• Insurance premiums are calculated by several factors like age, the type of insurance coverage, the amount of coverage, personal information, insurance history and other risk factors. These risk factors are influenced by the vulnerablity of the insuree. Reports show that upto 55% of the insurance premiums paid by insurees are not correct to their profile.

• The estimation of insurance premium is centralized and the premium value can be manipulated by middle-men and insurance brokers - leading to a financial loss.

• Insurance premiums are not personalized enough according to the insuree's details. There are only a limited number of factors to decide the insurance premium which often leads to miscalculation.

Features of Finsure

• Machine Learning model to estimate the insurance coverage and premium amount based on various factors fetched from Open Govenment APIs. The account balances are also aggregated over time for different states using PhonePe's Pulse API. The ML model uses fully connected dense layers of aggregated data from various data sources and predicts the best coverage and premium account based on the user's profile. (For example, a vehicle which is registered in a flood prone city will receive less coverage and requires more premium amount as they are vulnerable to be affected by floods).

• Blockchain and distributed storage of Insurance receipts using Inter Planetry File System. All the receipts stored in IPFS are immutable and can not be tampered. This can greatly improve the transparency in insurance sector.

Tech Stack