How Japanese banking giant MUFG is using AI

Mitsubishi UFJ Financial Group (MUFG) is going all out to use artificial intelligence (AI) in all parts of its business, moving beyond using AI as a tool to deploying AI agents that will work alongside its employees.

In doing so, the Japanese banking giant aims to become an AI-native company. This vision was laid out by Morito Emi, head of MUFG’s digital strategy division, at the MUFG Fintech Festival in Singapore, where he spoke about the progress of the bank’s AI efforts.

Even though MUFG has deployed AI tools for all employees, only half regularly use them, he said. To address the gap and prepare for the future, MUFG is not only doubling down on efforts to spur greater AI adoption internally, but also paving the way for agentic AI.

“We are entering an era of AI agents, one where AI shifts from being just a tool to taking on actual roles as digital employees working alongside humans,” said Emi. “What will matter most is how we make use of our unique data.”

To realise its vision, MUFG is focused on three main priorities: company-wide AI training, creating an architecture for AI agents to operate effectively and safely, and enhancing its data strategy, where it is rebuilding its business model to make its data more accessible to AI agents.

In July 2025, MUFG started the Hello AI @ MUFG campaign to promote the use of AI as a core part of its corporate culture. Over 6,000 employees took part in the first event of the campaign, a prompt challenge where staff compete to write the most effective instructions for generative AI.

A key part of MUFG’s efforts to build up its AI capabilities is working with top AI technology suppliers. In October 2024, it teamed up with OpenAI to give its employees access to ChatGPT Enterprise. OpenAI will also help with developing advanced AI use cases and provide dedicated support.

This was followed by a three-year deal in May 2025 with Sakana AI, a well-known Japanese AI startup that became a unicorn in less than a year, to help in decision-making. Their first joint project is to develop an AI loan expert that can deliver better recommendations and learn from feedback provided by human bankers.

Meanwhile, the bank has been embedding AI directly into its business operations, with around 60 advanced use cases in the works that are expected to save about three million work hours a year.

One example is an AI tool that automatically generates proposals for corporate clients, freeing up bankers for more client-facing work. Another example is a mergers and acquisitions matching function that uses AI to find potential acquisition targets that human analysts might miss.

Its trust banking arm has also been using an AI document reader to pull information from complex legal and financial documents, saving thousands of work hours each year. It is also looking to triple the number of small-business loans serviced online by fiscal year 2026 by using AI for credit analysis.

All in all, Emi said MUFG’s AI strategy is not about replacing human workers. “We believe the most powerful combination is not human or AI, but human and AI, learning side-by-side from each other,” he said.

MUFG’s digital strategy division was formed in April 2025 to drive digital transformation, including the use of AI, across its business. Besides building local digital teams in the US and other parts of Asia, the division is also working with companies such as ride-hailing giant Grab that it has invested in through MUFG’s venture capital arm.