The tech giant’s broad influence drives the company’s multipronged insurance play.
By Russ Banham
At the large global insurer and reinsurer AXA XL, Henna Karna earned plaudits across the industry as a visionary chief data officer. In November 2020, Google tapped Karna to craft its strategy for the insurance sector, whereby insurers, reinsurers, brokerages and other industry players would partner with the technology giant, using its innovative products and digital services to better predict and prevent insured losses.
One year later, Karna says, she and a team of Google data scientists and engineers are in full partnership mode, collaborating with dozens of leading insurance enterprises and software companies, among them Munich Re, Allianz, MetLife and Applied Systems. While these relationships are in their formative stages, Karna says she expects the partners’ use of Google’s cloud computing, AI, machine learning, advanced analytics, geospatial mapping, unparalleled search capabilities, and other cutting-edge tools will transform the global insurance value chain.
“Insurance is about hedging bets, pricing things now for a claim that may not occur for several years, which makes it really a data game,” Karna says. “That’s Google’s DNA. What we became known for is connecting information. Google is poised to do the same for insurance, connecting information to be purposeful for the business decisions made in the industry.”
Karna earned both a master’s degree and a doctorate in applied mathematics from the University of Massachusetts and an MBA from MIT. And she’s chalked up nearly 25 years of experience leading data initiatives at insurers AIG and AXA XL, where she implemented an award-winning data ecosystem called DEEP. Thus, Karna is no stranger to disruptive innovations or making good on her bold pronouncements. What distinguishes her efforts is their mission-driven nature. As she says, “I have an ambition to help humanity through technological advancements.”
Her personal goals are in alignment with Google’s corporate mission to “organize the world’s information and make it universally accessible and useful.” As the leader and general manager of Google’s Global Insurance & Risk Management Industry Solutions initiative, Karna says she is focused on making all risks better understood, more manageable and “dynamically priced to fluctuating exposures.”
A case in point is the opaque risks of climate change, cyber security, and pandemics and other infectious diseases, the three top risks identified by global insurance and reinsurance experts and the general public surveyed for the AXA Future Risks Report 2021. “Using Google’s technology, together we can unlock the data that resides inside the four walls of insurers, reinsurers and brokers and pull it all together to better predict and prevent losses to insure them at the right price points,” Karna says.
While other cloud platforms like Amazon’s AWS and Microsoft’s Azure compete with Google Cloud in this sphere, Google Cloud appears to have gained the most traction to date. Multiple email exchanges to secure interviews with leaders at Amazon and Microsoft who are focused on the insurance sector proved fruitless.
The tech giants are not alone, of course, in eyeing ways to enhance industry underwriting and pricing efficiencies and capabilities, with a galaxy of smaller insurtech firms formed exclusively for this purpose. Moreover, many insurers, reinsurers and brokerages have recruited and hired teams of data scientists and engineers to do the same work.
But where Google may truly be a game changer is the matchless depth of its technological skill sets and tools. “They bring to the table the smartest computer scientists in the world, not to mention billions and billions of dollars invested over the years in creating the most innovative and powerful technology products,” says Taylor Rhodes, CEO of Applied Systems, a global cloud-based provider of insurance software and agency management systems.
Epic + Google
These attributes led Applied Systems in February 2021 to become one of Google’s early partners. Aside from Google Cloud’s advanced AI, machine learning and other data-science solutions, Applied Systems is leveraging the tech company’s BigQuery database to help its agency and broker customers gain greater access and insights into the data that resides within their management systems. The company also is using Google Apigee, an API (application programming interface) platform that enhances the exchange of data across third-party applications. “The partnership is all about combining Google Cloud’s technologies with our own deep insurance technology expertise to enable agencies and brokers to serve their clients better, at a time when the cost of insurance is going up,” Rhodes says.
By next year, Applied Systems’ flagship agency management system, Epic, and its portfolio of analytics products, will be running inside Google Cloud. “This matters a great deal,” Rhodes says, “since we won’t have to build everything that Google has already built. Our clients will be able to access their customers’ data [via Apigee] and apply Google’s advanced AI and machine learning tools to this data, turning it into insights that solve business problems, spot loss trends, and lead to the development of new products.”
He called the partnership a “transformative” event in Applied Systems’ history, explaining that it enhances the ability of its customers—mostly midsize and smaller brokerages and agencies—to become more valuable providers of advice, while also positioning them to negotiate with the insurance markets more advantageously.
“Our industry has let data run through our fingers, with data stuck in emails, texts, systems and silos,” Rhodes explains. “We haven’t invested as much as we need to invest in connecting brokers, carriers, reinsurers and the rest of the value chain to collect and analyze data. Google, on the other hand, has made billions of dollars of investments in the world’s best computer scientists to create the world’s most powerful computing products.”
Since these products already exist, it obviates the need for brokers and agencies to reinvent the wheel. Indeed, no one carrier, reinsurer or brokerage, much less midsize and smaller MGAs and agencies, has invested at the scale that Google has invested. In the current war for rarefied technology skill sets, they’re also unlikely to entice the best computer scientists to work for them.
That puts Google in an enviable position to help its partners better understand and serve clients, create new revenue streams, and solve intractable business problems, he says. “Brokers right now are confronted with advising clients on unpredictable and volatile risks like climate change, cyber security and future pandemics,” Rhodes explains. “By sharing data with a partner like Google Cloud, there’s an opportunity to improve the modeling, management and pricing of these risks. Coupled with the skill, expertise and passion of someone like Dr. Karna, we’re excited about seizing these opportunities.”
Confronting Natural Disasters Building by Building
In addition to helping brokers provide increased insights and value to their clients, Google is working to help improve underwriting and risk management. In the first half of 2021, insured losses globally from extreme weather events, like the February polar vortex in Texas, reached a 10-year high, eclipsing $42 billion, according to Swiss Re Institute. In the third quarter, another atypical weather event, Hurricane Ida, hit Louisiana as the second-most damaging and intense hurricane to make landfall in the state. It then traveled to New York and New Jersey to cause flash flooding and multiple deaths, contributing to aggregate insured catastrophe losses of $48.5 billion, blowing through the first-half figures.
Google is partnering with companies like Cape Analytics to better predict and prevent property damage and business income interruption losses. Cape Analytics uses Google’s geospatial imagery and deep learning algorithms to offer insurers more comprehensive and near-real-time intelligence on a building-by-building basis for underwriting purposes.
“Our whole business is built around the notion that insurance carriers have historically used a traditional actuarial mindset in evaluating and pricing property risks,” says Ryan Kottenstette, Cape Analytics co-founder and CEO. “However, this mindset is based on a discrete list of risk drivers that may no longer hold true.”
Kottenstette provides the example of a corporate office campus located in the wildland-urban interface, the transitional zone between unoccupied wilderness and developed land. “In determining the replacement cost of the structure, a property underwriter generally takes into account the square footage and the quality grade of the building materials,” he says. “That’s the old methodology. What’s missing is the follow-up—what needs to be done and can be done [after the insurance policy is bound] to prevent the risk of wildfire damage and destruction.”
Rarely will the insurance broker, or the carrier insuring the office structure for the past few years, return to pay a visit to the campus, Kottenstette says. The property policy is renewed without further appraisal or with just a cursory assessment. “Meanwhile, the building’s ‘defensible space’—the buffer between the building and trees, shrubs and other combustible materials—is likely to be very different as the years progress,” he says. “The same can be said for a multifamily housing development in Florida, as the landscape continually evolves for both natural and man-made reasons.”
With Google’s tools and capabilities, Cape Analytics has built a series of computer models “turning raw information into risk information” like a hail risk score, wildfire risk score or a roof quality score, Kottenstette says. Through a real-time API download, the risk information is provided to its carrier customers for underwriting and pricing decisions. “This is all about turning data that generally is not utilized or is underutilized into a set of risk predictors that simultaneously provide insights on loss prevention,” he says.
Cyber Protection in Google’s Cloud
Google also is positioning to help its partners address another one of the biggest risks confronting the industry—cyber security. In the first half of 2021, the number of ransomware attacks against corporations increased 288%, according to an analysis by NCC Group, on top of a steady uptick since 2018. With ransom payments in the six- and seven-figure range, and business interruption costs five to 10 times higher than the cost of these extortion demands, the financial impact has tipped many cyber insurance providers into the red on the line.
As several insurers exited the market over the past two years, aggregate cyber insurance and reinsurance capacity has shriveled. “Brokers are facing enormous challenges in terms of getting capacity,” says Scott Sayce, global head of cyber at Allianz’s Global Corporate And Specialty insurance solutions and consulting organization. “The consequences have been severe for many risk managers and other commercial buyers,” Sayce says.
Indeed, it is common for a $10 million cyber insurance policy at renewal to have a 50% sublimit, meaning that only half the financial limit is available to pay off a ransomware claim. Many insurers still providing cyber insurance have separated ransomware from the main policy and made it a distinct class of risk requiring co-insurance and substantial sublimits.
Premium increases in the 100% to 200% range are customary, assuming a broker can extract a quote from the markets. Policies often include exclusionary language related to the insured’s use of certain email platforms and software types considered vulnerable to a cyber incident. In turn, this usage implies the business has substandard cyber security and related controls. If the issues are not remediated by policy renewal time, the insured may be forced to “go bare”—without cyber insurance.
Allianz has partnered with Google to overcome these challenges. “Most of our clients having migrated their applications, workflows and data into a cloud platform environment,” Sayce explains. “Google can help us manage these risks in the cloud through the cloud. We can provide solutions where our customers’ cyber exposures actually reside.”
He’s referring to a unique three-way partnership with Google that Allianz and Munich Re announced in March, a collaboration they call the Risk Protection Program. The overarching goal is to provide granular levels of cyber-security information on an insured using Google Cloud’s platform.
The program consists of two parts: a diagnostic tool called Risk Manager, enabling Google Cloud customers to measure and manage their cyber risks and receive a report on their level of security; and a new “specialized and more comprehensive” cyber insurance product called Cloud Protection + for these companies, says Tom Kang, who heads Allianz’s North American cyber practice.
“Using the Risk Manager tool, a Google Cloud customer seeking cyber insurance sends its diagnostic report to us,” Kang explains. “We then leverage the report in our discussions with brokers to assess the customer’s security posture, which determines their eligibility for Cloud Protection +. By having clearer visibility into the customer’s cyber controls, it enables us to better understand and predict the risks, resulting in better coverage and price.”
Since the diagnostic report details security vulnerabilities, it further suggests ways for insureds to prevent cyber-related losses. “We’re getting a view of an insured’s cyber risks through Google Cloud that otherwise would not be possible,” Sayce says.
The More Partners, The Better The Data
Google believes its insurance partnerships can only add to the community’s ability to address multiple systemic risks. “We see insurance not as a set of discrete products as much as an ecosystem that can be optimized by partnering brokers, carriers, reinsurers, agents and insurtechs,” Karna says. “Historically, we’ve gone through complex exposures like pandemics; there are parts and pieces of this data that we can draw from in assessing the myriad risks.
“But this is more than ‘let’s insure the next COVID.’ What we are looking to achieve and have already achieved with our partners is encouraging them to lean on us as an advisor when existential issues occur. This is the power of an ecosystem, the power of the network effect.”
In other words, the more industry players that partner with Google Cloud, the more use cases they bring to the table, increasing the collective value of these contributions to the benefit of all parties. “An insurance ecosystem accelerates the strength and capabilities of what this industry can do to reduce the protection gap we have in areas like climate change, cyber risks, and overall human health and wellness,” Karna says. “That’s our mission.”
Will an ecosystem require a reassessment of historical competition in the industry? “Perhaps, and that’s a reasonable thought,” Karna says. “Somewhere along the line, competition in the industry became more about pricing and less about what could be delivered to the customer of greater value. If you look at every part of the industry, the carriers, reinsurers, brokers, MGAs and so on, they all aspire to give back more than just a product. I know this because I talk with them constantly. An ecosystem offers us the opportunity to change the basis of competition to the betterment of all parties.”
Five years from now, as Google adds to its insurance partnerships, Karna says, Google’s partner customers can expect to understand ever-more complex risks, become more selective in the risks they bear, and be more strategic in the management of reserves and investment portfolios. This trajectory also comprises enhanced cross-selling and up-selling opportunities, emanating from a more holistic understanding of the customer’s risk profile, she says, while reducing partners’ data acquisition costs and overall expense ratios.
“Why reinvent the wheel when it is already moving forward?” Karna says. “We’re providing proven ways to accelerate the industry’s value proposition. By intelligently leveraging data, analytics, AI and machine learning, partners can exponentially increase their ability to predict and prevent losses, improving their underwriting, pricing, claims and reserving.”
For brokerages, carriers, reinsurers and other industry participants to obtain these perceived benefits, they may need to partner in an ecosystem with Google or, presumably, Microsoft, Amazon or another cloud computing provider. “I’m convinced,” Karna says, “this is how the insurance industry, an industry that has so capably served humanity and capitalism across the centuries, will continue to provide inestimable value.”
Russ Banham is a Pulitzer-nominated financial journalist and best-selling author.
TECH PLAYERS FOCUS ON NAT CAT MODELING
Google is not the only tech company focused on providing better data for underwriting property damage and business income risk. Multiple data and analytics efforts to this end are under way across the insurance sector.
“Historical data on hurricane frequency and severity are not entirely reliable by way of predicting future losses, requiring insurers and reinsurers to more closely scrutinize other data pertaining to a building’s structural integrity and the geographic costs of materials and labor to repair and replace damaged structures,” says Ketan Pandit, chief information officer at global insurer QBE North America.
QBE North America has armed its property underwriters with an array of cloud-based data and analytics tools provided by cloud provider Microsoft Azure that have been “democratized for non-technical uses,” Pandit says, “to rapidly gain insights into possible property losses based on the severity of different catastrophe events, like wildfires and hurricanes.”
San Diego-based insurtech startup FutureProof Technologies is another company focused on creating “the next generation of climate-smart property catastrophe models,” says Nick Lamparelli, the tech startup’s insurance strategy advisor.
“The property insurance market has been using catastrophe models for a couple decades now, which have been very valuable in preventing insurer insolvencies after hurricanes Andrew and Katrina. Those models are good at summarizing the types of events that can happen, in terms of where they can broadly occur and at what frequency. But they don’t generally offer precise projections of property damage losses at a particular geographic location—the vulnerability of a specific building.”
Having created analytic models using algorithms that account for a property’s age, structural quality, roofing condition, and other factors, FutureProof can enhance underwriting accuracy, Lamparelli says. “Instead of a one-to-many cat model, FutureProof provides a one-to-one model, aggregating actual carrier loss data and using modern statistical techniques to predict loss on an event-by-event, single-site basis,” he says.
FutureProof recently received $3 million in seed round funding led by venture capital firm Innovation Endeavors. “We saw an opportunity to use our tools to select, price and manage exposures,” says Alisa Valderrama, co-founder and CEO. “We believe that improved pricing of climate risk will create incentives to invest in climate resiliency measures.”
Another insurtech providing advanced technology and data solutions to enhance property underwriting and pricing is reThought, a managing general agency focused specifically on commercial flood insurance. “We launched the company in 2017 to serve mid-tier commercial property owners that for the most part don’t have flood insurance,” says Cory Isaacson, the MGA’s co-founder and CEO.
ReThought was formed to fill this gap in the mid-tier sector, which generally includes buildings valued from $2 million to $1 billion. Three specialty insurance markets provide capacity for the program, with support from Swiss Re and Munich Re. Isaacson says reThought generally offers higher limits of financial protection and more complete coverages to property owners.
“Our technology platform, which is built on top of traditional flood catastrophe models, is fed hundreds of thousands of attributes on each building, including hydrological, topological and geospatial attributes,” Isaacson says. “We then filter this data through our [pricing] engine.”
The upshot is a more refined and accurate assessment of a building’s potential replacement cost in the event of damage or destruction, says Isaacson, who provides the example of a large corporate campus complex with a $400 million schedule of buildings. “The owner told us that one cat model estimated the [flood] loss at $40,000 and another had it at $5.8 million,” he says. “Our formula had it at a little over $1 million. We’re closing the gap on flood insurance.”