The AI wave continues to transform various industries, and Taiwan's financial sector has officially entered the era of "AI agents." On the 5th of this month, Fubon Financial Holding, in partnership with Microsoft Taiwan, hosted the first Copilot Studio hackathon in Taiwan's financial industry. The event, using low-code tools, encouraged employees to independently develop AI agent applications. The event attracted 21 teams and 65 employees from the financial holding company and its subsidiaries, showcasing the innovative potential of financial AI, moving from passive assistants to active agents.
From AI Assistant to AI Agent
Han Weiting, general manager of Fubon Financial Holding, said that according to the "Chief Economist Outlook" released by the World Economic Forum (WEF) in May this year, nearly half of economists believe that AI will have a disruptive impact on business by 5.
This hackathon not only continued the application of M365 Copilot last year, but also further introduced Copilot Studio, pushing AI from a simple personal assistant to an agent that can proactively complete specific tasks and serve departmental needs.
Li Qian, General Manager of Microsoft Taiwan's Enterprise Business Group, noted that Copilot Studio has been adopted by over 16 companies worldwide. Through its low-code platform, businesses can quickly build contextual AI agents, with applications ranging from intelligent customer service to automated operations and risk detection. For the financial industry, this means AI tools are no longer just auxiliary tools, but are now directly integrated into the core of services and operations.
Copilot Studio Practical Application
Copilot Studio can combine custom knowledge sources, customized answers, and Power Automate processes, and be deployed on a designated platform to help companies further optimize their business processes.
Fubon has participated in the Microsoft M2023 Copilot Preview beta program since 365 and has imported thousands of licenses to date, with over 30% actively using them. Internal surveys show that most employees believe Copilot saves 3 minutes to 5 hours of work per week, with some even saving over XNUMX hours. The introduction of Copilot Studio symbolizes the advancement of AI from a "support tool" to an "agent role."
Diverse and creative topics demonstrate the potential of AI
The participating teams proposed specific solutions to their respective pain points based on their respective departmental needs:
• Fubon Securities has developed “AI-automated thematic investment portfolio generation”, which allows users to obtain customized investment advice using natural language input.
• Taipei Fubon Bank has created an AI tool to solve the problem of unstructured scanned documents in new employees' health checkup reports.
• Fubon Life Insurance has designed AI that can autonomously learn to identify new types of surgeries to help make claims review more efficient.
• The financial holding company’s ESG team focuses on sustainable project management, improving collaboration efficiency through AI-automated emails and cross-departmental tasks.
These solutions not only demonstrate innovative thinking in financial services, but also demonstrate that AI can already deliver practical value in professional scenarios.
Building an AI ecosystem for the financial industry
In the past two years, Fubon has held three generative AI hackathons, generating over 23 team proposals and successfully incubating 7 application services. In July of this year, the company brought together its subsidiaries, including life insurance, banking, property and casualty insurance, securities, and investment trust companies, to form a "Generative AI Application Promotion Team" to accelerate cross-sector collaboration. In September, Fubon will also launch its fourth annual Fubon TechDay, an innovative technology matchmaking event, inviting innovative companies to collaborate on the development of practical financial AI applications.
Opinion: AI is getting serious, and the challenge has just begun
Fubon's collaboration with Microsoft not only allows financial industry employees to personally "train" AI agents, but also signifies that Taiwan's financial industry has moved from the experimental stage to large-scale application. However, for AI to truly be implemented, challenges still need to be overcome, including data governance, privacy compliance, and cross-departmental collaboration.
Whether AI agent services can become a "new colleague" in the daily operations of the financial industry depends not only on the tools themselves, but also on the reconstruction of corporate culture and processes.
