With proliferation of insurtech, insurers have access to more and better data than ever before
AI and machine learning will allow insurers to interpret data to fundamentally transform the industry
Innovations in distribution and customer engagement will raise questions for brokers and agents
Insurtech industry faces regulatory hurdles at every stage of the value chain
The emerging insurance tech (“insurtech”) industry is growing at breakneck speed. While in 2011 insurtech startups received $140 million in funding from 28 total deals, investment increased more than tenfold in 2016, with investors pouring nearly $1.7 billion into 173 deals. Given the massive opportunity to innovate within the insurance industry, those trends are projected to continue.
On July 19 and 20, 2017, Akin Gump Strauss Hauer & Feld LLP participated in a conference in New York City—“Insurtech: Innovation, Corporate Strategy & Customer Engagement”—to discuss the future of insurtech. The conference featured representatives from startups, accelerators, insurers, reinsurers, venture capital firms, consultants, regulators and in-house and outside counsel, all of whom provided insight into what is next for the industry.
We break down our top four key takeaways that any companies or entrepreneur in the insurance industry should consider in business planning.
1. Insurers Have Access to More and Better Data Than Ever Before.
It was a common refrain: insurers now know more, about more things, with more certainty, than ever before. While traditionally insurers’ ability to price products, underwrite policies, detect fraud and manage claims has been restricted by relatively limited data (e.g., application forms or historical claims data), the amount and quality of available data is set to explode. Insurers will soon be gathering, interpreting and incorporating into their business models data created by the following:
• wearable devices that monitor health and biometrics
• automobile sensors (personal, rental and commercial)
• home sensors
• sensors for business operations and supply chains (e.g., crates and containers in shipping, cranes for construction or warehouses for storage)
• drones and satellite imagery
• complaint and review websites (e.g., Yelp, Glassdoor)
• U.S. government websites (e.g., data.gov, NASA, NIH 1000 Genomes)
• international organizations (e.g., World Health Organization)
• Amazon Web Services
• Google searches, Maps, location history
• Facebook, Twitter, LinkedIn, Instagram, etc.
2. AI and Machine Learning Will Allow Insurers to Interpret Data to Fundamentally Transform the Industry.
Vast amounts of data, when processed and analyzed using sophisticated Artificial Intelligence (AI), with processing power that continues to advance, will fundamentally change the industry. Numerous panelists and presenters provided concrete examples of how these forces will drive change across every insurance line, from auto to property and casualty to health.
In the auto industry, car sensors will lead to more usage-based policies so that consumers pay for only the miles they drive. Additional data will also give rise to incentive-based plans and pricing. For instance, insurers may incentivize consumers (through lower premiums) to obtain better “smooth-driving” scores. Initial pilots have revealed that such smooth-driving scores—based on sensor-gathered data about factors like hard-braking, acceleration and speeding—leads to better driving and fewer claims filed, a win-win for consumers, insurers and society. The opportunities open up further when gamification is introduced. Friends, or even strangers, may compete with others on the road for “smoothness” supremacy, with winners receiving prizes and discounts.
In the property and casualty space, insurers can incentivize (and even remind) their insurers to turn off stoves, close and lock doors, fix leaks, prepare for upcoming storms and even detect fraud or improperly priced risk. As one panelist explained, while an insured may not disclose that he uses his home to screen pay-per-view sports events for hundreds of paying customers every weekend, the fact that his door opens and closes 500 times each Saturday night will be a red flag to an insurer.
In the health insurance industry, insurance products may soon be more similar to virtual personal health assistants than to insurance of the past. These new products will not only minimize risk, but they will also reduce claims and costs: software will automatically switch an insured from CVS to Walgreens when medication prices rise, notes from “telemedicine” calls will be fed directly into online digital user profiles, and AI will analyze a consumer’s entire health profile and history—consolidated on one platform—against physician guidelines and treatment protocol to inform consumers not only of the risks they face, but which doctor to visit. Data might be used in less intuitive ways as well. As one panelist noted, by harnessing data from Google searches—for example, an increase in the search “what are flu symptoms?”—insurers may notify employers in advance of coming risks so that they may take actions (such as flu shots at work) to decrease claims and absenteeism.
3. Innovations in Distribution and Customer Engagement Will Raise Questions for Brokers and Agents.
In the hyper connected modern economy, consumers expect goods and services to be convenient and easily accessible. For consumers buying insurance, however, the experience is often confusing, frustrating and riddled with hard-copy forms and fax-machine-level technology. Submitting and settling claims is no better. In fact, the insurance industry has one of the lowest levels of customer satisfaction of any industry.
Given this, many startups have focused on improving the existing distribution and claims-handling models, offering AI-enhanced policy advice, digital purchasing platforms and lightning-fast issuance of policies. As one speaker stated, his goal was to take everything an insurance agent currently does—from hard-copy forms to telephone calls—and put it online. But innovation in this space raised questions from industry participants. In particular, will changes like digitization, automation and advances in chatbots and robo-advisors constitute a threat to brokers, or simply an opportunity to better service clients?
4. Regulatory Hurdles Abound.
Regulatory hurdles abound at every stage of the value chain. While data and AI allow insurers to more accurately price risk, regulation may prohibit specific factors from being considered, or may even prescribe the precise factors that must be considered. This problem is compounded by the fact that regulation of insurance is largely left to the states, resulting in 50 different regulatory regimes with which insurers must comply. For instance, while some states expressly permit credit scores to be considered when rate-setting for property and casualty policies, numerous other states apply strong limitations. While some states expressly permit genetic data to be used in the life and disability space, others expressly prohibit it. The same level of nonuniformity can be seen across many characteristics and all lines of insurance. While this complex system is already difficult to navigate for insurance veterans, it may prove even more daunting for startups and less-seasoned companies.
Another potential regulatory roadblock is what one panelist referred to as the “accuracy vs. determinability conundrum.” That is, as AI becomes more sophisticated and accurate at pricing risk, humans will be less capable of explaining or understanding it. This poses significant challenges for state actuaries and rate reviewers, who are tasked with understanding not only vast numbers of risk-classification factors, but also how each factor is connected to and influences the others. Rate filings of tomorrow will require significant cooperation between insurers and regulators to better understand these hyper sophisticated, AI-driven risk-pricing tools.
Distribution and claims-handling are also rife with traps for the unwary. Does providing free HR software to consumers who purchase insurance products violate anti-rebate laws? How many and which employees working in a brokerage startup must be licensed? When consumers purchase insurance products without interacting with a human, who is the licensed broker on the transaction? When AI is used to assess damage to a car, how will companies demonstrate that the approach is “reasonable” under state fair settlement practices laws? One startup speaker explained that, while the oft-quoted mantra “move fast and break things” is appealing, operating in a highly regulated industry like insurance complicates matters. Indeed, companies that fail to properly appreciate and engage with the regulatory framework may risk making costly (and sometimes disastrous) missteps.