CerebrumX announced a data-sharing partnership with Ford that will enable the company to provide real-time UBI (Usage-Based Insurance) data to insurance companies. This partnership applies to 2020 model year and later editions of Ford and Lincoln vehicles. The announcement adds to an impressive list of automotive OEM partners that CerebrumX has agreements with, including Toyota, Nissan and Stellantis.
UBI has grown in popularity significantly over the past three years (the recent COVID epidemic has drastically reduced car use) and allows insurance pricing based on actual kilometers driven. CerebrumX improves on this through its SaaS (Software as a Service) product that pulls ~250 different data feeds from vehicle CAN (Controller Area Network), infotainment and telematics systems. CerebrumX ADLP (Augmented Deep Learning Platform) processes and integrates these data streams and supplements them with AI-based assessments and insights for real-time customer access over 4G/5G connectivity. In addition to usage, insurance companies can also use this data to assess driver behavior, accident reconstruction, emergency dispatch and roadside assistance.
Compared to most competing products that use anonymized data, CerebrumX’s approach offers its customers vehicle-specific data and generates driver and vehicle ratings without the need for additional hardware or apps. This score helps insurers better assess risk and create more accurate, personalized policies for their customers, such as: B. Pay As You Drive (PAYD) and Pay How You Drive (PHYD) to support safe driving and optimize claims. Figure 1 shows the types of data inputs and outputs of the Cerebrum ADLP platform.
The value of the ADLP lies in creating insights for different customers who are dealing with specific applications in the ecosystem. Vehicle condition and driver rating are relevant for insurance providers. Fleet operators value insight into fuel and mileage patterns as well as proactive maintenance information to avoid costly downtime. As V2X and advanced ADAS become more prevalent, the types of data available are increasing, enabling deeper insights and more advanced applications (e.g. assessing driver engagement through haptic sensor data).
The ADLP platform is a cloud-based “synthetic sensor” that can integrate other data sources (e.g. data from in-cab and road cameras). However, capturing and transmitting continuous full camera data is expensive, introduces latency, and is likely to saturate connectivity bandwidth. Part of CerebrumX’s intellectual property (developed jointly with partners) is to detect salient events (e.g. driver distraction or an accident) and capture relevant time-stamped sensor data slices. Algorithms in the ADLP process this data and make it available to the customer in a time-synchronized manner for accident reconstruction and claims management.
As Figure 1 shows, CerebrumX’s SaaS products have other applications besides UBI, which has insurance carriers as end customers, including:
- Fleet management: uses connected data across fleet vehicles to monitor location and condition, driver safety, collisions and proactive maintenance. Fleet owners use this data to maximize vehicle availability and scheduling.
- After-market warranty and repair: Service providers can use real-time CAN data and alerts to encourage proactive maintenance and reduce repair costs
- Electric vehicle (EV) battery monitoring and charging: ADLP data products can be used by charging service providers to optimize site selection and power/battery needs and car users for trip planning.
- Roadside safety: Simplified access to alerts and location in the event of an incident to enable rapid dispatch of first responders with accurate information on vehicle occupants and status
- Traffic flow management: Collecting telemetry data from groups of vehicles enables traffic flow prediction and signal control to smooth the flow and avoid congestion and accidents
CerebrumX was founded in 2018 and its global headquarters established in Princeton, New Jersey in March 2020. It employs a total of 40 people (30 of them in India) and raised a $5 million Series A round in 2020, two of which are significant investors – LG Technology Ventures and Cerence. Based on agreements with OEM partners, >15 million vehicles are currently connected to its network. OEMs charge between $2-$6/vehicle/month for data access, with the market value of ADLP SaaS products ranging from $6-$10/vehicle/month. As volumes increase, OEM data access costs are expected to decrease, increasing the profitability of the SaaS business model. The company is revenue generating and currently has a backlog of >$20M in customer orders (7 signed customers in the insurance, fleet and aftermarket service industries) for delivery by 2024. It plans to expand into other regions and industries and expects to be profitable by 2024.
According to Sumit Chauhan, Chief Operating Officer at CerebrumX: “Modern cars generate significant amounts of data and have the ability to connect to the outside world. The availability of high-bandwidth connectivity, embedded compute, cloud computing, and intelligent edge and cloud computing are required to provide solutions for industries such as insurance, fleet management, aftermarket services, smart cities, etc. The overall market for relevant SaaS products and services is expected to grow to around 100 billion US dollars by 2025. CerebrumX looks forward to being a part of this journey and pioneering work down the road”.
As L3 autonomy (autonomous driving with a human driver ready to take control within 10 seconds) kicks in and L4 (full autonomy under certain conditions) beckons, SaaS products that use vehicle and other data like Using V2X, increasingly important for applications that rely on autonomy. For example, the safety of the autonomous propulsion stack, as well as vehicle health and refueling requirements can be qualified. Passenger data can be analyzed to provide important information for targeted advertising and the safety of passengers and vehicles. Communication bandwidth will likely be a challenge, as will processing actionable and stellar data from a much richer sensor stack replacing human perception.