Mizuho Financial Group is currently conducting trials with an IBM generative AI application designed to enhance recovery speed following IT outages.
In the event of a system disruption, accurate and timely recovery is critical for banking operations. However, when a fault is identified, operators often face a surge of messages and reports, complicating the process of diagnosing the issue and prolonging recovery times.
In a recent proof-of-concept (PoC), the generative AI solution achieved an impressive 98% accuracy rate in monitoring and addressing error messages over a three-month testing period. The PoC integrated patterns associated with potential errors into the AI, while also linking an application that assists in event detection with IBM’s advanced supercomputer, watsonx. This integration aims to streamline the recovery process and enhance response speed.
Additionally, the utilization of watsonx empowers Mizuho to manage internal operations more effectively, enabling on-site personnel to flexibly configure monitoring and operational menus when heightened security and confidentiality are necessary.
Looking ahead, Mizuho and IBM have plans to broaden the scope of the event detection and response proof-of-concept and transition it into production environments within the next year.
Mizuho has been a strong advocate for leveraging generative AI technology. The Japanese bank has collaborated with Fujitsu to incorporate generative AI into various workflows and tasks, including source code generation, system development, and maintenance. Last June, the bank took a significant step by rolling out Microsoft’s OpenAI platform to its workforce of 45,000 employees, thus embedding this technology throughout the organization.
For more insights on the evolving role of artificial intelligence in the banking sector, explore the opportunities and challenges at Finextra’s inaugural NextGenAI conference on November 26, 2024. Register your interest here.