November 19, 2026, 7:00 a.m. (JST)
[Japan] On November 17, 2026, Mitsubishi UFJ Financial Group (MUFG) announced a comprehensive partnership with Matsuo Research Institute and Third Intelligence.

In addition to advancing the research, development, and implementation of Japan’s next-generation AI—“ubiquitous AGI”—in financial services, the two organizations will collaborate to support the growth of startups incubated by the Matsuo Institute and to develop AI talent at MUFG.
The goal is to realize financial services that accompany customers throughout their lives and business operations by gaining a long-term understanding of their past interactions and circumstances.
During a press roundtable following the press conference, details emerged regarding the status of pilot projects in the inheritance field—an area where the company is leading the way—as well as challenges in integrating the technology into services.
MUFG is pursuing a shift from AI as a mere “tool” for answering questions to AI that understands customers, makes proposals, and “grows together” through repeated dialogue.At a press conference held on the 17th, MUFG Group CEO Junichi Hanzawa explained the company’s strategy to cultivate Japanese-developed AI in the financial sector and translate it into new customer value.
This partnership combines MUFG’s financial capabilities and customer base with Matsuo Research Institute’s expertise in AI research and talent development, as well as Third Intelligence’s proprietary AI development capabilities. MUFG had previously announced in 2025 that it would invest in Third Intelligence and sign a memorandum of understanding regarding collaboration; this comprehensive partnership expands the scope of that collaboration.
Inheritance AI Concierge: Approximately Six-Month Pilot Underway
The first concrete customer service to be launched is the AI concierge concept in the inheritance field. Through dialogue with users and their families, the service will help organize necessary preparations and identify challenges, while facilitating discussions among family members. The concept involves translating the key points identified during these discussions into concrete next steps and linking them to MUFG’s inheritance-related services.
At a press conference, MUFG Executive Managing Director Tadashi Yamamoto (Head of the Retail & Digital Business Division and Group CDTO) explained that while conventional AI tends to provide generic responses, inheritance matters involve circumstances that differ from family to family. He expressed the view that if the system can retain a record of family discussions and decisions over an extended period, it will lead to proposals tailored to individual circumstances.
During a press briefing, he explained that the company had been conducting verification tests on initiatives for inheritance planning for approximately half a year. He also mentioned the possibility of integrating the AI concierge into remote inheritance-related services currently under consideration. However, specific integration methods and launch timelines have not yet been determined, and the company is currently in the process of evaluating the results of the pilot program before making a decision. The customer-facing service has not yet been finalized.
While leveraging long-term memory, data protection is a prerequisite for practical implementation
Third Intelligence’s “ubiquitous AGI” is based on proprietary algorithms under research and development that differ from the Transformer architecture, which is currently the mainstream in generative AI. The company highlights three key features: “never forgetting” through long-term memory and continuous learning; “immediate understanding” through adaptation from small amounts of data; and “never stopping,” which aims for high-speed inference and dialogue.
At the press conference, the company also demonstrated an example of speech recognition technology that identifies technical terms and proper names based on the memories of users and organizations. However, during a follow-up interview, company representatives explained that the ubiquitous AGI is still in the research and development phase, and that the term is being used to describe a concept they aim to realize. This does not imply that general-purpose artificial intelligence has been completed at this stage.
Long-term recording of highly sensitive information, such as assets and family relationships, also poses security challenges. During the Q&A session, Mr. Yamamoto described the promotion of AI and security measures as “two wheels of the same cart,” and expressed the company’s intention to proceed with demonstrations while ensuring safety so as not to undermine the trust placed in it as a financial institution.The plan is not to expand the scope of implementation all at once, but to first conduct verification in the field of inheritance.
Envisioning “Lifetime Accounts” and “Mutual Prosperity Accounts”: Exploring Applications for Financial AI
As a medium- to long-term service vision, the bank has proposed a “Lifetime Account” for individuals—designed to support them through life’s milestones such as employment, home purchases, asset accumulation, retirement, and inheritance—and a “Mutual Prosperity Account” for corporations, which supports business management from startup through fundraising, growth, and business succession.Both are future visions that prioritize ongoing relationships with customers, and no decision has yet been made regarding their specific commercialization.
MUFG is also advancing collaborations with Sakana AI in the lending sector, integrating with OpenAI’s digital financial services, and partnering with Google on financial and non-financial services. Regarding this technology, Mr. Yamamoto expressed high hopes for its ability to provide personalized recommendations for complex inquiries by leveraging long-term memory.
During a press interview, he also mentioned the development of agent-based commerce with Google and explained a concept to connect business companies’ apps with MUFG’s financial AI agents. He stated that the long-term goal is to create a system where business companies can integrate with MUFG’s financial functions without having to develop their own financial AI.
Supporting Japan’s AI Industry Through Finance
Another pillar of the partnership is support for the Matsuo Research Institute and startups incubated by the institute. Through the assignment of dedicated staff, MUFG will provide financial backing covering everything from account opening and fundraising to business growth, M&A, and exits. It will also facilitate connections with client companies to help translate research findings into practical applications in society.
Furthermore, by leveraging the Matsuo Institute’s AI talent development programs and strategic advisory services, MUFG aims to become “AI-native” itself. The plan is not merely to adopt technology but to simultaneously advance the development of Japan’s AI industry and drive organizational transformation within financial institutions.
This partnership integrates the development of Japanese AI with its implementation in financial services. The next key focus will be whether the inheritance AI—which serves as the first litmus test for practical application—can successfully balance the ability to address individual circumstances with data protection.


















