Making sense of existing data is something that every financial firm struggles with, and Mizuho Bank Ltd is no exception.
The Japanese bank, which serves both retail and institutional clients, is using artificial intelligence (AI) and machine learning techniques to reduce costs for its existing business and to provide new solutions to its external users.
Tatsuya Shirakawa, senior digital strategist for the bank, said Mizuho began exploring initiatives using emerging technologies such as AI and the “activation of existing data” three years ago.
“Within the bank, there are diverse types of tasks. We had to determine how and where we can use AI, in the front, middle and back office, and whether it was to be in retail, institutional, or market operations. So first we decided where to use AI and as we did this, we had to think about what AI is,” he said.
Shirakawa, who was speaking at the Tokyo Financial Information & Technology Summit, which was held in Tokyo on July 4, said that to Mizuho, AI is a combination of data and machine learning.
Mizuho doesn’t yet have a systematic approach to AI, Shirakawa sad. The bank is first trying to enhance the literacy and understanding of its employees in AI and data. “They need to know the basics of those tools. We are trying to enhance the literacy of the members first,” he said.
Mizuho is currently using the latest AI algorithms to make use of unstructured data within the bank, which is a new capability for the bank. Among the tools it is dabbling in are image and video processing.
“Every day the bank is dealing with many forms. One of them is money transfer forms. This is manual work done by human beings. The format between forms is different too: it’s not standardized and it’s handwritten, so it can be hard for a machine to read. But using the latest AI recognition capabilities, data in an unstandardized form can be recognized. We are able to extract the necessary information and handwritten data using OCR [optical character recognition],” he said.
This process has automated 80% of that particular task, while the remaining 20% consists of manual checking and matching of the AI and OCR results with the bank’s existing customer base.
“By doing this, we have been able to improve our efficiency,” Shirakawa said.
In 2017, Mizuho, together with WiL LLC, which stands for World Innovation Lab, established Blue Lab to create new businesses based on technological advances. Blue Lab is based on the concepts of open innovation, creating platforms, global expansion, and an agile management style.
“AI is just a tool to enhance operations, but we first have to identify the problems. Another point that is important is that AI is a combination of data and machine learning, but data has great value. That’s a difference from the conventional IT investment because we have to now think about how to utilize the data of the user. By focusing on that, we work on AI and machine learning techniques and try to reduce costs for our existing business, and come up with new solutions to provide to external users,” Shirakawa said.
He said that Mizuho has not made the best use of a lot of data, particularly unstructured and investment data, but is now looking at how to collaborate with other financial institutions and banking groups to share know-how and information on improving efficiency.
The founder and CEO of Imperative Execution looks at how trade execution is changing and what that means for the buy side.Subscribe to Weekly Wrap emails
- Waters Rankings 2019: All the Winners
- Barclays Nearly Finished with First Major Quantum Computing Experiment
- Banks, Asset Managers Turn to Web Scraping to Generate Alpha
- The AI Ethics Dilemma: Banks Find a Fine Line Between New Tech and 'What's Right'
- State Street Explores Deeper Data Integration with Charles River