Voice robot Nikolai promptly helps citizens in the Novosibirsk region on hotline 122

The project Novosibirsk region "Digital Platform for Effective Communications with Patients" voice robot with as a central element became the winner of the next All-Russian competition of regional and municipal informatization projects "TRAC-IT.2020" in the nomination "."Health care

The pandemic that quickly broke into the life of the country instantly turned the service of voice communications of residents with medical institutions from fashionable high-tech fun to state need: problems with citizens' calls to medical institutions became the topic of a special meeting held in November 2020 by Deputy Prime Minister Dmitry Chernyshenko. As a result, those responsible for digital transformation in the regions are given a high priority task - to ensure the creation of regional Call Centers using voice chat bots technologies.

In the Novosibirsk region, over the past few years, a multi-channel voice self-service system has been developing systemically and dynamically. In 2020, the experience and achievements gained in previous periods during the implementation of the system in the regional Call Center of the MFC were scaled up to the health sector of the Novosibirsk region.

Since the fall of 2020, the intelligent dialogue robot Nikolai has been receiving incoming calls from citizens in the Unified Registry 124. In December, the solution was adapted for the number "hotline 122" by specialists of the Ministry of Digital Development and Communications of the Novosibirsk Region together with BSS and included in the Unified Register of Russian Software.

The Novosibirsk digital intelligent voice platform was presented at the All-Russian competition of regional and municipal informatization projects "TRAC-IT.2020" and received a high rating from the jury, presented by serious federal experts. It was separately noted that many architectural and service solutions of the platform are pioneering at the Russian level .

The goal of this project in the Novosibirsk region was to create a universal scalable digital platform for interaction with the population, which would allow using speech recognition and synthesis technologies, artificial intelligence to create a virtual operator that is practically not distinguishable in the comfort of communication and a specialized knowledge base from a human operator, "explained Anatoly Dyubanov.

At the same time, we tried to make the product as "friendly" as possible for a citizen in order to significantly simplify his interaction with the state, increase the level of trust in digital services, the minister emphasized.

Voice Service Performance

The acute problem of communications between residents and medical organizations of the region was the significantly increased waiting time on the line during the pandemic, which, in turn, could lead to the untimely provision of the necessary medical care. Today, this problem is largely helped by Nikolai. It is he who today is the first to answer all 100% of incoming calls. According to analytical data, in peak days the load on the contact center exceeds 15 thousand calls per day. In case the robot cannot deal with the request of the applicant, he transmits the call to the operator. During the entire operation of the digital platform, we did not lose a single call, people were assisted in any situation and all the questions were answered.

The system helps to solve the problem with the influx of calls during the pandemic and optimize the activities of any contact centers in the future. It is this architectural solution that will allow the region in the very near future to implement a single technological platform for all hotlines and departmental contact centers both for the authorities of the subject and for municipalities.

Key Features of a Digital Platform for Effective Healthcare Communications

The technologies of the implemented voice self-service system (hereinafter - GHS) make it possible to comprehensively realize the potential of omnichannel interaction with patients by organizing work at literally all stages of medical interaction using artificial intelligence and client analytics algorithms.

The GHS allows you to manage the queues and flows of patients, keep records, plan and use the resources of medical organizations, as well as conduct preventive examinations and medical examinations of the population. The system also provides integration services that can check the status of a citizen's appeal, prompt a work schedule, make an appointment with a doctor, etc.

The main functionality of the system:

• sound recording and maintenance of the telephone call database;
• definition of "top" topics of citizens' appeals and monitoring of trends of appeals;
• respond quickly to changes by identifying dialogues on new and non-specific topics. (For example, citizens have a need to find medicines - the platform will quickly output information about where to go about the availability of medicines in pharmacies);
• Systematic acquisition of statistics through the automatic phonogram clustering function for past and future periods;
• automatic evaluation and analytics of voice interactions with citizens in service channels.

The unified algorithm of the system allows you to optimize the process of updating information on the output consulting and information services, and also allows you to effectively work with incidents.

The regional Ministry of Health is constantly analyzing the dynamics of the use of introduced digital services. Work on the development and improvement of services (topics) is carried out based on the needs of medicine and the analysis of client tasks on the line of the contact center.

Realization of an omnikanalnost

According to the Ministry of Health of the Novosibirsk Region, during the introduction of the GHS into the work of the Unified Registration 124, there was a difficulty in implementing the fundamental requirement for the multi-channel system in the system. Moreover, the difficulty was not with the adaptation of the voice digital platform, but with the requirements for the implementation of these services in a complex and large unified medical information system of the Novosibirsk region (hereinafter - MIS). As a result, a fundamentally new architectural solution was developed, based on a single patient data model.

This model involves effective interaction with IIAs and interacting information systems at all points of possible contact with the patient by combining the functional capabilities of a single portal of public services, a medical organization registration, a contact center 124, an operational assistance service 122, a mobile application "Appointment entry," an electronic registration reg.nso.ru. For example, today the patient will be able to sign up using the public services portal, and cancel the recording - through the GHS.

Intelligent Voice Assistant Level

Before our eyes, the areas of application of artificial intelligence algorithms are constantly expanding, radically changing many aspects of labor and social communications. New technologies and opportunities are emerging. So, now they are gradually beginning to introduce NER (Named-Entity Recognition) technology into the work of robots, with the help of which speech modules are formed in such a way that communication with users becomes more and more like a human one, allowing you to convey emotional shades with the help of a "living language," said the Minister of Digital Development and Communications of the Novosibirsk Region.

The technological basis of the system is full-text recognition of fusion speech. The fused speech recognition software module operates using simulations on neural networks. As a result of automatic recognition, speech is converted into a text indexed file suitable for automatic vocabulary. Consider, for example, the request "I want to pre-sign up with a therapist for Thursday at 12:00." The system recognizes the essence of the request and uses the received information in generating a record for the current week at a certain time.

Speech analytics technologies structure the operator's conversations with the applicant, classify them into categories: complaints, feedback, repeated calls. In a special category, system developers highlighted emotionally stressful calls.

The division into categories gives both quantitative information (distribution of calls by topics, number of complaints and offers) and qualitative (growth of complaints of applicants about the quality of provision of a certain service or the work of a separate branch).

In addition, the system implements optimal means for sampling, searching and analyzing, provides the possibility of comparing different sets of phonograms and determining common, as well as unique words and expressions for these sets.

The built-in reporting module provides for the construction of various reports, graphs, charts representing statistical information on the results of phonogram theming. Visualization of data helps to identify the main patterns by indicators, processes, and employees.

If the patient turns to Nikolai on a question that goes beyond the borders of the healthcare sector, or the robot becomes difficult to recognize speech, he will transfer the call to the operator. According to preliminary forecasts, Nikolai will be able to independently call a doctor or make an entry for 50-60% of callers. No more than 40% of calls are transferred to the attendant.

In the system, there is a constant replenishment of acoustic material, which each time reduces the likelihood that the system does not recognize the speech of the caller, "says Anatoly Dyubanov.

What can I talk about with the robot Nikolai?

According to the developer of the system, the range of topics on which Nikolai provides information and advice to citizens is not limited. It depends on the customer's specialization of the contact center in which the robot is embedded.

So, if we talk about the MFC, at the moment the robot can advise on the procedure and features of providing the most popular state and municipal services (more than 40 services), for example, on obtaining a certificate for maternity capital and ways to manage capital funds, obtaining a driver's license, a passport of a citizen of the Russian Federation and others.

The robot can also advise on the mechanisms of preliminary recording at the MFC, provide information on the addresses and schedules of the branches, on the size of state duties, check the status of the appeal filed by the applicant and much more.

At the moment, work is underway to prepare for the withdrawal of about 30 more services on various topics of MFC services, the Government of the Novosibirsk Region said.

In the field of health, Nikolai knows how to communicate on target topics - an appointment, a doctor's call to the house. It will also be able to delete and confirm the record. The robot will clarify in a person the name, date of birth and specialty of a doctor. He himself will receive all the other necessary data from the medical information system. By default, a doctor's call to the house will be issued on the current date, and an appointment will be made to the clinic to which the patient is attached. Nikolai knows how to provide patients with simple advice: how to attach to a medical organization, receive preferential medicine, vaccination against covids 19, get a compulsory medical insurance policy, etc.

In addition, he will be able to make outgoing calls according to the configured call campaign, for example, to confirm an appointment, reminder that the medical examination period is approaching, etc. He also knows how to conduct telephone questionnaires of patients with the aim of subsequent monitoring of the quality of obtaining medical services.

The system optimizes the procedure for interacting with customers, and also allows you to obtain reliable analytical conclusions based on data for a better understanding of customers, including their preferences and needs, "emphasizes Anatoly Dyubanov.

The system also implements the functionality of text communications, which complements and expands the capabilities of voice self-service.

Personalization of communications with residents

The communication platform has comprehensive methods of intelligent data processing for extracting and using information that allows you to build individual interaction with each calling person. At the moment, the system does not store information about patients, all identification takes place according to the name and date of birth, which the patient announces, after which the GHS receives information from the medical information system using the API.

If the patient needs to make an appointment with a doctor, the system also clarifies the specialty of the doctor with the patient and again refers to the medical information system, and then records in the clinic to which the patient is attached, according to the data received.

In 2021, it is planned to update the system, within which the identification module will also change. The recognition of the applicant, the person with his consent, will possibly be carried out using the modules for verification and identification of the personality of the speaker by voice.

Thus, identification will be carried out with higher accuracy, and it will not be necessary to request personal data of a person in most cases, "emphasizes Anatoly Dyubanov.

Features of voice solution operation and implementation

The project team began to introduce the platform into the Unified Registration 124 system in November 2019. At the first stage, initial data were collected and a detailed examination of all business processes of the Unified Register 124 related to the provision of medical services to the population was carried out.

At the next stage, technical processes were implemented, as well as the integration of the GHS with the medical information system of the Novosibirsk region. In the process of integrating systems, it was necessary to change some of its algorithms of work. This, said Anatoly Dyubanov, turned out to be the most time-consuming and long-term stage of the project (it took almost six months). In July-August 2020, test operation of the system began.

During the development of the system, such an algorithm was thought out that allows you to quickly - literally online - adjust and update information for digital consulting and information services.

Opportunities to develop and replicate the communication system

The manufacturer of the digital consultant neuroplatform is the domestic developer BSS. Its solutions for speech recognition and synthesis, speech analytics, voice verification and identification of citizens are included in the unified register of Russian software and are registered in Rospatent. In the framework of BSS cooperation, BFT, a Russian developer of design solutions based on its own methodological and software products, carries out projects of digital consultants in different regions of the country. The solutions created on the basis of this platform are successfully functioning in regional and federal departments and services. They provide effective automatic remote service of citizens' appeals in voice and text channels, minimize the need for face-to-face visits to branches and offices, which is relevant during the pandemic, and contribute to the promotion and popularization of digital public services.

In particular, at present, such a joint project on the introduction of neurolatform is being implemented in the contact center of the Tula region. The solution for the MFC is replicated in the Kemerovo region, St. Petersburg.

In the Novosibirsk region, the possibility of creating a single call service center based on an interconnected voice system, covering the entire territory of the Novosibirsk region and integrated with external information systems, is being worked out. This integration will allow you to correctly manage call distribution, control the queues and waiting times of the caller, generate statistical information on operators, and provide unified configuration and monitoring management.

The GHS architecture provides an opportunity to expand the use and implementation of the system at any point of interaction and any industry where the task of omnical communications with citizens is relevant. In fact, the GHS has unlimited possibilities of seamless integration with other systems without significant refinement of the Nikolai robot. All changes depend only on the systems with which integration is to be carried out, "summarizes Anatoly Dyubanov.