National Credit Card Center Develops Local AI Agent, Aims for Proof-of-Concept by Year-End
The National Credit Card Center (NCCC) is developing a local AI agent product search service, focusing on business payment applications, with a goal to complete proof-of-concept by the end of this year. This initiative responds to the active expansion of AI Agent transaction applications by international card issuing organizations. Mastercard has already completed Taiwan's first identity-verified AI agent transaction in March, facilitating travel arrangements and payments through an AI conversational interface.
📋 Article Processing Timeline
- 📰 Published: April 15, 2026 at 20:20
- 🔍 Collected: April 15, 2026 at 20:32 (11 min after Published)
- 🤖 AI Analyzed: April 15, 2026 at 20:33 (1 min after Collected)
TAIPEI, April 15 (Central News Agency) - International card issuing organizations are actively expanding AI Agent transaction applications. Lu Hui-jung, General Manager of the National Credit Card Center (NCCC), said today that the NCCC is researching and developing a local AI agent product search service, focusing on the business payment sector, and expects to complete a proof-of-concept (POC) by the end of this year. If the verification proves feasible, the next step will be to consider commercialization.
The NCCC and Mastercard jointly held the "AI Era Payment Security and Data Trust Summit" today.
Lu Hui-jung revealed in a media interview before the summit that AI agents are becoming increasingly popular, attracting market attention. To align credit card application scenarios with future development trends, the NCCC has begun researching and developing a "local AI agent," focusing on the business payment sector, and expects to complete a proof-of-concept (POC) by the end of this year. If the verification proves feasible, the next step will be to consider commercialization.
Proof-of-concept is a small-scale implementation test for an idea or technology to prove its feasibility and potential value, commonly seen in the early stages of project development.
Regarding actual application scenarios, Chen Yi-wen, Taiwan Country Manager for Mastercard, explained that Mastercard completed Taiwan's first identity-verified AI agent transaction in March. Users input their demand for airport transfer services for a trip to Hong Kong through an AI conversational interface, and the AI agent searches for the most favorable and suitable solution. After user confirmation, the booking and payment are completed using the linked credit card.
Regarding security control, Chen Yi-wen explained that many aspects of AI Agentic Commerce require certification. First, it must be confirmed that the demand is initiated by the consumer. Second, the AI agent accepting the demand must be certified, not just any AI agent. Third, the transaction must be tokenized to confirm that the AI agent is acting on behalf of the consumer. Finally, if a dispute arises, there must be a complete resolution plan.
The NCCC and Mastercard jointly held the "AI Era Payment Security and Data Trust Summit" today.
Lu Hui-jung revealed in a media interview before the summit that AI agents are becoming increasingly popular, attracting market attention. To align credit card application scenarios with future development trends, the NCCC has begun researching and developing a "local AI agent," focusing on the business payment sector, and expects to complete a proof-of-concept (POC) by the end of this year. If the verification proves feasible, the next step will be to consider commercialization.
Proof-of-concept is a small-scale implementation test for an idea or technology to prove its feasibility and potential value, commonly seen in the early stages of project development.
Regarding actual application scenarios, Chen Yi-wen, Taiwan Country Manager for Mastercard, explained that Mastercard completed Taiwan's first identity-verified AI agent transaction in March. Users input their demand for airport transfer services for a trip to Hong Kong through an AI conversational interface, and the AI agent searches for the most favorable and suitable solution. After user confirmation, the booking and payment are completed using the linked credit card.
Regarding security control, Chen Yi-wen explained that many aspects of AI Agentic Commerce require certification. First, it must be confirmed that the demand is initiated by the consumer. Second, the AI agent accepting the demand must be certified, not just any AI agent. Third, the transaction must be tokenized to confirm that the AI agent is acting on behalf of the consumer. Finally, if a dispute arises, there must be a complete resolution plan.