Platform For Mutual Insurance Of Domestic Animals Using Smart Contracts Was Created In Russia
ReGa Risk Sharing has developed a platform for mutual insurance of domestic animals using smart contracts that also will provide identification and scoring with the help of photos.
The Risk Management Robot of ReGaPetBot, which is DAO on the Ethereum platform, will help a client to join the club of mutual support for pet owners. According to the company, the membership fee is around 350 rubles per month; a member will receive around 80% reimbursement of the cost of veterinary services in the event of sudden illness or accident with a pet.
Besides, ReGaPetBot using photos on social media will help an owner to find his animal if it is lost, and organize an online consultation with a veterinarian.
To get a club card it is enough to open a dialogue with the robot in one of the messengers: Facebook Messenger, Telegram, Slack, or on the robot site, you should upload your own photo, pet photos and transfer money to club party’s smart contract.
To receive compensation you should make an appointment to a vet through ReGaPetBot, re-take a picture of an animal in a clinic and send photos, a check to the robot. The robot will automatically compare the data obtained from the owner with the data obtained from the veterinary clinic and hold a reimbursement from the fund of the Club to club member's account.
Thus the Risk Management robot implements the distribution financial product that works without human intervention and centralized storage resources, which significantly reduces its cost, maintaining reliability and enhancing ease of use.
The interface of ReGaPetBot’s robot was developed using Microsoft Bot Framework and allows you to maintain multiple channels of communication with customers using a single application server.
The Risk Management Robot uses a hierarchical structure group of smart contracts on the Ethereum platform (Microsoft Azure Ethereum Consortium Blockchain). Each participant of the club placed in a group corresponding to his level of scoring. The level of risk and limits are monitored for each group online. If the risk level exceeds a predetermined level, the group could be disbanded and initial fees return to members of the group.
An animal is identified by the algorithm based on neural network, that allows comparing different pictures of the same animal with accuracy around 95%.
The solution architecture is below:
To construct a scoring model that allows classifying risks with the help of photos were used: Microsoft Cognitive Service and Azure Machine Learning. With the help of Microsoft Face API (MS Cognitive) were treated about 15 000 photos of potential clients and the model based on the algorithm of Two-Class Boosted Decision Tree that during the test gave around 73% accuracy in determining the quality of a customer.