Three Ways AI and Machine Learning Impact the Future of Insurance

Joe Orr, EVP Insurance, Clearlink

Joe Orr, EVP Insurance, Clearlink

Over the past 20 years, we have seen the tremendous changes in technology upend the trajectory of every industry. At the leading edge of this push to transform is artificial intelligence (AI) and the potential power in allowing machines to carry out certain routine and predictive tasks based on data we collect on customers.

"Insurance is not known as the most agile industry, but there’s ample opportunity to disrupt this narrative and create truly intelligent customer experiences with each interaction"

One example that allows us to see the potential in AI is Starbucks. Seventeen years after issuing its first loyalty card, the company has had over a decade to collect data on their customers’ drink preferences and seasonal choices. They’ve since cross-referenced this data with other customer information, and they’re now getting ready to use AI to make personalized serving suggestions to customers at the drive-thru window. We also see companies like Netflix and Amazon doing something similar when their algorithms make personalized, and often very compelling, purchase and viewing recommendations to users.

The insurance industry can similarly benefit from the use of AI, but carriers must be willing to invest money in a system that can take the data they collect and organize it, label it and translate it into a predictive sales tool. This isn’t so much a choice as a requirement, as client expectations and industry competition will force many insurers to integrate AI just to catch up to the competition.

Using AI to Transform Unstructured Data

Insurance carriers are able to collect more information on policyholders today than at any other time in history. From wearable tech like Fitbits to embedded auto sensors to drones, we know more about insurance policyholders and applicants than ever before—or we can know more if we have the technical ability to parse the data we are collecting.

Initially, collected data is unstructured. It’s not until that data is organized, labeled, mined and cross-referenced that it becomes structured data. AI does this—taking the results of data fed into an algorithm and giving us meaningful facts that can inform the sales process, policy issue, rates, product development, claims and other decisions.

Let’s look at auto insurance for an example. Right now, auto insurance premiums are in part based on a driver’s credit score. Credit scores give insurance carriers insight into a driver’s general behavior, which they translate into a predictor of their driving behavior. In other words, insurers assume that if a driver is responsible about keeping their bills paid on time, then they’re likely to be more responsible behind the wheel.

With the tools currently available, however, insurers can gather information about a driver’s actual driving habits. Carriers can use an algorithm to measure a driver’s tendency to hard brake, use directionals, speed and so on. However, it takes a fair amount of intelligence and time to analyze the various driving habits, correlate them to other drivers, and predict risk, making an investment in AI essential.

Removing Friction in the Sales Pipeline

Insurance carrier call centers are recording every call, gathering data from every customer interaction. And with the help of AI, the future may see us feeding this data into a neural network of machines, which will then be able to evaluate and correlate the information based on the outcome you’ve told the network you’re seeking. In sales, we might tell the neural network that we want to analyze the differences in data from calls that ended with a sale and those that did not. Not only will the neural network then be able to identify customer traits that predict a high likelihood of a sale, but it will also identify words and approaches used in successful and unsuccessful sales calls.

Mining this information will allow us to harvest the tactics of the most effective sellers to inform the sales process of all agents. We can discover the key terms, words, and phrases used in a converting call versus an unconverted call and analyze the successful interactions within a sales environment. Taking this a step further, we should even gain the ability to match prospects with the right sales approach and sales professional by analyzing their engagement level during different situations, giving us a far more effective method of distributing leads.  

As an example, an AI algorithm might recognize a customer’s frequent use of the term, “budget” as an indicator that they are price sensitive, thus helping an agent focus their approach so that inexpensive premiums are front and center, increasing efficiency and likelihood of conversion. And because AI is self-thinking, it will even find patterns and identify correlations that we can’t, giving us advantages in sales we’d never get on our own.

Through these advantages, AI may help carriers move sales predictability away from the Pareto principle and give insurers an increased yet highly predictable level of ongoing sales success. These sophisticated engines may also allow us to train agents with these processes from day one, so they are in the field, successfully, at the start. Since machine learning can pick up on things that people may not, it also opens up opportunities for more frequent cross-selling.

Meeting Evolving Customer Expectations

We already use AI for customer communication through chatbots, but the future holds so much more. Like Starbucks, Netflix and Amazon, insurance carriers can use AI to analyze customer needs and make useful suggestions for coverage updates and additions. The level of service, study, and nuance presented by AI may also clear the way for the rise of virtual agents. If the future of insurance is in a hybrid virtual/brick-and-mortar model, then AI gives us the tools we need to have a unified system and process for both virtual and field agents—which is good for success rates, persistency and compliance.

There are some who might argue that AI takes away a human element and disrupts the customer experience—yet AI needs the support of human intelligence to work properly. You cannot use AI to make a decision about a customer unless you can tell the regulator why you're making the decision.

Insurance is not known as the most agile industry, but there’s ample opportunity to disrupt this narrative and create truly intelligent customer experiences with each interaction. And while one of the most significant assets an insurer can have is customer data, they also must label, parse, evaluate and correlate it. It takes time for machines to learn, which means carriers must have the tools in place today if they want to realize the full potential of their customer data in real-time. Carriers that wait might not be able to catch up, leaving them in the past as a casualty of evolving customer expectations and competitor capabilities.

Weekly Brief

Top 10 Artificial Intelligence Solution Companies - 2020

Read Also



Steven R. Bauman, Head of Global Programs and Captive Practice in North America, AXA XL
Data-Driven Decision Making Starts with Partnerships between Business Leaders and the Data Team

Data-Driven Decision Making Starts with Partnerships between Business Leaders and the Data Team

Ajaz Akhtar, VP - Chief Data & Analytics Officer, Decision Sciences Center of Excellence (DSCOE), Utica National Insurance Group
Top 2022 predictions for the insurance industry

Top 2022 predictions for the insurance industry

Ray Miller,Director. Industrial Coverage,Charles Schwab
How transforming your communications ecosystem can grow revenue and profitability

How transforming your communications ecosystem can grow revenue and profitability

Dave Zamorski, General Manager Broadridge, Customer Communications, Digital Solutions
How to Organise Data in an Ecosystem

How to Organise Data in an Ecosystem

Patrick Wirth, Vice President Mobility Ecosystem, Baloise