3 ways tech is speeding up insurance claims

Having to file an insurance claim can be as stressful as the accident or situation that led to having to file the claim in the first place. Often the process is slow, cumbersome, and if you have to pay the bill upfront and wait for reimbursement, it can be a huge stress on your financial situation. While automation and digitization have expedited almost every aspect of business, the insurance agency has lagged, despite being an industry that now averages $5 trillion a year in gross premiums. This is quickly changing, however, as insurance providers recognize the need for innovation. Claims processing is one of the areas to see quick implementation and even quicker results. 

The McKinsey’s Insurance 360° benchmarking survey found that increased automation can drive up to 80% in cost savings for individual processes and according to their futurist report, 

“AI technology will reduce the overhead associated with claims by 70 to 90 percent, compared to 2018 levels. Automated customer service apps will handle most customer interactions through voice or text. Human agents will primarily focus on complex claims situations, with human interaction bolstered by analytics and data-driven insights. With most claims largely automated in 2030, claims processing times will be dramatically reduced from days to minutes.”

Zurich Insurance is an example of a company seeing those kinds of results. They recently test piloted an automated claim filing system and reported reducing the processing time from between 10-15 days to less than half an hour. The program also reported higher accuracy and customer satisfaction.

While numerous technologies can go into creating different processing tools, here are three of the main ways AI, ML, and NLP are speeding up claims processing and making it better. 

OCR- optical character recognition

Your paperwork can pass through dozens of hands from when you first file your claim until you see a payout. Often this is not metaphorical paperwork either, but the real deal, pen, and paper type. While some aspects of insurance claims and payouts are moving towards digitization, most documents in legacy insurance companies are still handwritten. For those times where it’s just not possible or practical to have everything digitized from the beginning- OCR, or optical character recognition can help speed things up and make the claims process more manageable. Instead of an agent manually typing out what someone wrote by hand, OCR can read the text using computer vision and automatically fill out the claim (read about what computer vision is and how it works here). 

The same technology can also be used for tasks such as signing up new customers since it can take info from customer IDs and be added to the customer profile in seconds.

The French insurance company AXA recently ran a deep learning pilot program to extract data from handwritten documents with a 96% accuracy rate (which is probably at least on par, if not better than any employee's ability to read their co-workers’ handwriting). These technologies often struggle with slow start because of the different required input (the way an American writes a 1 is different than the way a German would, and how many different ways can people write the letter ‘a’) but as the programs are used and adapted in different areas, the improvement process should increase rapidly. 

In addition to being faster and more convenient, having all of a company’s paperwork digitized increases transparency. Claims, how they are processed and how the company justifies the payout amont, are all done behind a curtain, so most customers don’t fully trust their insurance company to be fair. Allowing dashboard access for clients to look through their accounts and figures, might be a way to increase trust through transparency. 

Automated processing  

As McKinsey predicted and Zurich Insurance has seen, automation is the heavy-hitter when it comes to benefits from automation. Setting up automation processes is relatively easy and inexpensive, and shows the most value. In addition to being cumbersome and error-prone, the traditional claims management process can be expensive. According to a report by McKinsey & Company, the claims management process can eat up 50-80% of premiums’ revenue.

While there are many ways to automate claims, the one that currently makes the most sense is through mobile apps. Numerous auto insurance agencies, including Allstate, have adopted digital filing tools so that as soon as an accident occurs, the policyholder can start filing the claim and sending in photos from their phone. With their Quick Foto Claim, you can send photos, receive a quote and accept payment all through the app. In the not too distant future, drones might become a heavily relied upon tool as well. Insurance agencies could send out a drone to take photos of the accident and immediately file them to be processed. While auto insurance claims are probably the most common use cases of this technology, the same systems could be used to send in pictures of damaged luggage for travel lost claims or personal insurance for damaged items. 

Health insurance is another example of where technology can speed up processes. Fukoku Mutual Life, a Turkish insurance company, created a tool based on IMBs Watson that accesses all related patient medical files, mines them, then auto-calculates the payouts. Post-adoption, the company reported that staff productivity improved by 30%, and payouts were more accurate.

Fraud

Fraudulent claims are a significant problem in the insurance industry. According to a report from the FBI, American insurance companies lose $40B a year to fraud (that’s not even including health insurance fraud). When your industry sees that much fraud on a daily basis, it is to be expected that adjusters are going to take their time and scrutinize every claim, which of course, takes time. 

Data from the IoT (internet of things), including sensors from cars and smart homes, can make it easier to prove or disprove a claim in an instant. You might be familiar with these technologies from your own home or car; they often monitor things like temperature, pressure, and object position. Using this data, insurance companies can see right away if there is sufficient data to back up a claim or if it needs more speculation. By automating the more clear-cut cases, human agents are freed up to spend their time on more complex claims and everything moves faster. As more examples and data are collected, ML algorithms will be able to better classify and process claims based on severity and complexity.

The Turkish insurance company Anadolu Sigorta implemented a predictive fraud detection system and reported a 210% ROI in one year and $5.7M in saved fraud payouts. While this can seem one-sided, ‘insurance companies have yet another way to nickel and dime consumers,’ insurers should take this opportunity for savings to offer more competitive policies to their consumers. Part of the reason fraudulent claims are so high to begin with is a lack of trust between consumers and insurance providers. Customers feel taken advantage of and some want to take retributive action. If insurance companies can use these savings to offer better-priced options and offer transparent policies, that trust can be restored. Insurtech start-ups like Lemonade have shown that customers are eager to have insurance that works for them, instead of the other way around. In 2020, Lemonade made $94M in revenue, despite only being seven years old. Numerous insurtech start-ups are taking an ever-increasing size of the insurance pie because they understand the insurance market is changing. 

The insurance industry is facing a seismic, tech-driven shift. As more advanced forms of AL like deep learning are improved, the insurance industry, along with probably every other industry, will be almost unrecognizable by today's standards. If insurance companies are able to keep up, the growth will be well worth the investment. 

To read this article in its original publishing, click here