MediConCen logo

MediConCen: Increasing insurance claim processing efficiency and accuracy with Gemini

Google Cloud results
  • Realizes 98% field-level data extraction accuracy for medical insurance claims with Gemini

  • Reduces claim processing time from days to hours with 100% higher straight-through processing rate powered by Gemini and Vertex AI

  • Enables fast model fine- tuning for constant accuracy improvement with Vertex AI

  • Supports high-standard data security protection through Cloud Key Management

Using Google Cloud, MediConCen improved automation accuracy and speed of insurance claim processing to enable better services.

Improving medical insurance claim processing with automation

Processing medical insurance claims can be a long and tedious process. Traditionally, insurers need to read through a large number of files and analyze complex medical data, which is time consuming and prone to human error.

MediConCen was founded to address this pain point through automation. Using blockchain, the Hong Kong-based company has built a platform with the single source of truth architecture that allows insurance companies and medical institutions to exchange information seamlessly despite the different internal systems they use. The company has also implemented optical character recognition (OCR) to automatically extract text from images and digitize paper documents.

"Our mission is to automate medical insurance claim processing at the maximum level," notes William Yeung, CEO and co-founder of MediConCen. "That's why we've been constantly exploring and testing new technologies related to our industry."

As generative AI matured, MediConCen quickly recognized its potential to help advance the automation service it provides. Previously, however, OCR alone couldn't understand or summarize extracted text, meaning insurers still had to manually review every claim before making a decision. After comparing various generative AI models, MediConCen chose Gemini for its strong comprehension of medical content, as well as the enterprise-grade data security provided by Google Cloud.

"By combining Gemini's reasoning with specialized medical knowledge, it can reach more than 95% data extraction accuracy rate for insurance claim processing, while other models hit only 70-80%," explains Yeung. "Besides, Google Cloud provides unparalleled data security, which is paramount for our customers in the insurance sector."

MediConCen team members holding a phone and laptop
By combining Gemini's reasoning with specialized medical knowledge, it can reach more than 95% data extraction accuracy rate for insurance claim processing, while other models hit only 70-80%. Besides, Google Cloud provides unparalleled data security, which is paramount for our customers in the insurance sector.

Willian Yeung

CEO and Co-founder, MediConCen

Boosting insurance claim processing efficiency with 98% data extraction accuracy rate across languages

In April 2025, MediConCen launched its new insurance claim processing platform enabled by Google Cloud. To minimize the human review work required, MediConCen's DevOps team built an algorithm that acts as an AI manager through the agentic AI technology. The algorithm has broken down the process of handling insurance claims into hundreds of smaller tasks and orchestrates nearly 100 Gemini instances to perform them. This way, MediConCen is able to improve the field-level data extraction accuracy rate of its Gemini models to 98%, resulting in a 100% higher straight-through processing rate.

On top of that, MediConCen uses Vertex AI to fine-tune its generative AI models, further enhancing their comprehension accuracy rate for specific use cases. Yeung notes that the abundant computing resources provided by Vertex AI helps accelerate the fine-tuning process, which normally takes less than one day.

This means that MediConCen's new insurer customer can quickly leverage its automation service with high accuracy rates. Meanwhile, the multilingual capability of Gemini facilitates MediConCen's expansion to foreign markets, such as Vietnam, Indonesia, and Malaysia.

MediConCen team with inpatient claim management screen
Gemini and Vertex AI have significantly enhanced the automation level of our insurance claim processing service. Our customers now directly receive highly accurate claim summaries and decision recommendations from our system, enabling them to reduce claim processing time from days to hours and spend more resources on problematic cases.

Willian Yeung

CEO and Co-founder, MediConCen

Because of a higher efficiency and consistency in claim handling, MediConCen's customers have received less complaints from insureds. Furthermore, since MediConCen's Gemini models process and analyze all text and images related to claims, it can provide its customer with more comprehensive data and help them better detect potential fraud or abuse.

Preventing fraud while ensuring compliance and reliability

Data security and service reliability are crucial to MediConCen's insurer customers. The fact that Google Cloud meets the ISO27001 and SOC 2 Type II standards, and that customers' data isn't used to train Gemini, offers assurance. With Cloud Key Management, MediConCen gives customers full ownership and control of their keys that protect their data in Google Cloud through customer-managed encrypted keys, further strengthening security and confidence.

To ensure service reliability, MediConCen deploys its claim processing system on Cloud Run, which supports smooth CI/CD and high scalability. Yeung adds that since the company's adoption of Google Cloud, it has never experienced downtime of Gemini.

Next, MediConCen will continue improving its AI models' comprehension accuracy to support higher automation. It also plans to use BigQuery to detect fraudulent claims by identifying highly similar claim content.

"There's zero room for errors in processing insurance claims," notes Yeung. "With Google Cloud, we're able to realize impressive improvement in efficiency while ensuring accuracy and security. We're now well equipped to provide even better insurance claiming experiences."

MediaConCen team
With Google Cloud, we're able to realize impressive improvement in efficiency while ensuring accuracy and security. We're now well equipped to provide even better insurance claiming experiences.

Willian Yeung

CEO and Co-founder, MediConCen

MediConCen is an InsurTech company that automates insurance claims for insurance companies and medical institutions using blockchain and AI.

Industry: Healthcare and Life Sciences

Location: Hong Kong

Products: Google Cloud, Cloud Key Management, Cloud Run, Gemini, Vertex AI

Google Cloud