In this interview, Hari Balakrishnan, Founder & Chief Technology Officer of Cambridge Mobile Telematics, shares his insights on how the smartphone is one of the key components in distracted driving and how the same tool can actually be used to improve driving quality. Cambridge Mobile Telematics is a global company working with 35 organizations from over 20 different countries to accurately measure driving quality, deploy behavioral incentives to improve driving, and use AI to automate the claims process.
Keep reading to learn how Hari and CMT are investing in the future of better driving by expanding their global platform, developing new video analytics software, and improving driving safety in a world where self-driving cars are a soon-to-be reality.
What problem did you set out to solve when founding CMT?
In 2010, the concept of using a smartphone to collect fine-grained data to draw accurate inferences about vehicle dynamics or give direct feedback to drivers was unknown. Any attempts at that time to use smartphones to better understand user behavior were seen as inaccurate and generally unreliable, especially in the field of telematics.
When we started Cambridge Mobile Telematics (CMT), however, we were betting that the smartphone would be a critical, central component to the field in the future; more so, that although the smartphone would be the very device responsible for distracting drivers on the road, it could also help make them less distracted and become better drivers.
We also realized that many people do not recognize the dangers of driving while distracted. We were seeking approaches for what we viewed as preventable incidents. Our goal was to accurately measure driving quality and to create better drivers, making the roads safer for all.
Over the last nine years, CMT’s DriveWell platform has helped make roads safer by making drivers better in a world where crashes are rising because of factors like distracted driving. CMT’s rapid growth is fueled by a company culture that is deeply customer-committed, values collaboration, and values creativity via investment in research to improve current solutions and develop new ones.
CMT’s DriveWell platform provides insights on drivers and vehicle dynamics to insurers. What type of insights do you provide?
Many of the largest insurers in the world have adopted CMT’s DriveWell platform – a complete telematics and behavioral analytics solution for the connected car world that (1) accurately measures driving quality using mobile sensor data, (2) deploys a range of behavioral incentives to improve driving by reducing risk factors such as phone distraction and risky speeding, and (3) uses artificial intelligence on telematics data to automatically automate several aspects of claims management.
Using machine learning and signal processing, DriveWell accurately infers key metrics about mileage, speed, acceleration, driving style, distraction, and collisions. DriveWell understands a driver’s behavior over time, providing positive reinforcement for safe driving behavior: incentives such as gift cards and reduced insurance rates based on good driving influence behaviors. Meanwhile, our AI methods operating on sensor data can step in to save lives in the event of a crash.
Following a crash, it takes a lot of time and money for insurers and drivers to process a claim. CMT’s crash reconstruction technology applies AI techniques to telematics and contextual data, providing insights and decision-support capabilities to reduce the effort and cost. As a first step, crash reconstruction provides a comprehensive picture of the event, using processed telematics data, calculated crash indicators, and contextual information. Insurers can receive this information visually in the DriveWell portal or via an API. With this service, insurers can see machine-generated crash descriptions and details like severity rating, number of impacts, duration of impact, probability of vehicle hit location, weather and more.
As a result, insurers and agents are able to begin the claims process earlier and reduce manual efforts to document and analyze crash details.
There is a huge behavioral science movement in the tech startup community. Can you tell us a bit more about the impact and why it is important for the insurance industry to have such a deep understanding of behavioral science?
We have pioneered many innovations since our 2010 inception and spinoff from MIT’sComputer Science and Artificial Intelligence Lab. We deployed the first service to efficiently gather and process sensory data from phones for auto insurance in 2012; used phone sensors to measure phone distraction in 2013; and induced better driving with gamification in 2014. Together, these innovations created the category of “behavior-based insurance,” also known as “mobile usage-based insurance,” that enables the insurance industry to better price policies and save their customers money.
Using machine learning algorithms to assimilate information such as excessive speeding, harsh acceleration or distraction, DriveWell understands the drivers’ behaviors over time. It can then provide feedback, gamified ranking and positive reinforcement to encourage long-term driving style change. DriveWell’s impact on driver safety is undeniable and well documented:
The relative risk of crashing increases by a factor of 23 if texting while driving. Throughout the U.S., distracted driving occurs on over one-third of trips. After seven days of using the DriveWell app, we observe a 15% decrease in distraction events, followed by a 35% reduction in distraction after 30 days. These improvements can be sustained.
Data seems to be the driving force of your business. How will data continue to help insurers in the future?
We recently introduced new services including real-time impact alerts for roadside assistance and first notice of loss (FNOL) as well as crash reconstruction and claims assessment using mobile telematics data. These services benefit insurers, agents and drivers by starting the claims process earlier and reducing manual efforts to submit, document and analyze crashes.
CMT’s crash reconstruction technology allows insurers to save time and expenses by automating many of the key steps in the claims lifecycle. By using the technology provided through CMT, they can receive accurate, comprehensive impact and contextual data, improving and building driver relationships with additional safety services.
We also launched extensive vehicle fleet features for commercial insurers and ride-hailing companies, and best-in-class risk scoring models, which enable insurers to accurately price risk using telematics data.
Our actuarial risk scoring services emphasize crash risk factors like phone distraction and at-risk speeding to augment traditional telematics factors such as hard braking. The model provides a decile lift of 22x and is now approved by regulators in 28 states.