Rakuten Ichiba’s AI Shop Manager will allow merchants to serve customers 24/7
Mickey Mikitani, Chairman and CEO, Rakuten Group
No matter how much AI transforms Rakuten Group’s 70+ businesses, we must never lose our human touch. There are so many scenarios where customers’ overwhelming preference is to deal with a human being. At the same time, humans – especially those running small and medium sized enterprises – cannot process tens of thousands of requests 24/7, 365 days a year. We will soon launch a major update on Rakuten Ichiba that will allow our merchants to ensure that with the help of AI, they can do just that.
Announced at the recent Rakuten AI Optimism 2026, AI Shop Manager promises to bring significant time-saving support to our merchant partners across the country.
Our AI Shop Manager will provide merchants with an always-on, intelligent member of staff who can engage directly with customers in a personalized way.
Each Rakuten Ichiba merchant can have its own AI Shop Manager equipped with the shop manager’s profile, personality, and persona, significantly improving merchants’ ability to handle complex inquiries at scale. An AI acting as a shopping agent shouldn’t just execute transactions; it needs to understand intent, make proactive suggestions, and maintain the unique characteristics of each shop.
Built differently
Rakuten has a major advantage across its businesses: its huge amount of ecosystem data – roughly 3 trillion transactions every year, a number that will only grow – that constitutes a deep ocean of insight that we are turning into new value for our customers.
The true power of Rakuten’s approach lies in the ability to leverage deep user data without prohibitive computational costs. It allows us to spot patterns to gain insight on products or services across our ecosystem that may be of interest to users. Instead of simply fetching generic product specifications from the web, AI Shop Manager will leverage Rakuten’s ecosystem to interpret a shopper’s broader ‘persona’ and tailor its responses accordingly. A user shopping for golf clubs, for example, will be recommended an entirely different set depending on whether their history shows they are a seasoned player or a complete beginner.
In a similar vein, Rakuten is leaning heavily on its proprietary data and developing its own Japanese-optimized LLMs to avoid escalating costs and security concerns. Our vast depository of data means that we can do a lot of our Al development in-house. Relying on proprietary third-party Al is not merely expensive – as much as ten times more than in-house Al – so we will develop whatever we can in-house. At the beginning of 2026, we released our newest large language model, Rakuten Al 3.0, development of which is ongoing. Beyond internal use, we also plan to power select portions of services with our models.
While still in development, this LLM will observe how a user’s purchasing trends have changed over time – even usage spanning almost two decades. It can better understand intent by taking into account spending in specific periods and the factors that drive these decisions.
It doesn’t end there. Our LLMs will know more about the user based on their usage across our services. Raising this granularity by even a few percent points produces very large, compounding effects.
We will soon have very exciting updates on AI Shop Manager, a service that will bring a totally new experience to e-commerce in Japan.
