MongoDB is targeting South Korea's financial sector with expanded AI adoption while prioritizing data security and regulatory compliance, supporting on-premises deployments, and pursuing use cases such as fraud detection, document intelligence, and AI assistants.

The company introduced embedding models Voyage 4 and Voyage Multimodal 3.5, with Rerank 2.5 improving search relevance, and Voyage 4 enabling shared embedding spaces across model sizes to cut costs and latency.

Atlas Managed MCP Server lets AI agents run data queries, understand schemas, manage indexes, and perform control-plane tasks like cluster creation, configuration, and user access management through natural language, with OAuth-based access control and activity logging for audits.

Building a partner ecosystem is a priority to support implementation and deployment, pairing with product offerings through industry- and workflow-aware services.

Atlas Embedding and Reranking APIs unify the search layer by allowing embedding models to run directly in Atlas for vectorized search on MongoDB data, while preserving existing URLs, API keys, tokens, and model names to ease migration from external models.

MongoDB argues that real-time data access is the critical foundation for AI production, with data layer readiness identified as the main barrier to enterprise-wide AI deployment.

Public sector readiness includes procurement procedures and listing on South Korea's Nara Marketplace, enabling government adoption of MongoDB offerings for data modernization and AI initiatives.

South Korea is the primary market focus, with emphasis on real-time data integration from sensors and production systems in manufacturing, while plans aim to modernize legacy systems to broaden AI adoption across manufacturing, finance, and public sectors.

South Korean customer cases include LG Uplus using MongoDB Vector Search for an AI assistant, BC Card migrating an AI hot-deal service to Atlas, KEPCO KDN applying sharding to a large Meter Data Management System, Socar utilizing MongoDB Search for bookings, and Wrtn integrating AI character chat with MongoDB at scale.

The managed Atlas MCP Server eliminates the need for developers to host separate MCP servers by letting AI agents access and operate within MongoDB data environments via Atlas.