providers and a one-on-one meeting service will connect participants sharing common interests. You can register here.)
As much as we all wish coronavirus had never happened, it has supercharged innovation in the insurance industry. Some changes are welcome, many of which are likely to be permanent, and which will benefit all stakeholders in this critically important industry.
We are in the midst of a bionic revolution in all industries and especially insurance. Organizations are competing with a new mix of people and machines doing both cognitive and physical work, and those companies that become bionic fastest are winning because of superior customer experience and economics. This has led to a new Digital Darwinism — the evolution of businesses and business models that not only occurs across generations as it does in the natural world, but rapidly and many times within the lifetime of a single organization.
The speed of evolution creates imbalances — and in the insurance industry we see a sophisticated sales/distribution experience with an old-style issuance, service and claims experience. Until very recently, insurtech has been a story of front-end evolution. It is only now that enterprise investments in insurtech have matched the front-end.
How do you increase the clock-speed of your evolution? Our lens at Snapsheet is to collectively reimagine the operating model, operational capabilities and customer experience with five bionic design principles in mind. A design principle is a practical concept that helps executives imbue the organization with the right way to think about critical tradeoffs and solutions. For example, Steve Jobs hated styluses, so the iPhone created a user interface that did not need any type of stylus. This simple concept drove many actions in Apple.
Bionic Design Principle 1: Change the way you work, and change the way the work is managed: Reapportion cognition among people and machines.
The fundamental activity of the industrial revolution has been to take thinking and labor tasks “out” of people and put them “into” machines. It’s a one-way street optimized around one process, one answer, one method. In contrast, in the bionic age, everything has the potential to be smart. Processes can be agile; claims can be alive and know their own path and what data they need to complete their task. Claims can even be patient, knowing how long information requests should take and when it’s time to “ask again” for missing information. A claim can even route itself to the optimal process and resources using its understanding of its complexity, service level status, skills needed and available, licensing requirements or other attributes; for example, a claim can know if it’s a total loss auto claim and therefore has a shorter, faster process path.
This dynamic allocation of cognitive responsibilities is the most important design principle because its application allows for the entire claims ecosystem (or any ecosystem) to be better without having to be in full control of every step. For example, to implement our smart claim approach in auto, we do not require customers to choose the way they engaged with us or others. Rather, we allow the customer to choose and change the interaction mechanisms, and the claim learns how the specific customer or repair shop interacts and how long responses take. The claim acts accordingly. The process is not industrial and stringent, but bionic and adaptive.
When cognition is reapportioned, no time is lost, and actions are optimized against capacity—and as the system gets smarter it can evolve the process and support higher and higher level tasks. The design itself does not have to be all-knowing right out of the gate. It masters a specific challenge and evolves over time because it’s a living system, not a mechanistic one.
Advanced weapons systems in fighter aircraft have shared thinking responsibilities with pilots for many years. Sometimes the pilot is flying the plane, and the weapons system is protecting the plane; other times, the plane is flying itself, and the pilot is focusing on the weapons systems. This dynamic allocation of cognition means that the most capable thinker—person or machine—will be in charge based on context and competence, not rigid process. Think of claims as thinking for themselves: Each claim also knows when to “ask for help” and push the activity to a human when the claim lacks enough data, structure and rules to proceed. This adaptability enables productive integration of the complex, digitally enabled ecosystems that all insurers face. Put another way, it’s not just machine learning, it’s a learning machine.
Bionic Design Principle 2: Change the way you engage with the customer and change the way you work with others: Remove friction from the ecosystem—not just the customer.
Customers no longer tolerate friction. They want service anywhere, anytime, any way — and expect to have instantaneous status about every step of the process. In claims, the old way was to send out a person to look at the vehicle or visit the home. But if you’re creative, you can change the flow. We began Snapsheet by inverting the process—enabling customers to use their smartphones to send in photos of their accident so they could poll repair shops to deliver virtual estimates, instead of having a person visit the scene of the accident—thus removing a friction point and saving time. But as we implemented this innovation, we saw that we needed to do more. Some customers were not confident in their ability to take the right photos or did not want to use an “app”—so we had to enable omnichannel capture, allowing the customer to begin the process in any way the customer found faster or easier.
We’ve been chasing friction removal for seven years, and we expect to continue chasing it. It’s easy to say that “it’s not just about the technology,” but turning that observation into less friction for the customer takes a willingness to learn, adjust, reintegrate and sometimes reinvent major functions across the ecosystem.
Our ability to field smart claims that are aware of their context enables us to see how to close the gap between the desire and realization that happens in many places in the claims ecosystem. By looking at the whole ecosystem, not just the individual task, you can see entire areas of opportunity open up.
For example, we found that making payments—more than 75% of them still transacted by paper check—was an area we could redesign. Now we have the ability to provide payment and treasury functions instantaneously for all parties involved as soon as a step in the process—such as repair—is verified as complete. We asked participants in the ecosystem, “Do you want to get paid fast or slow?” No surprise, most customers say, “Fast!”
Another significant delay often comes from processing a total loss situation, which is often slowed by the complexity of issues having to do with identifying the lien holder of the asset and the title procurement process in a specific geography. This complexity requires a carrier to change the way it engages with other suppliers. In this case, working with other industry partners, Snapsheet is enabling faster exchange of information across multiple parties, initiating insurance and salvage workflows simultaneously and integrating new capabilities to remove that friction. In doing so, it’s possible to remove days, and even weeks from the process. This accelerates the time to sell the vehicle for the insurer and time to settlement for the customer.
Bionic Design Principle 3: Change the way you change: Innovate, don’t renovate—don’t start from scratch.
The most critical part of digital evolution is to change how you change. The idea of continuous improvement has been with us for many years—at least since the 1950s when Edward Deming led the quality movement. Most insurance companies have done an analysis of what it takes to redo the entire technology stack—and it is often an absurd number with millions and millions of dollars at risk for an uncertain payoff.
Fortunately, there are many new tools to help enable integration of legacy and emerging systems. The suites of robotic process automation tools help to provide what we call value-driven integration. When companies use these products or other integration/automation tools, they can often integrate multiple legacy systems with a secure container that can interface with multiple systems and then present the results on any device—providing new results, without the need for complete system overhauls or upgrades. This creates productive modules—of process, task and technology—together, which enables improvement with zero down time while delivering continuous enhancements.
With new API architectures, one can partition the work and make it more object-oriented, which enables more flexible task and process flow design within the organization and outside to suppliers. Properly used, this creates a much more agile and adaptable learning environment.
This can take time out of customer service, sales or other interactions by collapsing many dozens of steps and multiple screens over hours, to one screen and minutes, while kicking out a clean stream of audited data for machine learning to drive future improvement. This capture of cognitive capital drives near-term labor productivity and superior customer experience while providing a data asset for further improvement. Especially in the middle and back office, there are often greater opportunities to drive customer experience and superior economics.
Again, in the bionic age, the right way to think of it is as smart, evolving, alive systems. The most important design decisions are about the application program interfaces (APIs) and the modularity of the system. For example, Jeff Bezos declared that Amazon would use common APIs back in the early 2000s, which meant that the company could add or subtract functionality without interfering with the existing process and code base of the organization. This means the company can handle complexity at a much lower unit cost because complexity is contained within the modules. This is not true of many of his competitors, and you only need to go into Macy’s and try to find your order history with the store, to show the lack of APIs in that organization.
The innovate-not-renovate point of view relies on modular thinking. This may seem incremental to some, but some of the most important innovations in human history have been integration technologies that enable established ecosystems to interact productively. The international currency markets are such an innovation; so is the internet, whose very name states its “inter-networking” mission. In the bionic world the biggest bang for the buck is usually in innovating and integrating existing ecosystems, not starting from scratch.
Bionic Design Principle 4: Change the speed and efficacy of every decision:Use AI to pick up the trash and explore the stars
Evolution is often a process of simple and complex improvement. The fourth design principle is to use AI both for the low and the high end.
On the low end, AI helps to automate faster because it can gather new information and structure it faster. There are many mundane tasks that AI can help to improve. Proper analysis of photographs is a great example, and we have developed more and more capability to make sure we have the right angle and proper picture needed to move the claim forward in the process. Even simple uses of AI such as this can add up to vast improvements in speed and productivity.
On the high end, we are developing complex models that can continue to automate and refine estimates for more and more severe accidents. This other use of AI is like a telescope for the mind—allowing new insights into vast data sets and complex problems that would not be possible without it. We analyze millions and millions of accident pictures that feed a super-smart estimating engine and leverage our estimator talent eight to one compared with non-assisted, experienced estimators. We are also finding interesting patterns in repair shops, types of accidents, customer behavior etc. If we did not have the massive and growing scale of transactions, we would not have seen and been able to share these insights.
Bionic Design Principle 5: Change your economics:Take advantage of bionic economics— speed, scale & capital
Perhaps the most powerful force of Digital Darwinism is the speed of change in core economics. This is why our final key design principle is to use the business system with the best total economics, driven by speed, scale and new forms of capital. There is no question that this new mix of people and machines can make decisions faster. We’ve found that our staff are as much as five times as productive, but they are not super humans; they are simply people supported by great cognitive technology and training. The more data and transactions that any learning machine can ingest, the better its speed and accuracy.
Google Translate is so good because it was fed with billions of books and all the proceedings of the E.U., which publishes in about three dozen languages. This vast wealth of data made it learn faster. So, too, with claims. This speed not only decreases labor, it also allows us to provide answers faster and get through the entire process in less time. In claims, like many things in life, time only makes things worse— with impatient claimants creating additional service demands and the organization incurring added costs such as rental and storage.
In terms of scale, larger networks of computer power usually have superior economics. The cloud not only enables organizations to make their costs more variable, (e.g., if you need more capacity, simply contract for more without the need to buy new hardware and software) but also provides the ability to take advantage of the economics that support that network—such as the formidable buying power of the cloud provider, power efficiency, customized operating systems and software that creates other efficiencies. There are many end points on the network—another scale effect. Why do people usually go to Google? Because its network of links is larger and better than the competition. Why go to the second-best network? Likewise, cloud providers of critical business functions can create a world-scale network that is larger and better than all but the largest individual competitors.
Lastly, the bionic competitors gather and grow three types of capital: behavioral, cognitive and network.
Behavioral capital is the ability to track, analyze and model the behavior of any customer, device or service provider in the ecosystem. United Rental has over 70% of its assets—going to 100%—linked so the company can see the location, behavior and maintenance status of all assets. This is behavioral capital.
Cognitive capital is the store of AI, algorithms and other automated knowledge that can be used, improved and reused again and again. The better the store of cognitive capital, the more organizations can leverage existing labor and assets.
Network capital is how well a company is tied into the relevant networks for customer and supplier access. For example, no consumer products company can afford to not have a strategy for Amazon and Alibaba because they have some of the most extensive network capital in the world for consumer products. In bionic competition, organizations need to strategically use cloud providers that can bring these new forms of capital to bear to create superior economics.
There you have it. Five design principles that you can add to your strategy to move beyond siloed activities and on the way to deep transformation— speeding up your evolutionary process to meet the needs of a new Digital Darwinism and updating your customer’s experience along the way.
While every carrier manages claims operations in a slightly different way, there are three consistent technology setups currently in practice: Green Screen, Home-Grown and Modern. The back-end operational workflows for each of these practices are generally the same: The adjuster manually enters notes, manually sends emails or makes calls and manually ties documents from the document management systems to the claim systems. The challenge here is that the adjuster is the centrally intelligent component. Relying on an adjuster to connect various systems mires the adjuster in overly manual steps, leaving claims processing vulnerable to reduced speed, mistakes and inefficiencies – all of which lessen customer satisfaction.
While more common overseas and in smaller markets, green screen systems are still found in many claims operations today. The green screen is a simple claim database that only accepts user inputs from a text-based screen with minimal capabilities to integrate into any other systems. Adjusters are forced to use a separate document management system to store files and photos and use a separate email system for outward communications.
Carriers relying on green screen systems see inefficiency with data transfer. Adjusters have to hunt for documents that are not tied to a claim number, annotate the decisions they have made in the green screen system and communicate in a separate system to the customer. Most of the mindshare of the organization is spent on teaching the humans the rules of the claim and how to document their thoughts in the system.
Some organizations have managed to build their own systems internally over the years. In these systems, various IT projects over the years have been spliced together with complicated business rules that aim to reduce the human error and ensure legal compliance. Carriers with a home-grown system face significant IT spending to maintain their complex infrastructure. Even with a large IT staff, it is nearly impossible to launch new technology initiatives because change affects rules buried deep in the system. The result is a system that is expensive, inflexible, complex and generally oblivious to the customer experience.
Recently, carriers have consolidated their legacy systems into one modern platform. These setups require a large engagement with a third-party system integrator and many years of thoughtful planning and data migration. However, the output is rarely a truly consolidated system. Carriers with modern systems are bound to long-term, third-party support contracts and face many of the issues that home-grown carriers face. Complicated business logic is embedded in the software to try to avoid human errors, but it leads to complexity and rigidity that ensure internal compliance while ignoring the customer experience.
Carriers and Customers
As customer needs are changing, carriers’ technology should be changing, too. Today’s customers expect a seamless tech experience with clear communication, automation and the ability to input via apps, photos, phones and inboxes. There are several new tech solutions that aim to ease a challenge of current carrier tech configurations. At Snapsheet, we have already built software that eases nearly all of these customer expectations.
Here are the capabilities that are critical to advanced claims technology – all of which will help meet customer needs:
Cloud-Based Architecture: This feature is important for a flexible design, which eases the implementation. There is no data migration, no system integration and no multi-year project plan. Claims software is launched stand-alone around existing systems or as a full-on replacement. It enables carriers to track, with real-time precision, all of the customer interactions, how the customer engages with the claims process and how the adjuster is engaging with the customer. Immediate insights are gained and can be operationalized.
Intelligent Claims Files: Instead of relying on the adjuster to tie systems together and shepherd the customer through the claims process, the Snapsheet platform has advanced capabilities that understand the expectations of each step in the claims process and guide the customer through the appropriate actions. An intelligent engine coordinates the communications and documentation needs for each file and advises the adjuster when to take action. If all of the requested information is provided, the engine may choose to automatically move the work to the next stage.
Real-time metrics and operational transparency: It enables the carriers to track, with real-time precision, all of the customer interactions, how the customer engages with the claims process, and how the adjuster is engaging with the customer. Immediate insights are gained and can be operationalized. The result is an enhanced customer claims experience, led by automation and real-time customer engagement to provide a tailored journey through any claim in any language in any country.
Customized roll-out: Customization is key. Even with a single consistent platform, such as Snapsheet’s, it is important to customize implementation for whatever legacy IT configuration exists. This adds flexibility and ease-of-use to each project. Snapsheet’s recent strategic collaboration with Zurich is an example of taking a new software approach by putting the customer experience first. Various county entities in Zurich use each of the three software setups mentioned above. Snapsheet software can be leveraged across any configuration, activating software modules that smooth or plug efficiency gaps in the current process, or completely replace existing claims systems.
As we kick off 2019 and insurtech continues to expand, the industry will see even greater advancement in the technology space for carriers and claims processes. Automated systems are important to guide the customer through the correct claims journey and ultimately allow carriers more time to innovate.
Jamie Yoder, president of Snapsheet, talks about how his career helping advise insurers on how technology and information impacts their strategies has positioned him for a new role in putting that theory into practice to help Snapsheet serve its customers.