Underwriting
Top Trends in Underwriting

• Jeff Roberts
• October 2019
• 

Key Points

• Evolving Role: The role of the underwriter is being rewritten by the implementation of big data, predictive analytics, artificial intelligence and machine learning.
• How You Handle It: New sources of data are emerging daily, but how that data is analyzed and applied makes the biggest difference.
• Exponential Growth: The value of AI underwriting will surpass $20 billion by 2024 from $1.3 billion this year, according to a recent estimate by Juniper Research.

The long-held promise of big data, predictive analytics, artificial intelligence and machine learning have moved past the theoretical stage and are being widely implemented in underwriting.

And they are changing the very role of underwriters across the insurance landscape.

Best's Review consulted a number of industry insiders to gauge the developing trends in underwriting. One overriding theme emerged: Carriers and reinsurers are applying lessons learned from other industries and elsewhere in the value chain to use these emerging technologies to assess and price risk.

The result is the realization of the game-changing tools' potential.

“The big thing here is insurers are doing what they said they were going to do or what they wanted to do about three years ago,” said Samantha Chow, senior life insurance and annuity analyst at research and advisory firm Aite Group. “It's finally coming to fruition. That's the big story.”

• Four specific underwriting trends were consistently cited:
• The evolution of the underwriting workforce.
• The rise of nontraditional data and predictive analytics.
• The wider implementation of artificial intelligence and machine learning.
• The embrace of telematics in personal lines.

Not one is a brand new concept. Each is the novel use or natural progression of existing tools, or in the case of the evolving workforce, a direct result of them. And each is helping to increase efficiency and automation.

“A lot of these themes evolved in other industries first, and it's been a matter of discovering the right use cases within life insurance,” said New York Life's Joel Albarella, senior vice president and head of New York Life Ventures. “AI and machine learning are key themes. And before #AI was a tagline, before #Bigdata was a tagline.

“It reminds me of the iPhone. The iPhone was a collection of existing technology that was put together in a new way that was highly valuable to the end consumer.”

All four trends are interconnected.

New data points and enhanced methods in analyzing traditional data sources are offering greater insights and predictability.

Telematics are feeding insurers with real-time, dynamic data in auto that transcends traditional static information.

AI and machine learning are taking pools of data and helping carriers automate, while freeing up underwriters to focus on more complex risks.

And the traditional role of the underwriter is evolving into a hybrid position, incorporating traditional skills, data science and tech analyst expertise.

The result is consistent underwriting, the opening of new risk pools that were once uninsurable (think private flood in P/C and diabetes and other chronic illnesses in life) and the development of new and more relevant products.

“Traditional underwriting will remain important, but it has started to change,” said MetLife's Chris Smith, executive vice president and global head of operations. “All carriers are looking at this. How can we bring more innovation to the underwriting field?”

For some, the top underwriting trend is the search for new talent and skill sets.

“The evolution of our workforce is going to be the single-most important success factor in helping us create products relevant to people and to interact with consumers in the way they want,” said Swiss Re's Chris Behling, chief underwriter for life and health in the U.S.

For others, this underwriting revolution starts with data.

“It's really all about data these days,” said Swiss Re's Mike Hudzik, head of casualty underwriting for the U.S. and Canada. “What you can do with those data sets is most important.”

The case is the same in life insurance.

“The advancement of technology around computational power and analytics is having a massive impact on anything that's processing-, decision- and pricing-related,” Albarella said.

Meanwhile, artificial intelligence and machine learning are finally paying off on their underwriting promise.

And telematics in individual auto is producing new and more precise forms of data, informing real-time, usage-based coverage.

“Telematics in the United States is about ready to explode,” said Greg Donaldson, Aite Group senior P/C insurance analyst. “There are enough companies that are figuring out how to get carriers to interact with their consumers in different ways.

“And you're starting to reach a tipping point with the amount of on-the-road data that these carriers and vendors need to be able to really start to make an impact on the ratings side.”

Here is a breakdown of each of those trends.

The Evolution of the Underwriter

Underwriting is being rethought across all insurance lines. And the role of the underwriter is evolving along with it.

The required skill sets are shifting as the amount of data grows and processes continue to automate.

“For insurance companies, cultivating the right talent will be as important as deploying the right data and technology,” said Swiss Re's Mohit Pande, head of property underwriting for the U.S. and Canada. “The talent will have to keep evolving.”

That talent will have to be more dynamic, with a skill set combining data science, behavioral economics and “old-school underwriting with the ability to innovate and think outside the box,” Pande added.

As the industry's tools change—accelerated underwriting platforms in life; telematics, predictive mapping and satellite imagery in P/C—underwriters have to grow with them.

Some will shift to monitoring the algorithms and software performing some of the simpler risk assessments. Others will serve as liaisons between vendors and software development teams.

“You're going to have a new breed of underwriters that are going to need data skills and analytics skills,” said Chris Stehno, a managing director at Deloitte. “It's tough finding people for those roles.”

On the P/C side, underwriters are being asked to be more customer-facing with brokers and the agents. They're also tasked with better articulating their company's risk appetite.

“They're looking to develop underwriters who are more sales-oriented and more effective at managing those relationships,” said Matthew Carrier, a principal at Deloitte.

There are two ways to develop these new skill sets: Recruit a new kind of talent or retrain the current workforce and capitalize on their built-in underwriting expertise.

Insurers seem to be doing both.

Many P/C companies undergoing transformations are pivoting their operations to reorient current employees.

“Any company that is implementing this—which is most, if not all—is starting to realize that they have to figure out ways to get their employees to do different things,” Donaldson said. “You're not reading in the news about mass layoffs at insurance carriers.

“The reason is they really have done a good job traditionally of taking the human capital they have and focusing it on the problems that they need them to focus on. This is no different.”

However, recruiting new talent—especially data scientists—is a necessity to effectively work with evolving technologies and expanded data sets, Swiss Re's Hudzik said.

Many life carriers that have moved to accelerated underwriting also are retraining underwriters, Chow said.

“We as an industry have to have one foot in the current world and one foot in the future world,” Swiss Re's Behling said. “There's a culture clash there between our underwriters of today and the data scientists we're bringing on, but I think both sides benefit from it. It's a trend in our industry.”

'The Magic' of Data and Analytics

The next generation of data is here.

In life insurance, some companies are collecting real-time data gleaned from wearables. And electronic health records are providing a wealth of information once inaccessible or difficult to act on.

“One area we've focused on is how nontraditional sources of data can drive value for the end consumer,” said New York Life's Albarella.

Life insurers increasingly are investigating how existing data points can become more predictive as well as expanding into areas not being utilized to price mortality risk. Lifestyle and behavioral data elements are emerging.

“There are new sources of data popping up every single day,” Behling said. “Social media data. Data from sources that didn't even exist five years ago.”

Some already consider internet usage and spending habits. Others in the industry have talked about the books consumers read and even their Netflix browsing histories—anything that will inform a scoring or pricing algorithm.

In the property space, mapping technology, satellite imagery and drones are offering a new spectrum of data to measure risk. Meanwhile, customer reviews and employee satisfaction ratings are being used to underwrite commercial coverage.

But the industry's growing capabilities in data analytics, producing granular insight and risk prediction, might be more valuable than new data.

Carriers that can extrapolate their data into the future will be able to insure more consumers, create a more engaging process and design more relevant products.

“The magic is knowing how these new data sources integrate with one another and how you make sense of all that information,” Behling said. “It's less about more data and more about the algorithms that we apply to that data.

“It's about making the underwriting process more specific to the individual and being smarter about how we use the data.”

Analytics are allowing insurers to tap into new risk pools such as private flood. And increased adoption of predictive modeling is helping them hone in on risk selection and pricing with multiple variables.

For instance, most property policies require physical damage before business interruption coverage kicks in. But businesses often are forced to close following major storms or earthquakes, even when they don't suffer property damage.

New data and analytics tools are leading to parametric products that cover business interruption due to power outages and areas of restricted access.

And the private flood market has become viable thanks to data and computing power. It could generate $41.6 billion in written premiums for insurers from U.S. owner-occupied homes, according to a Verisk analysis.

“Historically, flood was looked on as an uninsurable risk,” Pande said. “Insurers could only get good data on a location's risk through an on-site survey. But the mapping technology and the data behind this has become so much better that insurers can assess and price better than ever before.”

Imagery data from low-flying aircraft, drones and satellites allows insurers to assess properties without on-site visits.

“The addition of these new data points, which are created by the advanced use of AI and imagery, is really enhancing the customer experience,” Donaldson said.

“Now they're able to take a high-resolution satellite image of the property and say, '47% of your property is covered by foliage. That presents a risk score of 78 out of 100.' It's still relatively new, but that's a trend that's really starting to gain some momentum.”

New business modernization is bringing a range of efficiencies and actionable data that can be incorporated into databases. On the life side, it encompasses electronic health records and converting scanned images such as PDFs.

“That's probably the biggest change in life insurance—changing the way that an application is taken and the data is collected and stored,” Deloitte's Stehno said.

The approval process is responsible for the biggest bottlenecks in the life space thanks to waits for medical records and lab test results.

EHRs not only reduce wait time, but also could allow insurers to skip some paramedical exams. It is resulting in savings (the cost of those exams) and improved consumer experience.

“We're bringing in data that we didn't have in the past as well as that longevity of data,” Stehno said. “Instead of just having my height and weight from my paramed, now I've got 10 years of history of my heights and weights. There's a lot of value in that extra longevity of data.”

A Tipping Point for AI and Machine Learning

For years, insiders spoke about the potential of AI and how it would transform operations.

But carriers were using it to power chatbots.

They are finally implementing AI effectively in underwriting, even if it remains in its embryonic stages.

“AI is coming up to a tipping point,” Smith said. “It's becoming more advanced. The engines are there. They've been underwriting policies, and you have a lot of built-in processes and data around it.”

The value of AI underwriting will surpass $20 billion in premiums by 2024, according to a recent estimate by Juniper Research. The estimated value for this year? Just $1.3 billion.

“It all comes back to finally being able to realize the potential of AI and machine learning,” Donaldson said.

They not only produce more consistent results, but they can free up underwriters for more important tasks.

Carriers in commercial and personal lines are learning to automate simpler underwriting risks, “freeing up the human capital, which is more expensive, more knowledgeable and has a lot more nuances to the way it can think,” Donaldson said.

“It seems like they're finally starting to not just talk about AI. I've heard from carriers lately that they're really starting to work with vendors to develop AI-driven decision engines that can start taking away some of the low-level decisions from underwriters.”

And it's only the beginning.

“The use of AI is in its infancy on the commercial side,” Hudzik said.

Telematics: Driving Change

Allstate's Drivewise. Nationwide's SmartRide. Progressive's Snapshot.

They each use telematics, collecting and analyzing real-time driving data to not only inform pricing but also foster safer habits and collect more granular and insightful information.

But privacy concerns, reliability issues and the hassle of installing a device or downloading an app have led to disappointing adoption rates of usage-based insurance. About 5% of North American drivers use it.

However, some in the industry think it's reaching a tipping point where personal lines consumers will embrace it in considerable numbers.

“If you look at the top 10 and even the top 25 personal lines auto writers, they all have some type of offering,” Swiss Re's Hudzik said. “And the couple that do not—for example, Geico—have significant capabilities that they might not have launched yet.

“Telematics has started to become a needle-mover on the personal side.”

Usage-based insurance has long been used in commercial auto lines, especially among large, self-insured fleets.

“Commercial lines auto is leading the way. But personal lines is coming along, and it's going to get better,” Donaldson said. “Adoption will start to increase very quickly.”

Underwriting is evolving in auto and across the insurance landscape.

“The real story is technology is enabling a high level of efficiency through automation,” Chow said. “The impact of making underwriting more automated and streamlined will result in better profitability and higher employee and higher customer satisfaction.”