JR Central to test multilingual JRTok AI travel assistant at Shinagawa Station

A phone guide for the Tokyo to Osaka journey

Central Japan Railway Company (JR Central) will test a multilingual AI travel assistant at JR Shinagawa Station in Tokyo from December 15 to mid March, giving foreign visitors a new way to get help before boarding the Tokaido Shinkansen. Travelers will scan station QR codes with a smartphone to open a dedicated mobile website, without downloading an app.

The service, called JRTok AI, will answer practical questions in English, Chinese, Korean, French and Spanish. Its subjects include train operations, ticket purchases, seat types, boarding procedures and large luggage, choices that can be difficult for someone using Japan's high speed railway for the first time.

Shinagawa is a major entry point to the Tokaido Shinkansen, which connects Tokyo, Nagoya and Osaka. The corridor carries large crowds each day, with particularly heavy traffic during the year end holidays. Visitors often need help with the same decisions: where to buy a suitable ticket, which seat to choose and where a large suitcase can go.

JRTok AI is intended to put that guidance on a modern phone when a passenger needs it. The assistant will also use location based prompts, meaning that the guidance it brings forward can relate to where a traveler is in the station or on board a train. During the journey, optional English commentary will explain history and culture along the route.

JR Central says the trial results will guide improvements and possible expansion to other stations and routes. The winter test will give the company a chance to assess the service during a demanding travel period, when visitors need quick answers and station staff face heavy demand.

How passengers will open the assistant

Travelers will start by scanning a QR code displayed on station signs or at information counters. The code opens the mobile site in the phone's browser. After choosing a language, a user can type a question or tap a common topic rather than search through several pages of information.

The home screen will present needs such as buying a ticket, finding a platform and handling luggage. Users can search in their own language, with answers intended to be short and clear. Where useful, the assistant can point to maps or official pages that provide more detail.

JRTok AI will draw on official guidance about ticket options, reservations, baggage rules and directions inside the station. When schedules change, it can direct passengers toward the correct platform and provide updates taken from public notices. Bringing those instructions into a conversation could help a visitor who knows what they need to ask but does not know which menu or sign contains the answer.

Station staff will remain available at counters. The assistant is designed to support their work, with travelers still able to ask a person for help when a question needs more attention.

Making sense of tickets and seats

Ticket questions are a central part of the trial. The assistant can explain the differences between reserved seats, cars without seat reservations and Green Class. It can also describe how to choose a service with fewer transfers, helping visitors connect their ticket choice with the journey they want to make.

Passengers will be able to ask about purchasing tickets at machines, at counters or online. Guidance can cover the use of QR codes and IC cards where those options are available. For international travelers, the service can explain SmartEX and smartphone ticket options.

Travelers collecting physical tickets can also be reminded to bring the same credit card used for the purchase. That is a small detail with a practical consequence: a person may understand the journey itself yet still need help completing the collection process before boarding.

Payments will continue through official channels, outside the chatbot. JRTok AI's role is to explain the available steps and choices. A traveler who needs to buy or collect a ticket will still use the relevant ticketing service, machine or counter.

Where luggage guidance can help

Large suitcases create another set of questions on the Tokaido Shinkansen. JR Central has specific rules for large luggage, and the assistant can explain the size categories and when a passenger needs a seat with an oversized baggage area.

Medium bags also need a suitable place. The service can explain where to store them and cover everyday etiquette, including keeping aisles clear, setting phones to silent and using the spaces between seat rows for bags of a suitable size. Clear instructions can help passengers avoid blocking shared space in busy cars.

For visitors unfamiliar with the railway, a luggage question may arise at several points: when choosing a ticket, before boarding or after reaching a seat. Guidance available on a phone can remain accessible through those stages, while location based prompts can bring nearby information to the user's attention.

JRTok AI will also offer optional English commentary about scenery and landmarks along the line. That part of the service is aimed at making the journey more informative for first time visitors, while practical travel guidance remains its main purpose.

Five languages, with commentary in English

English, Chinese, Korean, French and Spanish will be available for the trial's travel assistance. Together, they cover a large share of visitors using the Tokyo to Osaka corridor. English serves as a common bridge language for many travelers, while Chinese and Korean serve regional visitors who often travel between Japan's major cities.

French and Spanish extend access to many visitors from Europe and Latin America, including people making longer trips. Letting travelers ask questions in a familiar language can make it easier to describe a problem precisely, especially when the question concerns a ticket category or a luggage rule.

The cultural commentary on board will be in English at launch. More languages could be added if the trial shows strong demand. Starting with five languages for practical assistance balances access with the work needed to check answer quality, since more language options can make quality control harder and slow responses.

Japan's rebound in international travel gives that language support a practical role. Many visitors arrive with limited Japanese and encounter several unfamiliar choices at once. Signs and pictograms explain the basics, but questions about tickets, seat layouts and luggage spaces still reach counters and gates.

Why quick answers matter during busy periods

A multilingual assistant available on a phone can help fill gaps in understanding at any hour. Its value may be greatest when crowds are thick and staff have limited time: a passenger with a routine question can seek a quick answer, leaving staff more time for complicated cases.

Experience with language access technology in other fields offers a useful comparison. Hospitals have used AI to identify patients who need interpreter support so that limited staff can be scheduled more efficiently. Education researchers are testing chatbots that coach providers through sensitive conversations in controlled settings, while watching for mistakes and privacy risks.

Public transport presents a simpler use case than clinical care. The shared idea is to connect people quickly with information they can understand while keeping human help available. At Shinagawa, that means supporting travelers through everyday decisions and allowing station staff to handle situations that need personal attention.

The winter trial will cover long weekends and the year end holiday period, when trains are crowded and questions increase. That timing gives JR Central an opportunity to see how the assistant performs under heavy demand and whether it helps reduce queues at counters.

What AI can explain, and where care is needed

Modern chatbots use large language models, systems trained on huge amounts of text to predict the next word. That allows them to respond to many different ways of asking a question and to explain a process in everyday language.

A visitor might ask about a suitcase, a baggage reservation or where to put a bag. A conversational system can handle those different phrasings. It can also make mistakes when a question falls outside its knowledge, which makes the information behind an answer especially important.

Grounding answers in official documents and limiting responses to verified guidance are ways to reduce false or outdated advice. For a railway assistant, that matters because a short answer may guide a passenger's next action, such as choosing a ticket or moving toward a platform.

Research on generative AI describes a balance between potential productivity gains in services such as hospitality and tourism and concerns about bias, accuracy and privacy. A transport assistant should collect as little data as possible, avoid storing personal identifiers and explain how conversations are used to improve the system.

Human supervisors can review unusual or difficult questions and update the information the chatbot uses. That work can help the assistant respond more accurately to the questions visitors actually ask, while keeping staff involved in cases that need closer judgment.

What JR Central will look for in the trial

JR Central plans to review the results before deciding what to do next. Useful measures include how many travelers scan a code, how often they return to the service and how quickly they find an answer. The languages people choose can show where demand is strongest.

Accuracy can be assessed through feedback buttons and by checking samples of conversation logs for corrections. Repeated questions can also reveal problems beyond the chatbot, such as unclear signs or confusing steps at ticket machines. Improving those points could help passengers who never use the assistant.

The busiest periods will provide a test of both response quality and demand. JR Central can examine whether the tool remains useful when many people need assistance and whether routine questions move away from counters. If the service works well, it could become a regular part of station support alongside maps, signs and announcements.

The trial's main stages are:

• December 15: the JRTok AI trial begins at JR Shinagawa Station.
• The year end holiday period and winter long weekends: crowded services provide opportunities to test the assistant under heavy demand.
• Mid March: the trial is scheduled to end, with its results intended to guide improvements and possible expansion.

The exact March closing date and a timetable for any wider rollout have not been given. Expansion will depend on what JR Central learns about use, accuracy, privacy and performance during busy periods.

How the service fits into a phone based journey

Japanese railway companies have moved more information onto phones, including seat reservations and alerts about current operations. JR Central already supports online ticketing through SmartEX, and free Wi Fi is available on many Shinkansen trains. Multilingual digital signs and station maps are also common at major hubs.

JRTok AI adds a conversational way to reach information that may otherwise sit across several services. For a first time visitor, asking a direct question can be quicker than navigating menus. Future versions could connect with current delay notices, transfer guidance across different JR groups or city transport connections, with further languages considered as visitor patterns change.

Passengers using the trial should look for QR code signs near the Shinkansen entrances, information counters and ticket areas at Shinagawa. After scanning with a phone camera, they can tap the link, choose a language and start with a familiar topic such as buying a ticket or finding a platform.

Station Wi Fi is available in many areas, and free Wi Fi on most Shinkansen services can help travelers without a mobile data plan. Access to a connection will let them continue using phone based guidance during the journey.

Travelers should keep passport numbers, credit card details and home addresses out of the chat. The assistant provides guidance rather than handling sensitive personal information. Urgent situations, including a lost child, a medical need or an item on the tracks, should be brought to a station attendant immediately.

What to Know

• JR Central will test JRTok AI at JR Shinagawa Station from December 15 to mid March.
• Station QR codes will open a dedicated mobile site, with no app download required.
• Travel assistance will be available in English, Chinese, Korean, French and Spanish.
• The assistant will cover operations, tickets, seats, platforms, boarding and luggage rules for the Tokaido Shinkansen.
• Location based prompts will provide relevant guidance, with optional English commentary during the ride.
• Ticket payments will remain with official services, and station staff will continue to provide help.
• The winter test will cover busy holiday travel, with use, accuracy and feedback helping guide future decisions.
• Possible expansion includes more stations, routes and languages, subject to the trial results.
• Travelers should avoid sharing sensitive information in chat and contact staff immediately for urgent problems.