Claim Automation &
Automated Damage Assessment
Indonesia
Client
Automotive InsurTech Company
Short story
In Indonesia auto-insurance has long been a well established, highly competitive market, seeking to cope with this situation, insurance companies have begun to diversify their business both within and without their traditional core, opening up to new audiences and adapting digitisation to keep pace with competitors. Two of the major core business diversification based on digitisation are, underwrite better risks & at better prices, and reduce claims and claims management cost.
CLIENT
Client is one of the Insurtech companies that has provided claims processing and approval solutions to insurance companies. One of the important modules of the product was to assist underwriters with claims approval, saving time and detecting frauds.
The current process includes manual inspection of damaged car parts, estimating the cost and approving claims. Images of damaged car parts along with the documents are submitted for processing. This is a time consuming process and is prone to human error. Client’s product plugs in the existing claim processing system of the insurer and their mobile application. Pictures/video taken by authorised vendors are analysed for detecting damage with it’s severity automatically. It is an input to estimate claim’s amount saving time with improved accountability.
CHALLENGE
The claim process flow encompasses multiple stakeholders - from end customers to agents/vendors to the underwriter. Manual processes and lack of checks causes frauds, inefficient processes and lack of good quality data.
THE TOP 3 CHALLENGES:
-
Presence of unstructured data across documents: Data is captured by vendors/agents that includes claim & scanned kyc documents, pictures & videos of damaged vehicles.
-
Data incompleteness: Documents and pictures submitted are inconclusive to approve/reject a claim
-
Repetitive & distorted data: Sometimes data lacks human readability and is not scanned properly.
SOLUTION
Assistive claim processing system taking into account submitted documents, damaged vehicle images and information to reduce claim processing time, avoid frauds. It integrates into the existing process flow via APIs and an extended dashboard to give a 360 degree view of the claim to an underwriter. To ensure good quality data capture, ML models are running on agents/vendor mobile applications as an additional check if the right document or damaged part image is captured. Underwriters can quickly see damaged vehicle parts and it’s severity with an estimation of claim amount.
RESULT
-
Auto insurance claim approval time is reduced from '4-5 hours' to 5 min
-
Checks during document upload saves time by avoiding multiple iterations and improves customer experience