Chatbot "Moscow Metro" Alexandra answered 2.4 million questions

2022: 2.4 million questions answered

On November 18, 2022 Speech Technology Center Group (Speech Technology Center) , she announced that chat boat the Moscow Metro Alexandra, developed by her together with the team, Moscow Metro had answered 2.4 million questions since launch. You can write to Alexandra on the metro website applications transport , in the complex, and. social networks messengers

On behalf of the Mayor of Moscow S. Sobyanin, we are developing digital services and transferring some of the services online - this is convenient for residents and saves time. Chatbot Alexandra is one of the main assistants to passengers, since the launch she has answered 2.4 million questions about transport and received more than 100 thousand thanks from users. We will continue to update and train our bot so that trips around the city are simple and understandable, said Maxim Liksutov, Deputy Mayor of Moscow in the Moscow Government for Transport.

The CST Group has implemented more than 40 large-scale implementations of voice and text robots. The monthly number of requests from users who process CST robots exceeds 20 million. Robots are developing on the basis of a platform for creating intelligent virtual assistants, which supports hybrid classification - combines the advantages of machine learning and rules. The technologies that are used in the Moscow metro chatbot allow you to recognize text messages and process 80% of them without involving an operator, as well as create new topics and develop existing ones, noted Dmitry Dyrmovsky, CEO of the CST Group of Companies.

Alexandra can already tell you how to get a certificate for work or transfer the balance from one Troika card to another, how to leave an official appeal about the operation of transport or order support from the Passenger Mobility Center.

Alexandra also answers questions about bicycles and scooters, about the metro, MCC, MCD and ground transport, about bus stations. The virtual assistant can help with answers to questions about evacuation, parking and traffic management. In addition, the topic of answers to questions about Aeroexpress and electric trains was launched in test mode.

According to the CST, the project was the first implementation of a fully cloud-based solution of this scale, implemented in the public sector.

2020

Launch of Alexander's text robot in instant messengers and social networks

December 28, 2020 CST announced that Alexandra, an intelligent text robot for passengers, metropolitan metro developed by the CST group of companies together Sberbank Moscow Bank with the team, Moscow Metro will debut in and, messengers Whatsapp Telegram as well. social networks Vkontakte

The smart assistant works on the basis of the artificial intelligence of the CST group: it can recognize user requests, perform syntactic and logical processing of the information received, quickly explores the digital landscape and builds on its basis a natural dialogue with users - passengers of the Moscow metro, providing quick relevant answers to their questions.

During focus group testing, the virtual assistant automated 75% of requests without translating to the operator. Since November 17, it has been put into commercial operation - in the Moscow Metro application, where it demonstrated 88% of automation.

The level of automation of the virtual assistant continues to grow - as of December 2020, it is already more than 88%. Since the launch into commercial operation, Alexander - such a name for the robot was chosen by passengers - answers an average of 1254 questions per day, in total more than 45,000 questions were asked to the assistant. We continue the development of the project: hybrid classification technologies - a combination of rules and machine learning, which we laid the foundation, allow us to continuously improve the virtual assistant, develop vocabulary, the number of dialogue topics, take into account the current agenda, said Dmitry Dyrmovsky, CEO of the CST Group of Companies

We are glad that our virtual assistant very quickly found his audience in the Moscow Metro application, and the number of calls from passengers of the Moscow metro is increasing every day. We are sure that Alexandra will also quickly make friends in new communication channels for her - a social network and instant messengers, and this communication will become comfortable and useful for all Muscovites and guests of the city, noted Vyacheslav Tsybulnikov, Vice President of Sberbank - Chairman of Moscow Bank

The virtual assistant runs on the ChatNavigator artificial intelligence platform from the CST group. The robot integrates with the systems of the Moscow metro, and thanks to integration with services, it promptly receives information, for example, about the work being carried out and closed stations.

Introduction of chatbot for passengers

On December 3, 2020, the CST group of companies (part of the Sberbank ecosystem) announced that, together with the Moscow Bank of Sberbank and the Moscow Metro team, it had introduced a chatbot for passengers. During focus group testing, the virtual text assistant automated 75% of requests without transferring to the operator and launched into commercial operation on November 17. In industrial operation, the virtual assistant demonstrates 88% automation. The bot is available in the Moscow Metro application.

"Our metrobot is becoming one of the most technologically advanced and modern information client services. Already in the first days of work, the virtual interlocutor showed an excellent result: only in 12% of cases did the operator need help to answer - these requests were too long or without clear wording. The chatbot implements dialog scripts for more than 17,000 questions. Alexandra - this name was chosen by our passengers - can talk about fares for travel, operating hours of stations and transfers, while not only the metro, but also the MCC, MCD and this only at launch, then we will expand this list, "said Roman Latypov, First Deputy Head of the Metro for Strategic Development and Client Work.

The CST team created a chatbot based on its own product ChatNavigator - omnicanal a dialogue platform for creating intelligent virtual assistants. The chatbot has developed a universal connector (), API which, without the help of an operator, allows you to communicate with passengers in various channels, while the basic communication scenario remains the same for all channels, but each takes into account its own communication specifics.

The most popular topics of requests to the bot during testing on the focus group were: tariff policy and methods of payment for travel, travel benefits, metro services, as well as general issues on the quality of services provided.

"The peculiarity of the project is the placement of all components of the chatbot and platform in the cloud: this is one of the fully cloud solutions first introduced on the market. Hybrid classification technologies are used to develop the bot and more accurately understand the meaning of queries: a combination of rules and machine learning. This will allow creating new topics and developing existing ones, continuously improving the virtual assistant, "said Dmitry Dyrmovsky, General Director of the CST Group.

"We are confident that the chatbot, developed with the participation of our subsidiary CST, will make communications with passengers not only in the metro, but in the future and on other types of public transport, more efficient and comfortable," said Vyacheslav Tsybulnikov, Vice President of Sberbank - Chairman of the Moscow Bank.

Chatbot Development

The group of companies CST (part of the ecosystem) Sberbank is involved in the development of an intelligent chat boat one to automate responses to passenger requests. This was announced Moscow Metro on July 31, 2020 in the CST group of companies.

The chatbot will start working in the fall of 2020 on the mosmetro.ru website mobile application , in Metro, in Moscow, messengers as well as in the MosMetro account in one of. social networks

The CST team creates a chatbot based on its own product ChatNavigator - omnicanal a dialogue platform for creating intelligent virtual assistants.

In the period from March to June 2020, the CST group implemented dialogue scenarios in the chatbot on more than 7,500 issues.

During operation, the dialogs of the bot with users will be analyzed, and the "intelligence" of the chatbot will develop. Hybrid classification technologies will be used to develop the bot and more accurately understand the meaning of queries: a combination of rules and machine learning. Additional topics will be created and existing ones will be replenished.

In the future, the chat bot will be able to give reference information on the operating time of stations, transfers, crossings, while not only the metro, but also the MCC, MCD, intercepting parking lots. The chat bot will help not only order support from employees of the TsOMP (Passenger Mobility Center), but also automate the establishment of applications and responses from the service, and find lost things. In addition, the bot will be able to support an informal dialogue.

The CST group has developed a universal connector (API) in the chatbot, which, without the help of an operator, allows you to communicate with passengers in various channels, while the basic communication scenario remains the same for all channels of communication, but each of them takes into account its own specifics of communication.

Dmitry Dyrmovsky, CEO of the CST Group of Companies:

The peculiarity of the project is the placement of all components of the chatbot and platform in the cloud: this is one of the fully cloud solutions first presented on the market. Optimization of the transport infrastructure using modern artificial intelligence and machine learning technologies is a global trend. And the chat bot from the CST group is a tool that allows you to match this trend.

Press service of the Moscow Metro:

The rhythm of life obliges us to be fast and understandable in communication with passengers. It is important for them to get a full answer to any of the questions that arise, including on the trip, in a timely manner - this is also an indicator of comfort. The integration of developments in the field of artificial intelligence and the launch of a chatbot is designed to complement the developing metro infrastructure and improve approaches to passenger service.