In recent years, the insurance industry has increasingly been making use of technology to improve its services and capabilities. In this article, Eric Veron and Nobel Basser look at some of the latest trends and technologies being used by insurtechs to improve the speed and accuracy of insurance claims.
What is fuelling rapid insurance claims?
Insurtechs use technology to improve all aspects of insurance, from requesting quotes all the way to the settlement of claims. With the rise of Insurtech, the assessment of claims is more accurate, and the time taken to settle claims has nosedived – on 23rd December 2016, Lemonade set a world record for the fastest processed insurance claim in history. It received a claim for a $979 coat, checked the claim against the policy, ran 18 anti-fraud algorithms and made the payment - all in just three seconds and zero paperwork.1 It was so quick that it probably took you longer to read the previous sentence…
The question is, how has this been enabled? In this article, we discuss the technologies and trends that are fuelling rapid insurance claims, as well as some examples.
The three main trends improving the speed and assessment accuracy of claims are:
- AI (Artificial Intelligence)
Nowadays, there are sensors for a plethora of measurements. In your phone alone, the number of sensors reaches well into the double figures such as the magnetometer (compass), GPS, gyroscope, accelerometer, thermometer and pedometer, as well as many others such as the proximity, ambient light, fingerprint, facial recognition, microphone, and camera sensors. When there are sensors connected directly to the internet, they become IoT (Internet of Things) devices, which also include gadgets, appliances and actuators under the umbrella. Alexa and Google Home are prime examples. IoT devices are able to collect vast amounts of data and then transmit it. When used in insurance, they speed up claims, help reduce fraud, create policies tied to accurate measurements, and help enable agile underwriting.
One insurer that makes full use of sensors is FloodFlash, who hold the record for the fastest property flood claim, with just 9 hours and 44 minutes between the property being flooded and the policy holder receiving full settlement.2 All of this was done remotely, with no need for a claims inspector – instead it used an internet-connected sensor installed at the property, which measures flood depths within the property. As soon as flood waters reached the client’s trigger depth, this information was transmitted to FloodFlash’s HQ, who proceeded to validate the claim and pay the settlement.
In its most simple definition, Artificial Intelligence (AI) is about teaching computers and machines to perform tasks that would normally require human involvement. Artificial intelligence is seeing increased use, as deploying AI is cheaper than hiring humans to do the same task, while also doing it more quickly.
One of the simplest uses of AI is in chatbots. ZhongAn’s chatbot is responsible for 97% of its communications, with the ability to serve more than 300,000 users simultaneously. Across the insurance value chain, ZhongAn automates 99% of policy underwriting and 95% of claims settlement.3 This has helped significantly minimise costs, as fewer customer service agents are required, while at the same time rapidly increasing service speeds as customers spend less time waiting on the phone to be served.
Another more revolutionary example of the use of AI is from Tractable. Tractable uses AI to assess damage to vehicles from just smartphone photographs, which in turn helps estimate repair costs and speed up the claims process. Customers need only submit photos via their phones, and can settle their claims within minutes or, in some cases, during the initial phone call – preventing the need for in-person assessments. The data used to train Tractable’s AI tool comes from historical data from millions of accidents. P&C insurer MS&AD Insurance Group believe that their adoption of Tractable’s AI solution could accelerate the claims process by two weeks and save its adjusters up to 360,000 hours per year, while Tractable has already helped with over $1bn in auto claims for insurers like Covea and Ageas UK.4
Nevertheless, the growing usage of AI in insurance should not supersede the key principles of insurance, mutualization and hazard, by over personalizing exclusion analysis and underwriting, from the claims data analysis.
Blockchain technology is currently underused in the insurance industry but has potential for a variety of use cases. The strength of blockchain and the distributed ledger system lies in the core principle of every party receiving the same information – keeping all parties involved in the process and creating a transparent database while minimising fraud.
Marine insurance is one of the sectors which could benefit most from the implementation of blockchain and smart contracts. Currently, “there are a lot of paper contracts written between insurers, brokers, shipping companies and captains of ships.”5 When the risk parameters change, such as a change of captain and change in weather, all parties associated with the policy must update their paper records – introducing the risk of errors and fraud, which increase costs and take time to rectify. This is before we even consider the various jurisdictions and countries ships travel between – Hong Kong waters will have a different set of rules and regulations compared to British waters. Compliance of the different regulations in different countries can also be very difficult to handle when paper contracts are involved. The use of smart contracts helps minimise this complexity, reducing overall paperwork and ensuring all relevant parties receive the same data.
EY and Guardtime worked with Microsoft and major shipping company Maersk to develop a blockchain-enabled platform to enable better communication and transfer of data between the stakeholders in a marine insurance contract. Through a combination of smart contracts and telematics such as GPS sensors, Maersk can enjoy a faster claims process and more accurate premiums, developing trust between all parties. If, for example, a ship transporting goods from Singapore and flying an Argentinian flag ends up sinking off the coast of South Africa, telematic sensors will detect the sinking, while the ship’s GPS will have been keeping track of its position throughout this journey. Throughout the ship’s trip, sensors will transmit data to all relevant parties and the blockchain, ensuring that the relevant sections in the smart contract are constantly updated. Sensors detecting the sinking of the ship will send information which leads to clauses being triggered in the smart contract, adjusting for the appropriate jurisdiction and leading to the claim being processed and paid (and emergency services being sent to the ship’s last location to find the crew). This entire process is completely digital, with no human involvement required - a stark contrast to the vast amounts of paperwork flying around – and allows the shipping company to find out as soon as possible that a ship has sunk to duly notify its customers.
Overall, the use of blockchain ensures quicker and easier transmission of data, while also minimising fraud. This leads to much faster claims processing as all relevant data are been received and verified much sooner.
With regards to claims, Insurtech is not just speeding them up, but also removing fraud from the process and improving accuracy. This has been enabled by three main trends – telematics, AI, and blockchain. However, Insurtech is making waves across the entire insurance value chain, from helping create usage-based insurance products to enabling more accurate underwriting.
With the introduction of more data, technology, and telematics, an industry which has traditionally been reactive can now start to be proactive. An instance of this is through schemes that offer discounts to car drivers if they install telematics devices, which in turn encourage them to drive more safely with the knowledge their driving is being tracked and analysed.
Nevertheless, the growing usage of AI in insurance should not supersede the key principles of insurance, mutualisation and hazard, by over-personalizing exclusion analysis and underwriting, from the claims data analysis.