AI risk prediction tool for accurate real-time pricing for insurance

点击翻译按钮获取中文摘要

The bike market (including e-bikes) has experienced unprecedented growth in the past year, propelled by COVID-19 as a preferable alternative to public transport and government incentives to help tackle obesity and meet net-zero carbon targets (Cycle to Work, Fix Your Bike vouchers). As a result of increased demand (and as a consequence, value), bike theft and accidents are predicted to soar post-lockdown. Insurance is critical in getting people back on their bikes after accident or theft, however research shows a staggeringly low number of cyclists have appropriate (if any) insurance cover.A fundamental issue with current insurance is that the business model is flawed. Insurers price premiums based on customer claims history combined with a risk-factor of future expected losses (when/where theft/damage will occur), and do not account for customers' values, behaviours or changing circumstances. This results in: (1) Inaccurate and unfair pricing. Inherent inaccuracies and a lack of customer segmentation/understanding means many customers overpay, others are priced out, creating a lack of equality and inclusion; (2) Lack of transparency. Customers share data but do not know how it is used; (3) Distrust. Insurers are incentivised to reject claims in order to maximise profits, acting against customers' best interests. Premiums are paid upfront before anything happens, renewal prices increase with no justification and contracts are in-flexible despite changing circumstances/behaviours, highlighted during COVID where premiums remained the same despite changes in behaviour.Laka aims to overcome these inherent flaws, by offering a disruptive, fair, community-based insurance model. Customers join a group of people with similar risk profiles, and at the end of each month the group splits the actual cost of claims submitted, up to a personal cap. Laka get paid when claims are settled, providing complete transparency and aligned interests. Laka customers have, to-date, saved on average 25% compared to other insurance.Although already gaining significant interest and support, Laka's current approach to 'risk grouping' is based on limited data/understanding of the customer and therefore most suited to a particular type of customer risk profile ('bike enthusiasts'). To meet demand and become 'fit for purpose' and inclusive for all types of riders e.g. mass-commuter, e-mobility -- the two fastest growing segments, Laka aim to develop a novel AI customer 'risk' prediction tool to more accurately calculate/predict customers' real-time risk based on subjective datapoints, grouping them appropriately and generating individual pricing that is fair, fluid and reflects individual behaviours and circumstances.