LG Uplus Works With MongoDB to Expand AI Services and Modernize Architecture

LG Uplus, a key subsidiary of LG Corporation and a leader in mobile, internet, and AI transformation, today announced that it will work with MongoDB to expand the use of generative AI and accelerate its modernization strategy across the company. 

Recently, I had the chance to connect with Sang-yeop Lee, CTO and Executive Vice President of LG Uplus, and we aligned on a plan to work together to build an AI-ready data architecture by modernizing legacy systems across the company.

LG Uplus's decision to make MongoDB its first choice data platform for AI workloads comes after the significant business impact of its early AI projects with its Agent Assist service, which LG Uplus reports has already helped improve efficiency by 30%.

AI business impact

LG Uplus operates one of Korea’s largest customer service organisations, with 4,000 human agents handling more than 3.5 million calls every month. 

As with many call centers, LG UPlus has dealt with a number of inefficiencies, from slow response times to difficulty retrieving relevant information. To support human agents and to deliver better customer outcomes, LG UPlus built “AI Agent Assist”—a groundbreaking tool that features real-time transcription, automatic categorization, and intelligent contextual summaries.

Early in the tool’s design phase, the team realized that the rigid tables and fixed schemas of relational databases would be fundamentally incompatible with the complex, unstructured, and rapidly changing data required for real-time AI. To solve this, LG Uplus built Agent Assist on MongoDB Atlas, which is powered by the document model. In addition to being intuitive for developers to work with, document databases adapt easily to the wide variety of constantly shifting data that AI depends on. 

By implementing MongoDB Atlas Vector Search, LG Uplus could co-locate vector embeddings with operational data in a single, high-performance database. This integrated platform strategy eliminated the need for fragmented "bolt-on" vector stores, significantly reducing architectural complexity for the service.

The Agent Assist service is now able to identify customer intent instantly during live calls and suggest the best response guidelines to agents. Agent assist was developed in just four months and became a key part of the company's AICC (AI Contact Center). Along with the 30% improvement in resource efficiency, LG Uplus also reports that the service reduced average call handling time by 7%.

This project demonstrated how a high-performing team combined with MongoDB’s best-in-class technology can set a new standard for AI execution that has real business outcomes.

But excitingly, this is just the beginning. The next evolution of AICC will see the tool learning by analyzing context and emotions in real-time during consultations to become 'self evolving'. LG Uplus also plans to extend its enhanced data management model—currently used in AICC—to AI services across the enterprise.

AI-driven modernization

As enterprises scale AI services, legacy IT environments often present bottlenecks in performance, flexibility, and integration. LG Uplus is no different.

In order to help reach its AI ambitions, the Korean telco giant will work with MongoDB on a large-scale data layer modernization, including the strategic intent to migrate a number of on-premise applications to a microservices architecture, powered by MongoDB Atlas. Not only is AI a driver of modernization, it is also an enabler – MongoDB's AI-enabled modernisation services will help LG U+ modernize faster and more efficiently than traditional methods.

With MongoDB powering the data layer, this new architecture will help improve operational efficiency, speed up time-to-market, increase accuracy, and deliver better customer experience.

“Building on our experience and success with MongoDB with AICC, we plan to expand this data management framework across our core AI services, including AI agents," said Sang-yeop Lee, CTO of LG Uplus. "By establishing an environment where operational data can be leveraged more quickly and effectively, we will further improve the service quality that customers can directly experience.”