Insurance companies are shifting their focus in the artificial intelligence (AI) competition from merely adopting technology to how quickly and flexibly they can apply it to actual insurance services. Efforts are underway to streamline insurance payment processes and minimize customer inconvenience by implementing AI in areas directly experienced by customers, such as claims and assessments.
A prime example is Kakao Pay Insurance's overseas travel insurance. When customers submit receipts or other documentation for injuries or illnesses incurred abroad, the company's proprietary AI-based document recognition technology, known as 'Document AI,' identifies the type of document and the necessary information to process claims. In cases of flight delays, the technology analyzes relevant data in real-time to automatically issue payments without requiring customers to file a separate claim.
Kim Hee-jun, the Chief Technology Officer (CTO) of Kakao Pay Insurance, stated in a recent interview with Aju Economy, "In the AI era, companies must have the infrastructure and cloud-based service environment to flexibly process data. Ultimately, the ability to connect these capabilities to actual services will determine a financial institution's competitive edge."
The following is a Q&A with Kim, CTO of Kakao Pay Insurance.
- What is the immediate payment system based on optical character recognition (OCR) technology that you introduced for overseas travel insurance?
"OCR is a technology that recognizes text from images. However, it has limitations, such as significantly reduced recognition rates when images are distorted or contain handwritten text. Kakao Pay Insurance has integrated large language model (LLM) technology, which excels at understanding context and making inferences. When applied to the image recognition process, even if some characters are not recognized, the context of the document allows for accurate content inference.
For instance, in mobile phone insurance, we receive various types of receipts, not just standardized documents issued by Samsung Electronics Service. Previously, we had to pre-define and train on each document format to recognize them. With LLM technology, we have developed a system that can autonomously determine the type of document and recognize the information contained within it, as long as the necessary elements are present at a sufficient level. Importantly, we developed this technology in-house within six months, without relying on external commercial solutions."
- How much faster has the payment speed for claims improved since implementing this technology?
"In mobile phone insurance, claims can be paid immediately upon submission of sufficiently trained receipts. In fact, there have been cases where payments were made in as little as one second. For overseas travel insurance, we have taken it a step further by implementing a system that pays claims before customers even file them. We collect and analyze real-time flight operation data from various sources, including open APIs from airport authorities and global aviation information companies. When our internal algorithm confirms a flight delay, we automatically issue payments to affected customers.
In one instance, a customer received 40,000 won after discovering that their flight had been delayed only after the payment was made. This way, customers can receive payments without having to file a claim, and insurers can reduce the resources allocated to claim assessments."
- Why did you choose to develop this technology in-house rather than using external solutions?
"Relying on external solutions for core technology makes it difficult to respond quickly. If document formats change or customer feedback arises, it can take weeks or even months to implement those changes. In contrast, having our own technology allows us to address customer issues identified in the morning by making improvements by the afternoon. This responsiveness is a key competitive advantage. Therefore, we deemed it essential to internalize this technology and ensure that our in-house developers could provide it."
- How do you build the workforce and systems necessary for in-house technology development?
"As a digital insurance company, we believe that information technology (IT) is our core competitive advantage. More than 50% of our employees are developers. We also manage specialized areas like cloud operations internally, rather than outsourcing to external managed service providers (MSPs). Many of our developers have experience not only in traditional insurance and financial companies but also in general IT and service firms.
While it is common to separate developers responsible for core systems and those developing customer-facing channels, at Kakao Pay Insurance, these teams work together. Developers with insurance domain knowledge collaborate with those experienced in IT services, allowing for the rapid integration of new ideas and technologies into insurance services."
- What do you see as the technological competitiveness of the insurance industry in the AI era, and what is Kakao Pay Insurance's unique advantage?
"Just two to three years ago, many financial companies emphasized digital transformation (DT) as a process to create and utilize data. Now, most are discussing AI transformation (AX). Ultimately, companies need to have the infrastructure and cloud-based service environment to flexibly process data and the capability to leverage AI based on that. This must be connected back to customer experience and feedback to create a virtuous cycle.
As a new insurance company, Kakao Pay Insurance has a cloud-native infrastructure and internalized development personnel. Our core competitive advantage lies in our ability to flexibly combine this technology with the insurance domain to create new services."
- What is your ultimate goal?
"I hope customers perceive Kakao Pay Insurance as a company that offers innovative, flexible products and services that prioritize customer experience, distinguishing us from other insurers. Additionally, insurance often includes terms like 'effectiveness,' 'renewal,' 're-enrollment,' and 'expiration,' which can be difficult for the average customer to understand. We aim to make insurance more accessible and ensure that customers have choices."
A prime example is Kakao Pay Insurance's overseas travel insurance. When customers submit receipts or other documentation for injuries or illnesses incurred abroad, the company's proprietary AI-based document recognition technology, known as 'Document AI,' identifies the type of document and the necessary information to process claims. In cases of flight delays, the technology analyzes relevant data in real-time to automatically issue payments without requiring customers to file a separate claim.
Kim Hee-jun, the Chief Technology Officer (CTO) of Kakao Pay Insurance, stated in a recent interview with Aju Economy, "In the AI era, companies must have the infrastructure and cloud-based service environment to flexibly process data. Ultimately, the ability to connect these capabilities to actual services will determine a financial institution's competitive edge."
The following is a Q&A with Kim, CTO of Kakao Pay Insurance.
- What is the immediate payment system based on optical character recognition (OCR) technology that you introduced for overseas travel insurance?
"OCR is a technology that recognizes text from images. However, it has limitations, such as significantly reduced recognition rates when images are distorted or contain handwritten text. Kakao Pay Insurance has integrated large language model (LLM) technology, which excels at understanding context and making inferences. When applied to the image recognition process, even if some characters are not recognized, the context of the document allows for accurate content inference.
For instance, in mobile phone insurance, we receive various types of receipts, not just standardized documents issued by Samsung Electronics Service. Previously, we had to pre-define and train on each document format to recognize them. With LLM technology, we have developed a system that can autonomously determine the type of document and recognize the information contained within it, as long as the necessary elements are present at a sufficient level. Importantly, we developed this technology in-house within six months, without relying on external commercial solutions."
- How much faster has the payment speed for claims improved since implementing this technology?
"In mobile phone insurance, claims can be paid immediately upon submission of sufficiently trained receipts. In fact, there have been cases where payments were made in as little as one second. For overseas travel insurance, we have taken it a step further by implementing a system that pays claims before customers even file them. We collect and analyze real-time flight operation data from various sources, including open APIs from airport authorities and global aviation information companies. When our internal algorithm confirms a flight delay, we automatically issue payments to affected customers.
In one instance, a customer received 40,000 won after discovering that their flight had been delayed only after the payment was made. This way, customers can receive payments without having to file a claim, and insurers can reduce the resources allocated to claim assessments."
- Why did you choose to develop this technology in-house rather than using external solutions?
"Relying on external solutions for core technology makes it difficult to respond quickly. If document formats change or customer feedback arises, it can take weeks or even months to implement those changes. In contrast, having our own technology allows us to address customer issues identified in the morning by making improvements by the afternoon. This responsiveness is a key competitive advantage. Therefore, we deemed it essential to internalize this technology and ensure that our in-house developers could provide it."
- How do you build the workforce and systems necessary for in-house technology development?
"As a digital insurance company, we believe that information technology (IT) is our core competitive advantage. More than 50% of our employees are developers. We also manage specialized areas like cloud operations internally, rather than outsourcing to external managed service providers (MSPs). Many of our developers have experience not only in traditional insurance and financial companies but also in general IT and service firms.
While it is common to separate developers responsible for core systems and those developing customer-facing channels, at Kakao Pay Insurance, these teams work together. Developers with insurance domain knowledge collaborate with those experienced in IT services, allowing for the rapid integration of new ideas and technologies into insurance services."
- What do you see as the technological competitiveness of the insurance industry in the AI era, and what is Kakao Pay Insurance's unique advantage?
"Just two to three years ago, many financial companies emphasized digital transformation (DT) as a process to create and utilize data. Now, most are discussing AI transformation (AX). Ultimately, companies need to have the infrastructure and cloud-based service environment to flexibly process data and the capability to leverage AI based on that. This must be connected back to customer experience and feedback to create a virtuous cycle.
As a new insurance company, Kakao Pay Insurance has a cloud-native infrastructure and internalized development personnel. Our core competitive advantage lies in our ability to flexibly combine this technology with the insurance domain to create new services."
- What is your ultimate goal?
"I hope customers perceive Kakao Pay Insurance as a company that offers innovative, flexible products and services that prioritize customer experience, distinguishing us from other insurers. Additionally, insurance often includes terms like 'effectiveness,' 'renewal,' 're-enrollment,' and 'expiration,' which can be difficult for the average customer to understand. We aim to make insurance more accessible and ensure that customers have choices."
* This article has been translated by AI.
Copyright ⓒ Aju Press All rights reserved.