Quantum AI said on Nov. 12 it has launched 'llodi', designed to support automation of long-term compensation tasks from claims intake to review, payment and follow-up management.
The company said claims reviewers have manually checked and entered large volumes of documents, including medical certificates, receipts and detailed medical expense statements. It said policy terms and review guidelines are scattered across multiple places, and criteria vary depending on the accident date and applicable timing, requiring reviewers to check each item themselves.
Llodi links five task areas: intake and data entry; accident and contract pre-verification; review and benefit calculation; approval and notice automation; and linking with AI fraud detection systems (FDS). It classifies various claim documents, including medical certificates, receipts and detailed medical expense statements, extracts key information, and organises materials needed for review. Based on this, it supports automation of tasks from review and benefit calculation to approval, customer notices and follow-up management.
Llodi also records the basis for each AI decision. The company said this allows reviewers to check the policy terms applied to a claim together with the grounds for the decision.
Quantum AI also stressed it has developed its own document recognition technology that uses image processing based on an overlapping segmentation method that divides documents into multiple overlapping areas for reading. It said the technology recognises documents with high accuracy even when text is broken or overlapped, or when documents are rotated.
Quantum AI CEO Seong-jip Choi (최성집) said benefit reviews are the area with the most difficult judgments in insurance operations because medical, contract, policy terms and calculations are tightly intertwined. He said the company has produced results by first applying to long-term compensation task automation a method in which AI analyses various data and document images, presents the grounds and results first, and humans conduct the final review. He said it now plans to expand this across the insurance value chain.