AI for Attracting Millennial Insurance Customers – What’s Possible

Dylan Azulay

Dylan is Senior Analyst of Financial Services at Emerj, conducting research on AI use-cases across banking, insurance, and wealth management.

AI for Attracting Millennial Insurance Customers - What's Possible

Machine learning has far-ranging applications in the finance space broadly from document digitization to document search, chatbots to fraud detection. The insurance space in particular, however, stands to benefit from AI and machine learning applications in a few unique ways. They could help insurance firms with a challenge that’s at the forefront of the insurance world: attracting and meeting the needs of millennial customers.

Insurance Nexus hosted a video interview between three insurance experts: Viktorija Kaminski, Senior Risk Manager, Europe at Uber; Elena Rasa, Chief Underwriting Officer at Zurich; and Samantha Chow, Senior Analyst at Aite Group. They provided some insights into how insurance carriers are using technology to develop new products that could attract and retain millennial customers.

Graham Proud, Head of European Product Development at Insurance Nexus, said something that tees up this discussion well: “Those [insurance firms] that prove themselves to be ambitious and embrace change and are willing to put the customer at the forefront will be the most successful.”

According to Chow:

There’s a huge gap in demographics. You’ve got the baby boomers…but you also have this very large generation, the millennial generation…and they’ve been raised in a very different environment. Their needs are very different because their surroundings are now very different. Their behaviors are also very different and continue to change as a result of all the options they have. When I talk about options, what I’m really talking about is the options of how they want to communicate with people and how they want to buy.

In this article, we run through some of the ways that insurance carriers can use AI and/or data science to attract millennial customers, peppering our insights with quotes from the insurance experts that took part in Insurance Nexus’ video interview.

Readers should get out of this article a firm sense of both the challenges that established insurance carriers face when implementing new technologies and the rewards they could stand to gain if they do.

We start our analysis with an application that’s dominating the AI conversation in insurance: conversational interfaces, or chatbots.

Conversational Interfaces

Although Chow concedes that not every millennial wants to interact with a chatbot or purchase insurance over the internet, it’s hard to deny that millennials are used to online purchases of many kinds. Ultimately though, she says, “They want guidance.”

Insurance firms might invest in conversational interfaces to cater to millennial customers that might be looking to research products on a carrier’s website or perhaps even purchase so-called “on-demand insurance.” Insurance carriers are already providing ways for customers to purchase insurance products on their smartphones, and carriers like Lemonade, which sells product liability insurance, exist entirely online.

Although chatbots seem to be the most prominent AI use case in insurance—insurance giants like Progressive, Allstate, and Geico all have chatbots of their own on their websites—building one that can actually serve a millennial customer’s needs is no small endeavor.

Natural language processing applications such as chatbots require volumes of historical customer support tickets that come in via email or direct messaging applications like Facebook Messenger. This poses a challenge to established firms in a variety of sectors, but older insurance carriers may have particular difficulty with implementing chatbots due to these data requirements.

They simply don’t interact with their customers via email as often as a retail brand might, for example. Even if they’ve been speaking with customers over email for the last decade, they may not be organizing these interactions in a way that’s useful for training a natural language processing algorithm.

That said, established insurance carriers do have options. There are rare NLP vendors that offer unsupervised learning solutions, which mean they don’t require historical data on the part of the insurance carrier. These solutions right now, however, are mostly good only for routing support tickets into appropriate categories.

Developing Insurance Products for the Gig Economy

Perhaps more pressing than updating their communication channels, insurance carriers may want to consider updating their products to fit the lifestyle of today’s workers. “Insurance providers must first focus on people…creating tailor-made products for gig-economy workers,” says Kaminski.

Less and less people are finding themselves at large companies that provide them the benefits that older generations may have been afforded upon their employment. Instead, they’re more and more self-employed. Their working hours and paychecks are not fixed and guaranteed.

If an Uber driver’s car breaks down or for some reason they can’t make it out on the road for a night, he won’t be bringing home any money. A freelance writer isn’t guaranteed clients, and a light month could mean the difference between paying her family’s bills or not.

This has ramifications for insurance carriers that are trying to attract their business. Kaminski highlights the challenge this poses for established insurance carriers when she says, “[insurance carriers] are all familiar with a traditional white-collar contractor, but we’re not really in-tune with what real gig-economy independent workers’ needs are. They’re completely different. They’re super cash flow sensitive.”

Faster Claims Payouts

One way that insurance carriers can attract these kinds of contractors is by offering faster claims payouts. Although claims processing automation is extremely nascent, there are a few AI vendors that are offering claims filing automation at the very least.

Oftentimes this means that customers can file their claim on a smartphone app that may or may not involve a chatbot. Sometimes, that chatbot can route customers with questions it can’t answer over to a claims specialist that can speak with them in real-time.

Once the customer files a claim, however, it’s up to the insurance carrier to adjust it and determine the customer’s payout as quickly as possible in order to retain the customer’s business. Some AI vendors offer claims adjustment automation, most often for the auto insurance space. These applications in some cases involve machine vision.

Customers can take a picture of their car’s damage, for example, and upload the image to the insurance carrier’s app. A machine vision algorithm trained on historical payouts for similarly-damaged cars would then suggest a payout to a claims adjuster that can either approve or reject it.

This could significantly speed up the time it takes for the customer to receive their payout check or direct deposit. In addition, it could be a strong selling point for the insurance carrier in a few years, but again, machine vision for claims automation like this is still in its infancy.

Personalized Insurance Products

Millennials are used to messaging, advertising, and media that’s highly targeted to their preferences. Having grown up in a world that’s so heavily connected, they’re also both aware of their options and used to getting things fast. This may translate to the kinds of insurance products they want.

We already discussed how insurance carriers can offer more digital communication channels and faster claims payouts; in this section, we discuss how the design of insurance products themselves can meet the needs of millennial customers.

Chow discussed how millennial customers tend to look at health, wealth, and life insurance as one and the same. This affects the way they search for it and buy it. According to Chow, insurance carriers are moving to more “holistic” products as a result, products that insure a customer in ways that would traditionally be exclusive to health, wealth, and life insurance individually.

This is important because it gets to what millennials are really searching for: security. Chow made the point that millennials that are actually seeking employment are looking for employers that offer benefits. This includes health insurance and signals of long-term security.

They seem to want to get their most pressing needs met in one place, and Chow said, “[insurance carriers] have to meet them on their terms or we’re not going to win their business.”

Although AI isn’t going to make a decision for an insurance carrier on whether or not they should offer these bundled insurance products, machine learning could help allow insurance carriers to offer them. This comes down to data.

Bundling Products and Underwriting Automation

The insurance carriers that start collecting and organizing their customer data sooner rather than later will likely be the ones that win in the market as AI becomes more ubiquitous, as our CEO outlines in his AI Zeitgeist series.

The underwriting process can be extremely thorough and time-consuming. AI-based underwriting automation could be one key for allowing insurance companies to offer bundled insurance products. For example, an underwriter at an auto insurance company might know a few dozen critical factors to pay attention to that might give her a sense of whether or not to underwrite a customer’s policy.

An underwriter at a life insurance firm might look at several dozen completely different factors to make his decision.

Some companies are working on training AI software to make decisions like a salesperson, and this requires top salespeople and data scientists to collaborate and do what’s called feature engineering. In essence, they break down the way in which these salespeople tend to make decisions on which leads to prioritize, for example. This same concept can apply to underwriting.

To tie this all together, insurance underwriters in a variety of sectors (auto, life, health) can collaborate with data scientists to create the framework of machine learning models that can “make decisions” the way those underwriters might.

If the insurance carrier has collected the kinds of data points that underwriters from different sectors would generally use to decide if they’re going to underwrite a customer’s policy or not, then the carrier can use this data to train the machine learning model.

In theory, this could result in a machine learning software that speeds up the underwriting process for bundled products because it could suggest the premium that a customer should pay for a bundled product given the data the insurance carrier has on them.

Alternatively, an insurance carrier could build a machine learning model that uses data from past customers with bundled policies to determine if a new customer would have positive or negative ROI if they were to purchase a bundled product. The model’s output could inform whether or not the insurance carrier offers that customer a bundled product.

That said, vendor solutions for underwriting automation are few and far between and aren’t always as credible as we like to see. At this time, a machine learning model for underwriting may be best kept in-house, which would require data science talent.

This is an arduous process, and insurance carriers that aren’t prepared to take on the risk of building a machine learning software that actually doesn’t achieve ROI should not do this.

The Internet of Things

Rasa said that Zurich is already using data to offer customers life and health insurance, although it’s unclear if Zurich offers these as a bundle. Specifically, Rasa discussed how Zurich is using “devices” to do this:

With the use of a lot more devices…We’re moving into prevention…instead of giving the coverage when the disease appears, we try to track the development of the disease over time and give the customers the right products to make sure that he can feel more secure, followed, and that someone is taking care of them.

The internet of things (IoT) is one method of data collection that insurance giants like Progressive are already using to collect driving behavior data on customers. Progressive’s Snapshot program allows customers to open an app on their phone or install a device in their car that collects telematics data that indicates how sharp the driver’s turns are and how fast they’re driving.

As a result, Progressive can offer customers personalized policies based on their driving habits. Carriers can offer safer drivers lower premiums and distracted drivers higher premiums.

Kaminski echoed the potential value of IoT in auto insurance when she says, “[millennial customers] want to use their vehicle as they choose with insurance coverage that adapts.” This goes for home, life, and health insurance as well.

In the health insurance sector, wearables can be a way for customers to provide insurance carriers with data on their lifestyle and the state of their health. This could allow health insurance to offer lower premiums to customers with healthier lifestyles, which may attract millennial customers, who tend to have relatively healthy lifestyles.

We discuss how IoT devices could achieve ROI for insurance carriers in our report on the possibilities of IoT data in insurance.

Concluding Thoughts

Proud reiterated, “Insurers are tackling issues with legacy systems, processes, communication, culture, enterprise architecture…it really is a huge challenge but with huge rewards.” Although we only briefly discussed the challenges that insurers face in the era of AI in this article,

Proud rightly points out that a company’s culture can be make-or-break when it comes to attracting and retaining the data science talent necessary for building the machine learning applications we discussed above.

That said, AI will become accessible to more mid-sized and large insurance carriers over the course of the next few years. Carriers may want to think about their AI strategy in order to win market share with millennial clients as they age and so-called “Generation Z” clients as they enter adulthood and consider their insurance options as customers that are even more digitally-native than the millennials.

We discuss how insurance carriers can prepare for machine learning at their company in our report on enterprise adoption of AI in the insurance sector.


This article was sponsored by Insurance Nexus, and was written, edited and published in alignment with our transparent Emerj sponsored content guidelines. Learn more about reaching our AI-focused executive audience on our Emerj advertising page.

Header Image Credit: Tenstickers