Since artificial intelligence (AI) was invented, people have been worried they would become too smart, become self-aware, and take over the world like Cyberdyne’s Skynet system in the Terminator franchise. Well, it’s been more than 70 years since Alan Turing’s famous paper on thinking machines, and we’re still telling them what to do – even with the rise of AI in insurance claims.
However, computers have replaced humans for certain tasks, even those we were sure required a human touch. But even though Artificial Intelligence can write music, sometimes you still need some Mozart. This principle applies to almost every field—AI can execute tasks, but the chasm between executing defined tasks and managing complex project is still vast.
So, what does AI in insurance mean for claims professionals? Will AI take over for claims inspections and offer helpful customer service, carefully determine damage, and make accurate recommendations to insurance companies? Or will people continue to be necessary throughout the claims process?
We believe the solution lies in the middle and is not in a binary choice between all AI or all human. But before we get ahead of ourselves, let’s define AI and what kind of AI in insurance claims we are dissecting.
What is Artificial Intelligence (AI)?
Put simply, AI is a field that uses computers, machines, and datasets to mimic the problem-solving and decision-making capabilities of humans. From the beginning, computers have been great at gathering and storing information and making simple calculations. But with AI, computers can now take the gathered information, analyze it, and come to rational conclusions.
“Artificial Intelligence” is often used as a catchall term, but here are some distinctions within the AI landscape:
Machine Learning – The science of making machines “learn.” It also has the ability to plan, act, understand, and reason.
Deep Learning – “A type of machine learning that uses multi-layered neural networks to learn.” (This is what drives Loveland Innovation’s IMGING.)
Cognitive Computing – A subset of AI that focuses on simulating human thought processes. It uses “natural language processing and machine learning to enable people and machines to interact more naturally.”
Data Science – “The interdisciplinary field that incorporates statistics, mathematics, computer science, and business analysis to collect, organize, and analyze large amounts of data to generate actionable insights.”
Now let’s talk about how AI in insurance is being used, with a specific focus on claims and underwriting.
How Insurance Carriers Use AI for Insurance Claims
Back in the day, we used hand-written and hand-drawn inspection reports. Now we can capture property data with phones, apps, and AI (specifically deep learning) and store findings on the cloud. This new cross-collaboration approach allows insurance adjusters to standardize inspections and stay organized.
Essentially, this means we’re not replacing people with technology to perform inspections and file claims. Instead, adjusters can use AI in insurance claims to help them work faster and safer. AI also removes manual tasks through automation, increases consistency across claim data, and decreases costs.
For example, this is how IMGING from Loveland Innovations uses AI to make the claims process:
- Automated – IMGING automates much of the claims scoping process with AI-powered damage detection. Adjusters save minutes per house, and hundreds of hours a year just through the time savings of AI-aided visual inspections.
- Consistent – No matter how many times or who inspects a property, you’ll get the same data every time using IMGING’s AI which is available using IMGING on a mobile device or a drone.
- Cost Effective – AI helps streamline the inspection process, so the amount of time an adjusters spends at a property is reduced, which decreases LAE.
- Safer – Pairing AI with drone-based image capture allows adjusters to adjust claims while safely on the ground.
How Insurance Carriers Use AI for Underwriting
When it comes to underwriting, insurance carriers vary widely in their approach. For example, carriers, like Hippo and Lemonade, have taken people out of the equation altogether by using data science. Here’s what Daniel Schreiber, co-founder and CEO of Lemonade, had to say about their philosophy, “We believe in replacing brokers and paperwork with bots and machine learning, and we now have the backing to unleash this formula across new products and geographies.”
Companies like Lemonade and Hippo are finding success with a strictly AI strategy. These companies have also struggled in a difficult insurance environment. However, this tactic is riskier for premium, non-traditional, or historic properties. That’s why high-end insurers use IMGING for risk and loss control functions.
When it comes to high-value underwriting, you can’t replace people. For IMGING and companies like Betterview, the best approach combines traditional underwriting performed by people and AI data. A real person provides tried-and-true information gathering methods and AI in insurance underwritng provides new data that wasn’t available in the past or was extremely costly to obtain. The two sources of information—traditional assessment and AI data—provides a comprehensive view of the assumed risk for the lender.
How Insurance Carriers Use AI for Customer Experience
There are a lot of different aspects of the customer experience. But for our purposes, we’ll talk about how AI in insurance is used in sales, customer service, and claim settlement time.
Insurance Sales
Today, it’s all about customization. People only want to pay for what they need, and insurance websites and apps are offering more customization features that provide tailored and less expensive insurance options with limited friction. For certain types of insurance, AI has actually increased efficiency and customer satisfaction during the purchase process.
Time-to-Settle
Taking out the middleman (i.e. a person), will speeds up the settlement process for small, obvious claims. As mentioned above, Lemonade boasts their bots’ ability to file and pay claims without any human intervention. However, when severe damage from recent storms is not obvious and an inspection is in order, human interaction and understanding can’t be replaced.
How Insurance Carriers Use AI to Prevent Fraud
Did you know insurance carriers routinely report $80 billion in fraudulent claims? That’s kind of a big deal, which is why fraud prevention is a priority for every insurance carrier. And the best way to prevent fraud is with good underwriting and thoroughly investigating claims.
Underwriting
Performing aerial and handheld inspections of insured properties before extreme weather and accidents can protect carriers against fraud. Again, when it comes to high-value properties, it’s important to combine assessments performed by people and AI data.
Claims
While it’s convenient for people to file claims through a website or app, it’s important that insurance carriers ensure the claims are valid. Detecting insurance claim fraud or inaccuracies requires a lot of time, effort, and money. And, in some cases, AI can help make this process more efficient.
In a matter of seconds, AI can analyze and cross-reference both internal and external databases to detect anomalies. But there are also things that only a real person can detect when assessing a claim, which is why it’s important to employ insurance adjusters in addition to AI data for all but the most obvious of claims. True advances will occur not when simple AI is used for anomaly detection, but when deep learning models are consistently leveraged to save time while relying on adjusters’ judgement and expertise to settle claims and provide excellent customer experiences.
There’s no doubt that AI in insurance and other new technologies can improve operations and customer experience. And while some insurance carriers have figured out how to replace people with AI, there are still many instances that require both AI data and human insights and instinct to reliably asses insurance claims, underwrite customers, provide good customer service, and prevent fraud.
