Google Cloud and Shinhan Investment, Complete AI Agent Platform Construction

This page was machine-translated and may differ from the original. View original

Google Cloud and Shinhan Investment, Complete AI Agent Platform Construction

Published2026.08.18 09:24

Hybrid Orchestrator Self-Design, Over-the-Counter Derivatives Processing Time Reduced from 2 Hours to 5 Minutes

A case of applying AI agents to actual business systems in a financial regulatory environment has emerged domestically. Shinhan Investment, in collaboration with Google Cloud, has completed a next-generation AI agent platform that complies with financial security regulations such as network segregation while integrating with internal core business data and systems. Practical results have been confirmed, including reducing over-the-counter derivatives contract processing time from 2 hours to 5 minutes per case.

Google Cloud announced on the 18th that Shinhan Investment has built an AI agent platform based on its cloud infrastructure. Shinhan Investment established an AX headquarters to promote AI transformation (AX) by 2026, selected 'AX coordinators' by department, and conducted business-led agent development, generating over 500 agents. This platform was built to organically integrate these agents with internal key business data and financial systems.

Application of Hybrid Orchestrator and Data Flywheel Structure

While utilizing the Gemini Enterprise Agent Platform, Shinhan Investment self-designed and applied a 'hybrid orchestrator' structure tailored to the financial environment. For tasks requiring accuracy and consistency, such as fund transfers, the system processes based on pre-defined rule sets, and calls LLMs such as Gemini only in areas requiring reasoning, such as contract analysis, thereby securing both stability and flexibility, according to the company. Additionally, the company has implemented a 'data flywheel' structure that accumulates business processing results and feedback in BigQuery and similar platforms for continuous agent utilization.

In terms of security, the company went through relevant procedures including innovative financial services approval, CSP safety assessment, cloud usage reporting, and security due diligence. The company maintains network segregation through dedicated lines between the internal network and Google Cloud and VPC service controls, and has applied LLM utilization area segregation and data loss prevention (DLP) solutions.

Over-the-Counter Derivatives Contract Processing Time Reduced from 2 Hours to 5 Minutes

The platform is currently operating as a closed beta for some employees and is preparing for official launch. The representative achievement came from the over-the-counter derivatives contract processing task. Previously, 3 to 4 staff members manually reviewed unstructured contracts of approximately 50 pages and then manually entered about 50 items into the internal system. After introducing the agent, the process was automated from email selection to system entry, reducing the processing time per case from 2 hours to 5 minutes. Shinhan Investment plans to increase the automation rate of agent-based tasks to over 70% focusing on tasks targeted for automation.

Noh Hyun-bin, Head of AX Planning at Shinhan Investment, stated, "We were able to complete the platform construction in just 3 months based on the Google Cloud environment, whereas the existing development method would have required a minimum of 1 year or more."

Lee Seon-hun, CEO of Shinhan Investment, said, "We will pioneer a new direction for Korea-style financial AX based on the agent platform."

Ruth Sun, President of Google Cloud Korea, said, "This case can provide direction to financial institutions looking to apply AI agents to actual operations in regulated industries."

#클라우드 #LLM #AI 에이전트 #구글 클라우드 #업무 자동화 #Gemini #신한투자증권 #금융 자동화 #하이브리드 오케스트레이터 #금융 규제

본 기사에 대한 정정·반론·추후보도 청구는 보도 청구 안내를, 그간 게재된 보도문은 정정·반론보도 모아보기를 참고해 주세요.

명세환 기자

All articles by this reporter

관련뉴스

Google Cloud Supports AI Security Innovation with East Security's "Al Yak Family Care"

Google Cloud supported the innovation of ESTsecurity's AI family security platform 'Alyac FamilyCare'. The company implemented an AI-native security system that proactively responds to social engineering attacks through a Vertex AI, Gemini, and BigQuery-based architecture.

2025.11.25by 명세환 기자

Google Cloud Introduces Real-Time Commentary Service for MLB Apps… Analyzes Game Context with Gemini

Google Cloud has partnered with Major League Baseball (MLB) to introduce 'MLB Scout Insight,' a generative AI-based real-time commentary service. The service is integrated into the 'Game Day' feature within the official MLB app, providing real-time information on player matchups, statistics, and unique facts during games. Utilizing Gemini, BigQuery, and AlloyDB, the system pre-generates insights tailored to the day's game flow and connects relevant information within two seconds of an actual event occurring. The service also incorporates a concept of "unexpectedness" that selects content likely to captivate fans, going beyond simple statistical delivery.

2026.04.06by 명세환 기자

Google Cloud Provides Conversational AI to Weverse Customer Support

Weverse has applied Google Cloud's conversational AI technology to its customer support system to address the increasing volume of global fan inquiries. The new system was built as an automated response mechanism to handle 24-hour processing of inquiries related to ticket reservations, MD purchases, and platform features. Following its earlier migration of the data analytics platform to Google Cloud BigQuery, Weverse has now begun utilizing AI technology in the customer support domain. Since its introduction in March, the system has processed inquiries from 245 countries and regions, with plans to increase processing efficiency throughout the year.

2026.05.06by 배종인 기자

NetApp and Google Cloud Expand Integrated Storage Services… Lowering the Burden of AI Data Movement

NetApp and Google Cloud have officially launched new data management and storage capabilities to support enterprise AI workload operations. This announcement focuses on unified support for file and block workloads and streamlining data movement across multiple clouds. Both companies are committed to enabling enterprises to operate high-performance workloads such as databases, HPC, EDA, and VMware in the Google Cloud environment without requiring significant modifications to existing application architectures.

2026.05.19by 배종인 기자

NetApp and Google Cloud Officially Launch Integrated Storage Service

NetApp and Google Cloud have officially launched an integrated service that enables file and block workloads to operate in a single storage environment. The two companies plan to reduce enterprise cloud migration and AI workload operational burdens through the Flex Unified service based on Google Cloud NetApp Volumes and NetApp Data Migrator. This announcement is significant in that it enables enterprises to operate high-performance workloads such as databases, HPC, EDA, and VMware in cloud environments without major changes to existing application architectures.

2026.06.04by 배종인 기자