Imagine a scenario. A client’s car meets with an accident and a telematics IoT device connected to his car’s OBD-II port “senses” the actual damage and files a claim with an auto insurer. The entire process – detecting damage, estimating cost and filing claims – takes place automatically. Within seconds, the insurer’s chatbot pops up on the client’s smartphone to take action on the claim, which includes accepting or rejecting the claim and processing repairs, replacements and reimbursements. This is the future, and it is already here. Barring a few WIPs, the entire claims automation use case can be realized today – thanks to artificial intelligence (AI).

Based on our experience of working with several Insurance companies, we understand that they handle more than thousands of claims per month. The three key challenges of claims processing in any insurance firm are:

• Reducing claims turnaround time (TAT): TAT is considered as the most important factor that decides the loyalty of a client to an insurer. Unfortunately, traditional ways of processing Insurance Claims that involve manual claims management and settlement can lead to insufferably long TAT where clients have to pay upfront from their own pocket. The next scenario involves either waiting for a long time for the claim to be settled or being denied. While 9% of hospital expenses are denied initially, only 63% of them are recoverable.
• Reigning in operational costs: Traditionally the claims processing center is the most labor-intensive department in insurance. Since these are human brains processing truckloads of insurance claims everyday, it cripples operational efficiency and increases chances of errors and inaccuracies, thereby frustrating the clients. Therefore, it is regarded as the largest cost center. According to a CAQH report, manually managing claims processes costs providers $4 more than managing them electronically. The same report highlights that electronic claims management solutions can save healthcare providers $7.9 billion yearly.
• Detecting instances of fraudulent claims: Fraud cases in the US insurance (non-health insurance) industry has a yearly worth of more than $40 billion, amounting to an estimated 10% of overall claims expenditure. And needless to say, manual claims management runs a high risk of failing to detect fraud. In most cases, the loss is converted into higher premiums, costing the average US family somewhere between $400 and $700 yearly. This leads to client attrition.

Note that in all the three pitfalls, the ultimate sufferer turns out to be the insured. Now, not many insurers can take that risk as:

• Insurers can’t survive competition without satisfying and delighting their customers
• Customers are no longer putting up with shoddy insurance services. They want faster support, customization, lower premiums and so on. 
• Failure to deliver a superior claims experience to customers will push them away to a bunch of new insurtech firms armed with the two key needs of the hour – proper technology and customer-focus. 

According to a recent report, IT, analytics and customer experience are the top 3 business functions that might contribute to an improved claims process. This indicates that insurers are gradually realizing the untapped potential of data and AI’s capabilities in harnessing it to offer the best possible value to clients and insurers.

AI holds the key for insurers to completely transform the claims experience for the new-age, tech-savvy client. While it delivers value across the entire insurance lifecycle, AI’s contribution in claims processing is perhaps the most impactful for insurers. With its team of cutting-edge technologies like machine learning (ML), natural language processing (NLP), internet of things (IoT), optical character recognition (OCR), robotic process automation (RPA), computer vision, etc., AI opens up a whole new world of possibilities for the claims experience in insurance.

Here’s a look at how AI transforms key areas of the claims journey for an insurer.

Key tech: NLP, OCR, RPA, intelligent automation
FNOL usually comes in the form of text, PDFs, images or speech through emails, audio calls and multimedia messages. NLP can easily scan the data, search for related policy, generate structured datasets based on claim reports, and analyze it to assess payouts, flag liabilities, identify recovery opportunities and detect fraud. While structured data can easily be handled using RPA, intelligent automation does the trick when data comes semi-structured or unstructured. Documents can be scanned and indexed automatically using OCR. On the other hand, chatbots can enable real-time communication with customers to keep them updated about their claims status.

Key benefits: Operational efficiency, freeing up time for human employees to focus on other crucial areas.

Example: Elafris, Inc.’s chatbot enables customers to submit a claim simply by answering a few questions and clicking on damaged areas in car drawings. Once a claim has been submitted, the chatbot communicates that a claims adjuster will follow-up, and also shares information of local repair shops. FNOL is a matter of minutes without the claims adjuster stepping in.

Key tech: Computer vision, ML
Once the client clicks a picture of the damage to his car or home or other property and sends it to his insurer, AI kicks in. Computer vision helps insurers by scanning those images, running them against historical and trained datasets of similar images, recognizing patterns, and then finally estimating cost by assessing the extent of damage. Apart from mobile images, footage from drones, sent to do aerial surveys of unreachable areas, can also be analyzed using AI. Moreover, loss estimates can be made using augmented reality (AR) as well.

Based on the analysis, chatbots can communicate the triaging of claims by offering payouts, repairing information, replacements, reimbursements or even rejecting the claim. Any possible fraud cases can be flagged to a human counterpart for deeper analysis and follow-up by an AI-powered chatbot too. AI can also analyze pictures of external environments like weather and traffic conditions to rule out fraud claims, and identify subrogation opportunities in a claim.

Key benefits: Smoother processing, reduction in errors, omissions and inaccuracies.

Example: Tractable’s software helps insurers automate claims using machine vision. Agents can upload images of damaged properties with a payout quote. Tractable’s system compares the uploaded images against a dataset of images of various severities of damages and associated payouts. If the estimated payout by the agent is more, Tractable informs him so that the agent can adjust the actual payout accordingly and avoid claim leakage.

Key tech: NLP, ML
By analyzing historical data, previous claims, customer social media profiles, etc., NLP can build predictive risk models to assign scores or risk ratings to various claims. This helps in detecting fraud cases in the insurance industry. Intelligent automation can claim verification, reserve updates and offer real-time communication during the adjudication of claims process.

Key benefits: Reduction of fraud costs for insurers, which can translate to lower premiums for customers.

Example: French startup Shift Technology uses a SaaS approach to use big data and machine learning to detect patterns of fraud insurance claims, preventing claims leakage and increasing efficiencies of insurance claims handling teams.

Key tech: RPA, intelligent automation
RPA along with intelligent automation helps out in this area with automated cash transfers, invoice generation, validation and check, data exchange, auto follow-up and so on. Chatbots play a crucial role in real-time communication and following up with clients.

Key benefits: Faster TAT

Example: Lemonade’s claims bot AI Jim reviewed a claim, cross-referenced it against the policy, ran 18 anti-fraud algorithms, approved the claim, sent wire instructions to the bank for a cash transfer, and informed the client about it – all within a world record of 3 seconds!

An automated and enhanced claims experience results in true customer delight. We partnered with one of our clients involved in the commercial insurance market to build a claim management solution, with necessary analytics and litigation engine in place. The scalable engine of the solution was able to scout through billions of records in less than 60 seconds and raise critical alerts. The solution led to generation of early claim settlement and customer alerts that saved time and money, ultimately leading to high customer satisfaction.

AI and its irrefutable role in the claims lifecycle in insurance has a direct impact on customer experience. With faster TAT, lower errors and inaccuracies, improved service, real-time communication and support, the satisfaction and retention of clients takes a big leap. Moreover, business benefits like operational efficiencies, lower operational costs and reduced fraud cases get transferred to the customers in the form of a smooth user experience, lower premiums and more scope and time to provide personalized offerings.

Just like claims, to know more about how AI impacts other functions of insurance and unlocks superior experience for customers in the insurance value chain, download our report here.