.jpg)
Tokyo Metro, partnering with Allganize, will deploy generative AI solutions to enhance customer-facing chatbots and customer service operations. Using advanced LLM technology, this initiative aims to improve response accuracy, streamline lost item inquiries, and automate email replies, improving both customer convenience and operational efficiency. It's a pioneering railway AI system.
Tokyo Metro will collaborate with Allganize , a provider of all-in-one LLM solutions that revolutionize corporate productivity with AI, to introduce a system utilizing generative AI for customer-facing chatbots and customer service center operations. This initiative aims to improve customer convenience when making inquiries and to enhance operational efficiency. The introduction of a dedicated generative AI system for both customer-facing chatbots and customer service centers is the first such initiative by a railway company.
Background of the initiative:
Tokyo Metro's customer service center handles a wide variety of inquiries from customers via phone, email, and chatbots, with the number of inquiries reaching approximately 250,000 annually. Through the implementation, the goal is to provide prompt responses to various customer inquiries.
Overview of the Initiative Utilizing Generative AI:
1. Enhancing the Functionality of Customer-Facing Chatbots
The current chatbot on Tokyo Metro's official website has been designed to respond to customer inquiries based on predefined FAQs. However, it has struggled to address the increasingly diverse range of customer questions. To address this, generative AI will be utilized in addition to FAQ functionality, enabling the chatbot to generate appropriate responses using information from the official website and other sources. This will expand the scope of inquiries the chatbot can handle and improve customer convenience.
2. Improvement of Inquiry Methods for Lost Items
In traditional email inquiries regarding lost items, multiple confirmations were sometimes required to gather sufficient information for the search. Moving forward, the chatbot will allow customers to input necessary information in a conversational format tailored to the characteristics of the lost item, greatly reducing the effort required from customers and customer service employees.
3. Development of a Support Solution for Responding to Emails Sent to the Customer Service Center
For inquiries received via email, generative AI will be used to understand the content, search for the necessary information, and draft response proposals. By automating the series of tasks previously performed manually by operators, operational efficiency will be improved, enabling quicker responses to a greater number of inquiries.
Tokyo Metro, in collaboration with Allganize, will continue to listen to customer feedback and strive to provide higher-quality services, making the subway easier to understand and more comfortable to use.
Development of a Dedicated System for Railways and Call Centers Utilizing Generative AI:
This initiative will leverage the generative AI applications from Allganize's Alli LLM App Market to achieve the following:
1. Enhancement of chatbot functionality,
2. Improvement of inquiry methods for lost items, and
3. Development of a support solution for responding to emails.
1. Automatic Response Generation and Display of Reference Information Using LLM (Large Language Models) and RAG (Retrieval Augmented Generation) Technologies
By utilizing the Alli LLM App Market, the range of issues customers can resolve immediately will be expanded. The Alli LLM App Market allows for the selection of the most suitable model from OpenAI's GPT series, Microsoft's Azure OpenAI models, and Google's Gemini series, depending on the application. It also implements a proprietary RAG system, currently patent-pending, to improve response generation accuracy. This system is capable of understanding documents, including charts and handwritten content.
The current chatbot on Tokyo Metro's official website is FAQ-based and limited to responding to general inquiries, resulting in a restricted response rate. The Alli LLM App Market incorporates advanced AI technologies capable of accurately understanding the meaning and intent of freely entered questions, thereby improving self-resolution rates. For questions that cannot be addressed by FAQs, the AI will search for appropriate information from the Tokyo Metro Group's official website and the LLM will automatically generate responses. This will expand the scope of inquiries the chatbot can handle and improve customer convenience.
2. Improvement of Inquiry Methods for Lost Items
Previously, email inquiries regarding lost items often required multiple follow-ups to gather sufficient information for item searches. Going forward, a chatbot will be used to allow customers to input the necessary information in a conversational format tailored to the characteristics of the lost item, reducing the effort required from customers. Additionally, this will improve the operational efficiency of customer service center operators, enabling quicker responses to a larger number of inquiries.
3. Email Response Support Solution
For inquiries received via email, generative AI will be utilized to automate understanding of the content, searching for the necessary information, and generating response drafts. The information referenced will include the Tokyo Metro Group's official website, records of responses to similar past inquiries, and various documents. Based on this, the system will be able to automatically generate appropriate response drafts for a larger number of inquiries. This will improve the operational efficiency of operators, enabling quicker responses to a greater
number of inquiries.
About the All-in-One Generative AI and LLM Application Platform "Alli LLM App Market"
The Alli LLM App Market is a generative AI and LLM application platform that provides all-in-one solutions for corporate use, including "over 100 pre-built LLM applications without the need for prompts," a "no-code LLM app builder," "integration with company data," and "security."
As the use of generative AI and LLM applications in business operations grows, challenges such as "uncertainty about which tasks to apply them to," "difficulty in creating effective prompts," and "lack of results with LLM applications that are not tailored to specific business operations and data" are becoming increasingly common. The Alli LLM App Market comprehensively addresses these issues, enabling effective utilization of generative AI and LLM in enterprises.
LLM applications can be used without requiring prompts and come pre-equipped with various tools that can be immediately applied to business operations, such as "automatic response generation from documents," "contract clause review," "drafting customer response emails," "summarizing specific documents," and "report creation." The platform also integrates with multiple internal data sources and business systems, optimizing not just individual tasks but entire workflows through generative AI and LLM.
The platform supports multiple LLM models, including OpenAI's GPT series, Microsoft's Azure OpenAI models, Google's Gemini series, and Anthropic's Claude 3 series, allowing users to select the model best suited for their specific needs. New application creation, customization, and data integration can be performed easily without coding, making the platform accessible even to those without specialized knowledge of LLMs. For clients with specific needs, the platform also offers development of industry- or company-specific small-scale LLMs (sLLMs) and provides on-premises versions.
About Allganize
With the vision of "automating and optimizing all business workflows through AI," Allganize provides all-in-one LLM solutions for enterprises globally, focusing on Japan, the United States, and Korea. Founded by members with extensive experience in the global market and supported by domestic and international investors, Allganize operates from its bases in Tokyo (Japan), Houston (USA), and Seoul (Korea).