SIGNATE Participates in FSA's Generative AI Pilot Project for Regional Financial Institutions

SIGNATE Inc. (Headquarters: Chuo-ku, Tokyo; President: Shu Saito), a comprehensive AI consulting firm that supports AI adoption from strategic planning to execution and talent development, announced its participation in the 'Pilot Research Project for DX in Regional Financial Institutions Using Generative AI,' planned by the Financial Services Agency (FSA) and conducted by the Financial Data Utilization Association (FDUA).

The project aims to create use cases for generative AI and establish utilization models that account for risk mitigation, in collaboration with financial institutions and AI providers.

Drawing on its track record in formulating strategies for generative AI and developing AI/DX talent for financial institutions, SIGNATE will contribute to this project by developing AI agents through a competition format, strongly promoting the societal implementation of generative AI.

## Overview

To promote the healthy utilization of generative AI among regional financial institutions and support their operational efficiency, this project will conduct pilot experiments on AI usage in banks, creating use cases for services such as customer support. The findings, including use cases and risk mitigation methods, will be organized as guidelines and development assets and released to support widespread adoption among regional financial institutions across Japan. (Expected participants: approx. 100 institutions)

## Expected Use Cases

### Customer Inquiry Support (AI Agent)
Based on a conversational AI agent that can handle broad inquiries in natural language, this will enable response generation using multiple data sources through agent collaboration. This aims to implement risk management for customer-facing generative AI, improve customer convenience, and reduce the workload of bank staff and call centers.

### Business Feasibility Evaluation (AI Agent)
Business feasibility evaluation is a high-level task assessing a company's future prospects and sustainability. Challenges include variations in evaluation quality, information collection/organization, and labor-intensive manual corrections due to inconsistencies in input data formats. The project aims to resolve these through AI agents with universal mechanisms independent of format, thereby supporting sales staff decision-making and improving operational efficiency.

## Project Structure

Given the rapid pace of generative AI progress, effective AI agent development requires diverse technical approaches. Therefore, this pilot will use a competition format to have multiple approaches compete, reducing failure risk and identifying best practices. Furthermore, by incorporating diverse opinions from financial institutions through surveys and releasing open-access-based deliverables, the project aims to avoid vendor lock-in.

## Drafting AI Agent Implementation Guidelines (Tentative)

As AI agents are still in the early stages of adoption, their introduction and risk management could be burdensome for regional banks. Based on pilot experiment findings, these guidelines will outline approaches for use case deployment and risk management, promoting healthy AI adoption and strengthening 'regional financial capacity' through operational efficiency.