Osaka, Japan’s second-largest city, has begun testing autonomous AI agents in its local government to help alleviate the country’s declining workforce.
Osaka Prefecture has launched a public-private consortium to experiment with AI agents that provide administrative support and multilingual services.
The prefecture will bring together expertise from a consortium including Google Cloud Japan, telecommunications provider NTT West, Microsoft Japan, and Osaka Metropolitan University. This test assesses whether AI can accurately and independently streamline administrative processes based on predefined rules.
Osaka Prefecture Governor Hirofumi Yoshimura said the initiative was aimed at creating a “more convenient and prosperous society.” In Silicon Valley, AI agents are a technology that should be scaled, but in Japan, the main concern is minimizing disruption through standardization.
Japan’s AI agent boom
The new consortium in Osaka Prefecture follows a number of large household businesses that have begun deploying AI agents. Itochu Corporation, one of Japan’s largest food and beverage manufacturers, and carmaker Mazda are testing AI agents in automated payments, internal audits, and customer service.
Japanese software testing firm SHIFT and data analytics firm TDSE are also exploring payment ecosystems that leverage autonomous AI agents. According to TDSE, the proof of concept aims to initiate trades, validate requirements, and work with other systems to execute payments without direct human intervention.
In fact, a recent industry survey found that 35% of Japanese companies have already deployed some form of AI agent, and 44% plan to do so.
Japanese companies’ ambitions to develop AI agents are mostly reactionary and “defensive” measures. This represents economic acceptance of autonomous AI as a productivity tool amid labor shortages, rural depopulation, and declining tolerance for foreigners.
A hotbed of innovation? Think again.
Japan is not trying to win the race for the largest AI agent model. We are taking a slower, more deliberate, and more risk-averse path. Lux, a Tokyo-based accounting software provider, isn’t convinced that AI can do everything.
Chatbots have emerged as the latest sales and customer interaction tool in Big Tech and FinTech. However, the current capabilities of backend chatbots do not make life easier, according to Shinichiro Motomatsu, the company’s director and chief AI officer.
“If you try to handle expense reimbursement entirely through chatbot workflows, it’s probably going to be a hellish experience,” says Motomatsu.
Their main concern is implementing a system that increases the operational burden on a team that is already at capacity.
gray area of responsibility
Japan wants to safely deploy AI agents within real organizations. We focus on minimizing mistakes to maintain trust.
Japan’s approach is not a failure of imagination, but a deliberate response to how organizations actually work, according to Lux’s head of AI.
“At each stage, you need to evaluate whether the technology is really helping users. If it’s not, you should not hesitate to withdraw,” Motomatsu said.
Motomatsu said AI agents should be treated as goal-driven tools rather than stand-alone technical goals.
He believes it is much more realistic for AI agents to function as partially autonomous actors, as accountability is a big gray area.
“If something goes wrong, you can’t simply explain that AI made the decision,” Motomatsu says. “Someone within the organization has to be held accountable for the results.”
Human-centered AI agent
Lux’s chief AI officer emphasizes that AI does not eliminate the need for well-designed workflows and checks and balances. He argues that the real value of organizations lies in supporting, rather than replacing, the structures that make them work.
“AI agents are not magic; they do not eliminate the need for rules, processes, and human judgment,” Motomatsu says.
Osaka Prefecture’s push to introduce AI agents reflects the governance-first approach that is ingrained across Japanese companies. The plan is to formulate practical guidelines that can be imitated by local governments across the country by the end of fiscal 2026.
The prefecture aims to create a framework that outlines clear rules for what AI agents can do, how their actions will be monitored, and when human intervention is required.
There are also benefits beyond efficiency. Atsushi Yoshida, chief AI transformation officer at Hitachi, Ltd., suggests that AI agents could take over repetitive tasks and make way for the cognitive space.
“Just because AI allows us to do more work at a faster pace doesn’t necessarily equate to progress,” he says. “What matters is how people use the empty space.”
He describes this as a “white space” and argues that it can be a source of innovation, reflection, and decision-making.
Rather than introducing technological developments, Japan is designing AI agents to address corporate problems. In a business culture that traditionally prioritizes control and standardization, companies are wary of systems that obscure human judgment in mission-critical processes.
