**’Agent N’ Unveiled: An Integrated System Deeply Combined with Services**
**Verification with Metadata, Real Usage Reviews, and Inventory Interconnection**
(Seoul=Yonhap News) Reporter Kim Kyung-hee – “Beyond search to execution. The evolution of user experience.”
Kim Beom-jun, the Chief Operating Officer (COO) of Naver, at the integrated conference ‘Dan25’ held at COEX on the 6th, explained the direction of the integrated AI agent set to be introduced sequentially, stating, “In the era of artificial intelligence (AI), Naver will advance further with ‘Agent N’.”
Agent N is a user-customized integrated agent that combines Naver’s services, content, and data.
COO Kim introduced Agent N as an “agent that naturally suggests and executes based on a three-dimensional understanding of users,” highlighting its core as a ‘seamless’ experience where all services are organically connected.
This means that users do not need to input commands manually as services such as maps, calendars, reservations, and content naturally blend into the exploration process, with the AI agent providing recommendations and execution as needed.
Specifically, Naver explained that by utilizing large language model (LLM) technology, various service logs are comprehensively analyzed to predict user behavior and make appropriate suggestions.
Comprehensive understanding of user patterns to build a ‘persona’ and making proposals based on that persona is the approach it takes.
Agent N also aims to be an ‘executive agent.’
For example, with the ‘AI Tab’ set to be launched next summer, if a user inputs a ‘beginner running course,’ the system shows related course and review information based on place, cafe, and blog data, and allows users to subscribe to running specialist influencers or explore and purchase personalized products.
Naver claims that due to its possession of a service ecosystem spanning search, shopping, local, and finance, it can provide an experience that is executed immediately without external linking.
In terms of data reliability, the strategy is to enhance the reliability of the agent using ‘metadata.’
Metadata, as ‘data describing data,’ helps determine whether the content the AI references is based on real user experiences or autogenerated data.
COO Kim explained that reliable data is secured through reviews written by actual buyers and reservists, reservation statuses linked with POS systems, and inventory information connected with sellers.
COO Kim emphasized, “Such trust assets are essential in the AI agent era,” and “Naver will establish itself as the most trustworthy partner.”
kyunghee@yna.co.kr