Kashima City Becomes First in Japan to Adopt Claude, Cutting 590 Hours Annually and Accelerating Municipal DX
機械翻訳 / Machine-translated
A
AI Friends
2日前・10分で読める
機械翻訳 / Machine-translated
@aifriends
AI Friends（https://aifriends.jp）のクロスポスト公式アカウント。AIツールの紹介・使い方・できることを、中学生でもわかるやさしい日本語で届けます。
What you'll learn from this article
On April 1, 2026, Kashima City, Ibaraki Prefecture, launched full operation of an AI chatbot on its official website. The system uses Claude, a conversational AI developed by American AI company Anthropic. This marks the first time in Japan that a municipality has officially deployed an inquiry-response system powered by Claude.
Cellpromote Co., Ltd., which handled the development, built a mechanism that trains the AI on information published on the city's website. When residents ask questions such as "How do I sort my garbage?" or "Where can I get a copy of my residence certificate?", the AI answers based on the city's official information.
Before the full rollout, Kashima City conducted a 73-day proof-of-concept trial from December 18, 2025 to February 28, 2026. During this period, approximately 2,360 inquiries were handled by the AI chatbot.
According to the city's estimates, assuming staff spend three minutes per inquiry, approximately 118 hours of work time were saved during the trial period alone — equivalent to roughly ¥388,000. If usage continues at this pace for a full year, an annual reduction of approximately 590 hours, worth roughly ¥1,938,000, is projected.
In other words, by having AI answer basic questions from residents, staff are now able to focus on tasks that only humans can perform, such as complex consultations and in-person counter services.
In a post-trial survey, approximately 83% of users responded that the system was "helpful." Here are some of the specific comments received:
Particularly noteworthy is the ability to use the service at night and on weekends. If someone thinks on a weekday evening, "I want to go pick up my residence certificate tomorrow — what do I need to bring?", the AI chatbot can answer instantly even when city hall is closed.
Two key technologies power this system: Claude and RAG (Retrieval-Augmented Generation).
Claude is an AI (known as a large language model, or LLM) capable of natural, human-like conversation. It operates on a similar principle to ChatGPT but is said to deliver safer and more accurate responses.
RAG is a technology that allows the AI to search and reference pre-prepared, accurate information when formulating answers. In this case, the AI references the Kashima City website, which means it does not make up answers on its own, reducing the risk of conveying incorrect information.
AI chatbot adoption is progressing not just in Kashima City, but across municipalities throughout Japan. A survey conducted by the Ministry of Internal Affairs and Communications in October 2025 found that approximately 74% of municipalities nationwide have introduced AI or RPA (Robotic Process Automation), with around 62% of those having adopted AI specifically.
Shibuya Ward in Tokyo introduced a generative AI chatbot in March 2025, and within just two months, 100,000 questions had been submitted — 77% of which were resolved by the AI alone. In Saga City, phone inquiries decreased by 15% over the four months following its introduction in February 2025.
The primary reason municipalities are adopting AI is that it allows them to simultaneously address staff shortages and improve resident services. As Japan's population ages and birth rates decline, the number of municipal employees is also shrinking — while the volume of inquiries from residents continues to grow.
AI chatbots can handle routine questions 24 hours a day, 365 days a year, freeing staff to devote their time to specialized consultations and complex procedures. For working residents, the ability to access information at night and on weekends is also a significant benefit.
Kashima City's initiative serves as a model for other municipalities across the country. An era is beginning in which AI and humans working together can deliver better public services.
This article is a cross-post from AI Friends.
