Artificial intelligence, Machine Learning, and Deep Learning are more than futuristic concepts. These technologies are impacting the insurance industry in a significant way right now and this impact is likely to increase in the near future.
The idea of Artificial Intelligence (AI), Machine Learning (ML), and Deep Learning (DL) may fascinate consumers who enjoy talking to their digital while admiring a Nest® thermostat. But for the insurance industry, these terms are business-changers that affect products and services offered and interactions with consumers and other industry partners.
The definitions of these terms may be a bit confusing to the uninitiated (see sidebar). But at CCC, these technologies come into play every day as we help insurance companies, collision repairers, OEMs, and consumers make sense and use of the vast amounts of data becoming available through connected cars and technologies.
Already, telematics data is moving us toward usage-based insurance models, and has changed the way consumers interact with their vehicles. And, Deep Learning models will revolutionize our decision-making ability and the way claims are processed and paid.
Every Picture Tells a Story
Less than 10 years ago, we humans had to walk into a bank and hand a paper check to a teller to have it processed and deposited. It is now commonplace for consumers and businesses to use a mobile app to upload a photo of that check to said bank for deposit, all from the comfort of their home, office, etc.
So, when will AI and related technologies be in place for a consumer or claims adjuster be able to initiate a claim by simply snapping a photo of a damaged automobile? CCC has been in production with machine learning since 2011, helping customers determine the most appropriate appraisal channel for a vehicle and improve their identification of total losses. What’s next – and coming very soon – is the next wave of AI, DL, and ML.
For example, soon insurance adjusters that use CCC could take that picture and a neural network, a form of AI, would help them detect and classify the damaged areas. DL algorithms would be able to predict the most effective appraisal channel for that insurer based on a single vehicle photo. This type of model could drive decisions about where a car will be towed: to a repair facility or to a salvage yard. Today, we are over 85% accurate in predicting that an insurer will declare a car totaled through analysis of a photo of the damaged vehicle, and we are working on improving this accuracy.
Finally, a “market basket” analysis will kick in. Consider: If you’re shopping online and you have bread and peanut butter in your cart, you may see a suggestion for “Jam” or “Jelly” because the technology in place can see patterns of these things being purchased together. Similarly, CCC is working with AI, or pattern recognition technology, to support the collision repair process. In the future, if a collision repairer orders a front bumper with an energy absorber and a license plate frame through the CCC ONE® platform, the system will know that there’s a great chance a hood front seal or a front bumper cover clip may also be needed to complete that repair and suggest that to the repairer. We think of this as a potential job aid as it helps to suggest parts that are also likely to be part of a repair, helping an estimator to write the estimate quickly and efficiently.
Real-Time Behavior Modification
But what about longer-term, future applications for these advanced technologies. CCC is always in the lab testing new applications that may power solutions one day.
Take our Distracted Driver project, this proof of concept project used ML and DL models with an internal-facing windshield camera to observe simulated driving behavior in real time. Movements can be detected and analyzed to better understand driver actions and reactions while behind the wheel. Potential business use cases for this technology include: educating teenage drivers on their driving behaviors and risk factors; monitoring fleet and bus drivers; and combining behavioral data and telematics data to identify risk trends in a graphical report (ex: “you drove two seconds with your eyes closed going 55mph. That equals 160 feet.”)
AI + ML + DL = Better Claims & Safer Drivers
You don’t need to be a data scientist to take advantage of these business cases, because CCC is committed to helping insurers think through and apply the latest technologies. For decades, we’ve provided automotive claims, repair, and more recently, telematics solutions that help drive efficiencies and enable our customers to make smarter decisions. We assist our customers by putting their data through our models to help identify time- and cost-saving decisions every day. Our technology is state of the art, and we keep investing into making it better every single day. The possibilities for our industry are exciting and are what drives CCC’s innovations forward.
Artificial Intelligence, Machine Learning, and Deep Learning are not interchangeable terms, but they are related. Here’s how:
AI is the oldest of these terms and generally refers to machines that exhibit human intelligence. AI dates back to even ancient time in stories and calculating devices but recent examples include autonomous driving, Google Translation, and IBM Watson. According to TechEmergence, AI soon will shorten commutes via self-driving cars that result in up to 90% fewer accidents, more efficient ride sharing to reduce the number of cars on the road by up to 75%, and smart traffic lights that reduce wait times by 40% and overall travel time by 26% in a pilot study.
ML is the subfield of computer science that gives computers the ability to learn without being explicitly programmed. An example of this is an email services’ ability to remove a vast amount of spam from your incoming email while rarely including a personal email in that spam detection. Also, the ridesharing app Uber, uses machine learning for ETAs for rides, estimated meal delivery times on UberEATS®, computing optimal pickup locations, as well as for fraud detection.
As seen in a recent Forbes article, DL focuses even more narrowly on a subset of ML tools and techniques, and applies them to solving just about any problem which requires “thought” – human or artificial. Deep Learning can be applied to any form of data – machine signals, audio, video, speech, written words – to produce conclusions that seem as if they have been arrived at by humans. Today, sensors and onboard DL analytics can teach cars to recognize obstacles and react to them appropriately.
Nest® is a licensed trademark of by Nest Labs, Inc. UberEATS® is a licensed trademark of UBER Technologies.