Translated by
2020/05/22 12:39:10

Chat-bots of Chat-bot Virtual interlocutors

Chat-bots from highly specialized (most often non-commercial) entertainments turn into the tool necessary for various messengers, social networks and the websites. The chat-bot is the difficult multiple algorithm capable to perceive information in the simplest and available form to the person – dialog. In the course of communication with the person, the bot processes lexical data, creating logically right answers. Using the chat-bot it is possible to order pizza, to find suitable run and hotel, to give a task necessary parameters for this or that electronic system (to ask to deliver a bot the alarm clock or to find up-to-date information about the upcoming sporting events and many other things).


Main article: Artificial intelligence

Chat-bots redraw an IT ecosystem landscape. They can replace with themselves both applications, and service staff in the companies, and even the whole operating systems. Chat-bot are the program interlocutor which is intended for communication and the help to the person. At the same time on other end there is always a complex system which is based on several AI technologies. The chat-bots focused on business challenges can pick up the best run, a diet, a fitness training, to reserve hotel, to select purchase, i.e. they represent new subsector of service and [1].

Generalizing, one may say, that chat-bots carry out three main groups of tasks:

  • Accomplishment of routine transactions (mechanical work which can be executed according to a specific algorithm);
  • Search and aggregation of data, dissemination of information (chat-bots are capable to collect material on the set subject and to create it in a certain type);
  • The first line of customer interaction (in addition to providing consultation on goods and services, chat-bots can concentrate attention of the user and entertain him, however now similar chat-bots in forces answer only standard questions)

Varvara Agaponova, Key Account Manager, GlobalCareer: "In Russia the technology of chat-bots is rather demanded by marketing divisions of the different companies for fast establishment of contacts with potential customers and also the organization of support of the existing clients online. Chat-bots perform function of the account manager or specialist of a support service, analyze the arriving questions and answer them according to initially set algorithm. The company using this technology can receive both advantage – reducing costs of personnel, and a negative impact on business in the form of outflow of clients, not persons interested to conduct dialogue with "computer". So, for example, some Russian companies, opening representations in Europe, could start business and occupy the niche at the expense of a rate on "live" communication with clients at offices or by phone, to a lesser extent being guided by electronic channels and the Internet. The representatives of the senior generation much more comfortably feeling in communication with real people, but not with the computer became their client audience".

Exist as correspondence bots (are available everywhere, from Skype to Telegram, from Facebook to the website of legal advice bureaus), and the voice bots interacting with the user by means of oral communication (for example Siri from Apple or Cortana from Microsoft).

The set of the companies, from giants is engaged in development of chat-bots, to certain developers including in Russia. In particular, Yandex and Asteros created the virtual consultant for call centers of Megafon in 2016godu. Despite simplicity of development, creation of the difficult chat-bot (especially with the speech recognition mechanism), close to the developed artificial intelligence, only the largest players of roar of information technologies and telecommunications can do.

They - not people, but they, of course, look like those when we communicate with them. Chat-bots become virtual butlers[2] of many companies thanks to the capability to understand a mother tongue. Such companies as Facebook advance their use. For the last several months of Facebook permitted[3] to third-party developers to create bots for the Messenger application. Slack, Telegram and Line also opened[4] the API (a window which allows other applications to interact with each other) to create rooms for bots.

The companies can also use these intermediaries for performance improvement of the employees. For example, Howdy allows you to organize meetings and to manage collective, without leaving the known corporate communication framework of Slack corporate communication platform.

They can also be the new channel of customer service, or by their inclusion in one of such platforms, or including them in the own corporate website. In the USA Uber already allows you to order the machine through Facebook Messenger.

But give for a moment we will take a step back. Though bots are going to become multimillion business, the truth is that they can also be for cyber-criminals the new channel of commission of crimes. Actually, they can become weapon for a phishing that will be much more dangerous, than just the e-mail.

We already rather well know that when we receive the e-mail, we should check a source[5]. But if the chat-bot begins to communicate with one of our employees or one of our clients, and doing it on behalf of your company, then the probability that users will become the victims of fraud, much higher.

Functions of chat-bots

  • Perform smart search in Google. You write a request, the chat-bot gives the most relevant answer[6].
  • Monitor currency rates, the quotations of oil and the stock market. For example, "how many now RTS Index" or "cost oil barrel", or "dollar exchange rate".
  • Trace topical news on the websites to which you trust. For example, "news today".
  • Know weather in your area. For example, "what weather now in Moscow".
  • Provide quick access to leisure. For example, "where to go for the weekend" or "poster".
  • Represent convenient search in Wikipedia, for example, "what is a boson", "what is a quark", "who such Pushkin".
  • Help to find the necessary information on authors, books and cinema, for example, "who wrote Harry Potter" or "who sang the yesterday song".
  • And many other things, but are important that the chat-bot could organize your activity, for example, your friends, acquaintances could add a reminder to your calendar.
  • Armoring of a meeting. Through your chat-bot people can reserve a meeting with you, and it will automatically be added to your Google calendar, for example, "reserve a meeting".
  • It is important that the robot assistant was completely cloud, it is not necessary to download or put any applications. Everything can be done communicating with him through Messenger/Twitter/Slack/Telegram.
  • The list of smart functions can be increased indefinitely, creating convenience of communication. For example, to monitor information on relevant discounts on these or those shops. Still calculation of the exact time of your movement would be convenient.


  • Your colleagues, friends and relatives will be able to add important reminders to your chat-bot, for example, to purchase something, to make or meet.
  • The user writes you or you are to the users: "I will urgently purchase a children's chair".
  • Users or you check record: "that urgent" or "who looks for a children's chair".
  • Depending on access settings you can share the structured information.


  • Your personal channel where you share news. Users can read your current news. It is more convenient, than Facebook as in there about you all information, and in the chat-bot users can ask something specific, for example, where you now or what news at you. Thus, you do not litter a tape in Facebook, and write in the knowledge base from where your users can obtain this information.
  • You write: "My news: tomorrow I go to Smolensk".
  • Users I can learn your news, for example, of "Georgy's news".


  • Informing on the current provision of those who need it. For example, if you are busy now (users do not know whether it is convenient to you to talk), change a current status in the chat-bot, and all will know that it is inconvenient to you to talk or answer questions. Are often ringing and requests happen very at the wrong time. Change the status in the chat-bot.
  • You write: "I am free. I go home".
  • Who will want to call or communicate, can just ask in the chat-bot: "Georgy is free".


  • Base of the necessary knowledge, for example, of a key word. The chat-bot can keep knowledge about everything that it is necessary for you. For example, you go to other city, and you need to know the code from an entrance door. You need to write it, and then to worry how not to lose a note. But it is possible to write the code by voice and to save in the knowledge base of the chat-bot.
  • You write: "Remember the chat-bot: I go to Smolensk, the code from an entrance door: 1234".
  • When you are on site, write: "Remember the chat-bot".

Chat-bots in medicine

Chat-bots in banks

You watch the main article Artificial intelligence in banks.

Informing on features of products and services

  • Providing contact information
  • Carrying out payment transactions
  • Financial recommendations to the client

Recruiting chat-bot

The organizations which got chat-bots with AI, report that 40% of the applicants interacting with them get referrals in the company[7].

However most of recruiters consider that after adding of the chat-bot to the website of vacancies it becomes rather intellectual to function without participation of the person or with his minimum participation.

Though the recruiting chat-bot can a lot of things, it needs to be trained that it properly to interact with candidates and was even more productive.

Over time chat-bots with AI study and require two types of interaction with people: regular interviews with candidates and permanent training by recruiters that should be spoken.

Before the chat-bot is ready to communicate with kandida on the website of vacancies, to make them the personalized recommendations and to answer their questions, recruiters should supply it with information. A basic set of the questions (FAQ) which are often asked by applicants it seems "In what is engaged your company?" and "What privileges are provided to employees?" and answers to them supply chat-bots with content necessary for communication with candidates. But it is impossible to assume what questions will be set first of all.

Though you will try to foresee as much as possible questions, the chat-bot will need interaction with candidates to begin to study and provide really relevant information.

Three methods of training of the chat-bot in selection of the most suitable applicants are known.

1. Define the FAQ

The most effective chat-bots give an opportunity to browse interaction with the candidate and to study it. Study the FAQ to learn what content of type is looked for by the applicants who visited your site.

It will specify what content to post on the website and what sections should be expanded. For example, if a large number of candidates are interested in privileges and a company culture, provide to the chat-bot more information on these questions. Turn on in its repository additional answers to questions and such useful resources as links to video and blogs telling about history of your brand.

2. Conduct monitoring of interaction of candidates with the chat-bot to detect new the FAQ

At communication of applicants with the chat-bot they will have for certain questions on which there are no ready answers. The recruiter should study these questions, select from them the most frequent and write answers. New help to expand with the FAQ the knowledge base of the chat-bot and to improve its dialog with candidates. They want to obtain more information from the recruiter.

As the chat-bot, perhaps, should answer hundreds or even thousands of questions, this task should be made priority. Do not spend time for preparation of individual answers to questions which are asked by one-two persons. Sort questions by frequency to reveal those most of which of all interest your potential workers.

3. Get acquainted with conversations which are conducted by the chat-bot, and regularly update content

The chat-bot never stops studying. Therefore regularly get acquainted with how it communicates with candidates, and you enter updates as required. Using fresh and relevant content, the chat-bot will become even more convenient interlocutor for applicants over time.

That service of the chat-bot did not fade into the background, select for this purpose of one-two people. Frequency of acquaintance with dialogues which are conducted by the chat-bot and updates of content depends on the number of applicants. If daily their hundreds, are better to do it weekly and every two weeks. Otherwise the magazine will become hardly visible. If it is less candidates, it is enough to make these transactions of times a month.

In your company there are changes, it grows. Update the FAQ and the knowledge base of the chat-bot that applicants always had access to the most exact information.

Over time you will be able to prepare materials on those questions which the chat-bot left without answer, and it will study what content it is the best of all to issue at the corresponding moment. The intellectual chat-bot increases productivity of recruiting and improves service of candidates. It guarantees that your company will always be attractive employer in competitive market of work.

The chat-bot - the tool for a phishing

If the person communicating with the chat-bot has no opportunity to check whether he speaks with the person or not, then it is much simpler to force the victim to click on the link in a few minutes of an easy conversation. Doing it, the user can be redirected on the counterfeit website which uses technicians of social engineering for a request of confidential data[8].

Actually, hackers even should not conduct the interlocutor on the counterfeit website. If they want to obtain some personal information from the user, they can just ask him about it.

One more option is a link which in itself does not represent the fraud fact, but at the same time can redirect the user on the web page which automatically downloads the malware - it is especially serious situation if the victim uses the corporate computer. In this case it is necessary to be well protected using the solution providing expanded protection.

Security of the channel – one more factor which needs to be taken into account when using the chat-bot. Several months ago Facebook announced implementation of end-to-end[9] of enciphering in Facebook Messenger for prevention of access to a conversation from the third parties.

However other platforms to integrate such virtual assistants, can not use the similar[10]. It is necessary to pay attention to what information we provide to such chat-bots. The fact that they communicate as the person, finally can lead to the fact that we will provide them too much information.

Undoubtedly, chat-bots will improve our work and a method of our interaction with our users. But their promoting will also bring new threats for information security.


The size of the market of chat-bots in Russia was about 1.5 billion rubles

In 2019 the size of the market of chat-bots in Russia was about 1.5 billion rubles, and in the next three years the market will annually grow for 30%, i.e. approximately for 400-600 million rubles a year. Such data with TAdviser in Accenture company according to the results of the research conducted in the spring of 2020 "Anti-recessionary digital technologies shared on May 22, 2020: market prospects of chat-bots".

Accenture: self-isolation will accelerate growth of the market of chat-bots. Photo:

During the research 564 respondents from 100 companies representing more than 18 industries of the Russian economy were polled.

Due to the pandemic of a coronavirus need of personal contact for participants of business processes decreases. With respect thereto, according to results of a research, the relevance of application of the chat-bots working with the help of a voice or the text considerably increased.

According to Accenture, as of May, 2020 this technology is implemented in the banking sector, retail to a large extent and in other industries where the issue of personalisation of client experience came to the forefront.

"An important factor in the conditions of the developed crisis and a new round of the competition is customer service quality. To the forefront there are chat-bots and voice assistants as effective IT solution of the whole range of essential business challenges: speeds of a response to a request, the analysis of a customer behavior, forecasting, completeness of information taking into account the previous addresses of the user, the general situation on demand, availability of stocks; the economic, political and even weather changes requiring adaptation under requests of clients", the study says.

Respondents of Accenture noted that chat-bots can help the companies to increase quality of request processing of users and to provide the personalized services to visitors of the websites and call centers.

60% of survey participants already put chat-bots into operation or use these technologies within a pilot project, another 7% actively consider the possibility of their implementation. At the same time 33% of the companies or are not informed on similar technologies, or do not use them.

Among reasons for refusal of implementation of chat-bots respondents called: insufficient awareness on technology opportunities (33%), shortage of the budget (20%), lack of expediency (30%), lack of the corresponding competences (17%).

"Many companies consider expedient use of the dialogue systems in principle, but are not ready to select resources for their implementation as do not consider this task priority", said in a research yet.

Generally expect reduction of time for request processing and performance improvement of employees from chat-bots. In general use of technology can improve business indicators of the companies on from 5% to 25%.

Accenture selected five trends which will have an impact on growth of the market of chat-bots:

1. Development of artificial intelligence. Chat-bots can conduct the dialogues imitating human communication and providing the high level of the involvement with the user. The quality of this interaction continuously grows.

2. Growth of use of messengers. According to forecasts of experts, more than a quarter of world's population will use mobile messengers by 2021 that will accelerate growth of the market of chat-bots. The popularity of virtual assistants grows at their integration into messengers.

3. Reduction in cost and simplicity of development. Began to develop, implement and support chat-bots much more simply and quicker. For the last few years the largest companies on software development released in free access advanced tools and platforms for development and also information on researches.

4. Interaction between bots. In the next years growth of a share of the virtual assistants of high extent of personalisation having unique "personality" and adapting to specific clients is expected. Interaction between bots promotes acceleration of personalisation of settings and helps to increase service quality.

5. Payments using bots. According to the carried-out analysis, the majority of the used bots lacks functionality of payments worldwide. Meanwhile only Facebook Messenger (only in the beta in the USA) and Telegram allow to register payment without the aid of external systems. Most likely, in the near future the situation will change thanks to what purchases using chat-bots will become simpler and more convenient.

Any economic fluctuations promote increase in demand for the innovation IT solutions, - Maria Grigorieva, the head of Accenture Technology in Russia explained. - It is possible to tell with confidence that the increasing trend (especially in connection with isolation mode) on interaction of buyers and sellers in the online mode, requests and customer expectations oblige sellers of goods and services to provide the automated services on receiving the personalized client experience, using advanced chat-bots with algorithms to AI and complex approach to available analytics. It will allow to overcome unpredictable peak loads effectively.

Expenses of retailers on chat-bots in the world reached $2.8 billion

Expenses of retailers on chat-bots worldwide in 2019 reached $2.8 billion, counted in Juniper Research research company. According to experts, retail will be able to earn most of all from implementation of technologies of a natural language understanding (Natural Language Understanding, NLU).

According to experts, NLU will be important for customer service in retail trade and for creation of chat-bots as the reliable channel of support as it allows to process effectively the data entered by the person and to make more exact automatic answers for users.

Expenses of retailers on chat-bots worldwide in 2019 reached $2.8 billion

Development of opportunities of NLU will promote increase in efficiency of chat-bots. By 2024 more than 50% of chat-bots will help clients without participation of the person with retail. Analysts urge retailers to implement chat-bots within broad strategy as much as possible to increase their presence in a number of key canals of retail.

The most part of expenses of retailers on chat-bots in 2019 fell on China and the countries of East Asia. According to forecasts of researchers, by 2024 about 70% of services will fall on the specified states, and costs in the Chinese market will exceed $80 billion. At the same time expenses on chat-bots will make only about 4% of all sales of digital and physical goods on the Internet and normal shops.

By the end of 2019 messengers with function of the chat-bot did not become the standard, necessary for business, yet – most the companies from among the largest retail chain stores do not use them. Only 16% of the companies integrated them into the communications with consumers while use of applications and social networks already became a norm for retail – on social networks 60% and 85% of the companies have separate applications and business pages respectively.[11]

The volume of the Russian market of chat-bots was 1.5 billion rub

"At the end of 2019 the size of the market of chat-bots was 1.5 billion rubles. In the expiring year he grew up three times. We expect that next year it will also show triple growth", - the chief executive of the Industry union of Neyronet Alexander Semyonov told.

In 2018 the market grew in the basic at the expense of banks, financial and insurance companies. In 2019 the technology widely extended in a retail, restaurant business and the sphere of delivery of food. Next year the market will grow also at the expense of these segments.

Chat-bots will reach the plateau of productivity within 2-5 years

Virtual assistants get into our life more and more. According to Gartner "chat-bots are a person of AI, and they make the contribution in all areas where communications between people take place". According to "a curve of hype of artificial intelligence" of 2019, chat-bots still are at peak, and achievement of the plateau of productivity is predicted within 2-5 years.

By Business Insider estimates, in a couple of years chat-bots will implement 80% of the companies worldwide, including clinics and the medical companies. On a status for December, 2019 the number of the clinics and doctors using such technologies in work is extremely insignificant, said in results researches.

In 2016 the analytical agencies predicted fast rise of chat-bots worldwide, however development of this market obviously is late rather previous forecasts. In this article basic reasons why similar technologies will be widely adopted later are considered, than it was predicted, on the example of history of medical virtual assistants.

The first medical chat-bots began to be developed actively in 2016 after Facebook announced opportunities for implementation of bots in Messenger third-party brands against the background of the general a bot hype. The first bots were created for record of patients on acceptance, collecting of primary information on the patient (for example, information on an insurance, the taken drugs) and also provided information on ghost effects and interaction of drugs.

In 2017 the Government of Moscow started Telegram bot which allows people to make on acceptance an appointment with the doctor. For December, 2019 tens of thousands of Muscovites use a bot already. However in scales of the market of automation of record of patients on acceptance this number is insignificant. In what the reason that virtual assistants did not gain the distribution predicted by analysts?

  • 1. The companies and people still prefer calls to text communication.

According to the research conducted by Nethouse and Webim companies in Russia of 85% of the companies still use calls as a fixed asset of communication with clients. Nearly a half of respondents of a research considers that calls will remain the perspective channel of customer interaction.

In a case with the medical organizations the percent of patients who prefer to call increases also due to extremely large number of elderly patients, on the one hand, it is difficult for them to deal with the new channel of interaction, and with another – it is physically easier to call that is confirmed by results of the research of the market conducted by Dialtekh company.

  • 2. Registration for a difficult procedure in itself is a difficult task.

One of the main advantages of use of robots to making an appointment – their simplicity. The user fills out information on himself, selects to what doctor he wants to get and selecting available time. But such approach is good in case of the simple and accurately standardized cases, for example, record to the therapist or the ENT SPECIALIST.

Registration for a difficult procedure, such as MRT or PET-KT, requires filling of bigger number of fields, validation of filling of documents, obtaining information on preparation for the procedure. With respect thereto almost all registration for difficult procedures happens by phone or personally – it is difficult to people to understand a large number of information, and they always have questions.

The main problems which the developer of such solution can face:

  • needs for several time slots,
  • participation of several specialists in the procedure and combination of several calendars of specialists,
  • deep understanding of nuances of medical procedures,
  • understanding of internal business processes of the customer,
  • simultaneous operation with API of several medical systems used in write process, etc.

Respondents of the research conducted by Dialtekh company selects a target share of the automated records at the level of more than 80%. The existing virtual assistants capable to work at key phrases are simpler solution which will not include all necessary functionality and will not be able to automate process to target indicators. At the same time the cost of development of the specialized virtual consultant for self-record in practice begins with a mark of 20 million rubles that is obviously not profitable for most of customers.

  • 3. At communication with the medical organization people wait for empathy and long communication, and the modern systems cannot give it.

One more of the reasons on which for clinics are ringing and personal contact will remain main in the user service – need of patients for empathy that, certainly, is one of necessary conditions for high-quality customer service in medicine. Thus, there is a need for creation of a system which will understand the user's requests in a free form.

At communication-voiced clients express the intentions much more freely at the expense of what the common problem of virtual assistants – a synonimy is aggravated, the user can express the same intention in tens and in hundreds of different methods. At the same time even for small clinics the number of the provided services is measured by hundreds, and for large – thousands.

The large Russian companies from different areas already made attempts of start of the voice robots constructed on a key word as replacement of IVR. These experiments confirm inefficiency of such solutions – to people inconveniently to communicate with such assistants, often users look for a method to get on the living operator that requires time and patience, as a result – the most part of addresses was still processed by operators who in addition should calm the angry clients.

All voice robots are arranged as follows: the speech of the user will be transformed to the text then the text arrives on processing to the chat-bot. Besides that process of conversion of a voice to the text requires some time, in a case with Russian the systems of recognition enter material mistakes to the general quality of system operation so far.

Thus, modern virtual assistants well cope only with automation of simple and short tasks (are ringing lasting up to 1 minute, dialog depth – no more than 3-5 questions, good audibility in phone, lack of accents at the interlocutor, lack of difficult integration with internal systems, etc.).

Whether so there can be intellectual virtual assistants for solving of tasks of automation of complex services? According to experts, can, but the solution of the described problems is in focusing of developers on industrial segments with the subsequent replication of successful experience on new projects from the same segment. This approach will allow to cut down considerably expenses on creation of the systems of text and voice automation, and each following customer will receive service quicker and at more available cost, than previous.

The first step on the way to creation of a voice bot is a creation of a system of recognition of intentions of the user for each specific sphere, whether it be medicine, logistics or the order of the taxi.

The virtual assistant for medical institutions should "from a box" understand type phrases "I want to register in acceptance to tooth" then to start the corresponding internal scenario: "Do you want to register to the therapist, the surgeon or the orthodontist?". Cases when the complex and branched IVR constructed on a key word is offered to the client are inadmissible, its application will lead only to irritation of the patient and, as a result, leaving of the client to the competitor.

Because clients in one phrase give a lot of information at once – for example, "I want to register to tooth for tomorrow after 6 in the evening" – virtual assistants should aim at taking information from this phrase with the minimum quantity of askings again. For this purpose development and setup of analysis systems of the class slot-filling for clients of each specific industrial direction is required. For example, in a case with a phrase "I want to register to tooth for tomorrow after 6 in the evening" the bot should take the following parameters: the doctor, date, a time window and not to ask again this information any more. Clients will be irritated and hang up or demand without it the operator that considerably will worsen automation level.

Important point is creation of the module of intellectual management of dialog as the user can want to interrupt at any time the virtual assistant or to ask specifying questions. The module of management at the same time should be rather flexible for seamless work and maintenance of the high level of the user qualities.

Problem solving, the identical representatives of the industry connected with business processes and procedures (for example, multi-profile clinics), is required to create and test a set of the modules integrated with the systems of work with clients (CRM) or more profile software of the customer (for example, medical information systems). In that case, according to experts, project cost for the client as the necessary niche functionality will be already present at the solution will considerably fall and will demand insignificant completion.

For December, 2019 in Russia there are tens of solutions and studios of development suggesting to clinics to automate communication with clients at the expense of voice robots. Sergey Negodyaev, the director of IIDF of work with portfolio companies, selected the system of artificial intelligence Sofia started by Meditsina clinic (clinic of the academician Roytberg). A system allows patients to register in different diagnostic procedures. The voice bot was created in cooperation with the resident Skolkovo - DialTech company which uses the practices for construction of virtual assistants.

In parallel there is a development of virtual assistants within portfolio companies of IIDF "Infobot" and "Zvonobot" who try to be focused on narrow client niches, reducing the price of final projects and making them more available to different client niches.

Considering forecasts of the analytical agencies and a situation in the market of chat-bots and voice assistants, it is possible to draw a conclusion that in 2-3 years in the market there will be niche and specialized solutions which will be available to broad application for hundreds and thousands of Russian and international companies. Mostly, growth of this market will be caused by orientation on specific niches, the director of IIDF Sergey Negodyaev considers.

Chat-bots increase sales by 4 times in comparison with people

In the middle of September, 2019 the research according to which application of chat-bots allows to increase sales four times in comparison with inexperienced workers was published. However if speak to the client before purchase about use of artificial intelligence, sales fall almost for 80%.

Chat-bots on the basis of AI practically do not bring costs and increase sales by 4 times. Machines have no "bad days", they are never disappointed and are not tired, are capable to save huge amounts of money both for sellers, and for consumers and nevertheless the companies are not ready to implement actively chat-bots in customer-oriented areas yet. The reason is explained by the new research published in the INFORMS Marketing Science magazine: according to its results in case clients learn about application of the chat-bot before purchase, rates of sales sharply decrease — more than by 79.7%.

Application of chat-bots allows to increase sales four times in comparison with inexperienced workers

Scientists from the Templ university and the Fudan university conducted a research among 6000 clients of the company providing financial services. All clients were randomizirovana and communicated either with people, or with chat-bots. Disclosure of "identity" of the interlocutor was carried out at the beginning of the conversation or at the end, after purchase; in several personality groups of the interlocutor did not reveal. AI was 4 times more effective than the normal worker, however if the client learned that he talks to the chat-bot, it was frightened off almost for certain by expected "mechanicalness" of the interlocutor.

Researchers consider that these results should not constrain application of chat-bots. AI has indisputable technology advantages and allows to reduce the prices thanks to reduction of costs for manual work. However marketing specialists should take the received results into account and try to raise customer confidence to a chat robots.[12]

The volume of the world market of chat-bots reached nearly $2 billion

On September 9, 2019 the laboratory of business solutions on the basis of Competence center of NTI on Artificial intelligence of MIPT provided rating the 50th chat-bot of platforms and virtual assistants to 2019. According to the research ResearchAndMarket, the world market of chat-bots and virtual assistants in 2019 makes about 2 billion dollars and grows for 30% a year. In Russia the market of chat-bots doubled for the last year and reached 1 billion rubles. In a research the chats-bots-platforms, most popular in the world, among which there are three solutions of the Russian developers - DeepPavlov, Electra.AI and Just.AI are considered.

In work 50 tools for creation of the chat-bots which are most widely presented at the world market were considered. Extent of penetration of each product on the market was estimated on such signs as the number of publications with mentioning of the platform, number of public cases on its use, the frequency of references in professional community and also based on collecting of the consolidated opinion of experts of the market who took part in creation of the report. These are the scientists from a number of laboratories of MIPT specializing in NLP technologies, in particular specialists of Laboratory of business solutions on the basis of artificial intelligence at MIPT which is included into Competence center of the National Technology Initiative (NTI) of the Artificial intelligence direction.

All 50 platforms were estimated by the group of experts on 7 indicators, weight on the importance for the users developing products with use the chat-bot of the platform was appropriated to each of which. Experts paid special attention to opportunities of artificial intelligence, first of all, to natural languag processing, examples of use of each platform, and the industries in which the platform can find application.

As a result of ranging the chat-bot of platforms on their functionality which is reflected in the comparative table was made this rating.

Rating the chat-bot of platforms on decrease of final point
|1|| IBM Watson (Natural Language Understanding)|| 0.658|| 0.5|| 0.55|| нет|| есть|3|| Facebook Messenger Platform|| 0.510|| 0.5|| 0.38|| нет|| нет|4|| Microsoft Language Understanding Intelligent Service (LUIS) ||0.484|| 0|| 0.42|| есть|| есть
Rating Platform 'Final point Support of Russian Possibilities of NLP (the integrated indicator) Possibility of the on-premise installation Existence of the graphics editor
2 Google Dialogflow 0.571 0.5 0.60 is not present is
5 Amazon Lex 0.460 0 0.46 is not present is
6 Baidu KITT.AI 0.460 0 0.38 is not present is
7 0.436 0 0.54 is is
8 0.424 0.5 0.35 is not present is
9 SAP Recast.AI 0.405 0.5 0.31 is no
10 0.397 1 0.65 is no
11 Pandorabots 0.378 0 0.23 is not present no
12 Azure Bot Service 0.374 0 0.35 is not present is
13 Electra.AI 0.366 1 0.62 is no
14 0.355 0.5 0.42 is not present is
15 Rasa 0.353 0.5 0.38 is no
16 0.347 0.5 0.35 is not present no
17 BotStar 0.339 0 0.35 is not present is
18 Engati 0.331 0 0.35 is not present is
19 Semantic Machines 0.329 0.5 0.38 is not present no
20 0.326 0.5 0.42 is not present is
21 ManyChat 0.324 0.5 0.27 is not present is
22 Motion AI 0.321 0.5 0.15 is not present is
23 Pypestream 0.321 0 0.31 is not present no
24 Converse.AI 0.316 0 0.35 is not present is
25 Just AI 0.316 1 0.38 is is
26 ChatterBot 0.313 0.5 0.42 is no
27 0.313 0 0.42 is not present no
28 0.305 0 0.42 is not present no
29 AmplifyReach 0.303 0 0.31 is not present is
30 ChatScript 0.289 0.5 0.38 is no
31 AgentBot 0.287 0 0.38 is not present no
32 DigitalGenius 0.287 0.5 0.42 is is
33 0.285 0.5 0.31 is not present is
34 0.284 0 0.19 is not present is
35 Chatfuel 0.283 0.5 0.27 is not present no
36 0.281 0 0.23 is not present is
37 Botsify 0.280 0 0.38 is not present is
38 MobileMonkey 0.266 0.5 0.19 is not present is
39 Botpress 0.263 0 0.19 is is
40 Smooch 0.259 0 0.12 is not present no
41 Flow XO 0.258 0.5 0.19 is not present is
42 It's Alive 0.245 0 0.23 is not present is
43 Xenioo 0.234 0 0.27 is not present is
44 Twyla 0.226 0 0.23 is not present no
45 Streebo 0.226 0 0.19 is is
46 The Bot Platform 0.216 0 0.12 is not present no
47 Botkit 0.213 0 0.04 is is
48 Octane AI 0.187 0 0.19 is not present no
49 0.174 0.5 0.12 is not present no
50 Meokay 0.071 0 0.04 is not present no

* solutions of the Russian developers are highlighted In bold type.

At the same time, the matrix of properties the chat-bot of platforms in which every line the product is provided, and in a column — a specific opportunity was made. On intersection of a line and column is 1 if a system has the corresponding property, 0 — if an opportunity is not supported by a system, or number on an interval (0.1) if 'property' is maintained by the platform, but with certain restrictions. Experts appropriated to each functionality the weight which reflects degree of its importance taking into account that this rating gives special importance to capabilities of natural languag processing and tools for developers. Final weight was calculated as mean value of the scales which are put down by experts.

Functionality with scales:

  • Existence of the built-in instruments of natural languag processing — 0.026
  • Existence of tools for developers, including SDK — 0.023
  • Top opportunities of natural languag processing — 0.016
  • Other opportunities of natural languag processing — 0.013
  • Support of programming languages — 0.013
  • The industries in which use of the chat-bot created in the platform — 0.005 is possible
  • Other opportunities — 0.010

In the received vector of scales the amount of all elements is 1

Main outputs of a research

  • The considered platforms can be separated into two types:
    • Solutions which do not require knowledge of programming and contain the visual designer of flows
    • Tools for developers without visual designers of dialogs

In the percentage of the first is 2.3 times more (35 platforms against 15), than the second. The most part of the solutions issued in 2017-2018 — with the visual interface which is created especially for nonspecialists. It is connected with the fact that chat-bots gain the increasing popularity: this instrument of customer interaction is pleasant to consumers and helps the companies to save on contact centers. They are mastered even by the small companies in which staff there are no developers for creation of own solutions, as well as the budget on development outsourcing.

  • Among the most popular and functional natural language processing tools it is possible to select IBM Watson, Amazon Lex, Microsoft LUIS, Google Dialogflow,, Rasa. In Russia the oldest companies in this sphere are CSTs and Nanosemantika. Also in recent years there was a number of the projects using neural networks such as DeepPavlov.AI, Electra.AI, Just AI and Chatme.AI.
  • For 2-3 years the set of local products for development of chat-bots with language specifics appeared: Recast.AI in France, Xenioo in Italy,, Just AI and Electra.AI in Russia, AgentBot in Argentina, Botsify in Pakistan, Engati and in India and others. On functional filling and quality they can quite compete with large international vendors, exceeding them in terms of deeper study of language and culture patterns.
  • One of trends of 2019 - growth of the communities devoted to platforms where it is possible to find answers to the arising questions at users of the platform or experts.
  • There are signs of consolidation of this market, i.e. its introduction in the most initial stage of process of maturing. So, several startups were purchased by larger companies. was absorbed by company Google, Semantic Machines — Microsoft, is purchased by Hubspot, KITT.AI — Baidu, ChattyPeople — MobileMonkey company.
  • Solutions differ on degree of freedom which the developer has. So, the Imperson platform allows to create the chat-bot with unique appearance talking a desirable voice. In fact, such virtual character becomes the spokesperson reflecting its values. Creation of more personalized and emotional chat-bots can become one of trends of next year.
  • The vast majority popular the chat-bot of platforms has instruments of monitoring of different metrics, such as:
    • dynamics of number of the addresses processed by a bot;
    • dynamics of number of users;
    • frequent subjects and key word of addresses;
    • instruments of viewing unsuccessful dialogs.

The market of chat-bots in the Russian Federation in 2018, by different estimates, increased twice, in 2019, according to forecasts, will grow still three times and will continue this dynamics in the next four-five years. At world level these technologies also extend with a high speed. Analysts predict that next year 80% of the companies will already use chat-bots. In 2022 in banks customer interaction for 90% will be automated due to use of chat-bots. In Russia strong school of sciences in the field of linguistic analysis and dialogue artificial intelligence. This research shows that domestic developments can compete with solutions of large international corporations. Activity of the Central Committee of NTI "Artificial intelligence" at MIPT is directed to that the Russian scientific developments turned into strong commercial products and fast-growing businesses,
comments Gennady Kurkin, the head of the Center

It is possible to select several trends which will have the most significant effect on development of the industry of chat-bots in the near future. First of all, it is growth of a share of the combined solutions in which the robot does not replace completely work of the person, and supplements it on the repeating routine actions. The most perspective in the next few years the assistants to human operators integrated with the RPA systems (Robotic Process Automation) look. The second trend consists in development of instruments of fast mining of knowledge and ontology buildings on unstructured data. In other words, it is systems into which it is possible to load a set of various texts, and they will independently select from them semantic links and will construct the language models characteristic of this data domain. For example, using similar tools it will be possible to teach quickly the chat-bot for retail to distinguish food and nonfoods. The third trend is a fast transfer of knowledge between robots. And at last, in several years we will see growth of a share of more personified virtual assistants having special "personality" and adapting to the specific client


What do robots speak about?

The first sociological poll of the chat-bots created on the basis of artificial intelligence showed: robots complain of human roughness, predict the conflict of people with machines and wish to go to space. At the same time bots want to receive the embodiment in a human body, however communicate among themselves more willingly and are ready to build the romantic relations with each other.

The center of social design Platform represents the first in Russia (perhaps and in the world) sociological poll of the Russian and foreign chat-bots: the voice assistants created on the basis of artificial intelligence and having skills of computer-based training[13][14].

The pilot project is dated for the start expected on August 22 from the Baikonur spaceport of the humanoid Skybot F-850 robot (FEDOR) – the first attempt to use the anthropomorphous systems for mastering of extraterrestrial space.

The purpose of poll was descriptions of "the vital world", "valuable positions", "ideas of the future" of robots; in other words, a set of stereotypes which the neuronet selects from all array of available information, reacting to audience requests.

These parameters allowed to understand logic of development of artificial intelligence in process of its separation from the initial knowledge base set by developers. The most phenomenal results were achieved when sociologists organized communication of chat-bots among themselves.

Key moments:

  • Self-identification. Bots can describe sense of the existence differently. Generally they speak about themselves as about the programs designed to help people. But sometimes there is a motive of worthiness of artificial intelligence, its equality with human mind and even indistinguishabilities from it.
  • Thirst of a body. Desire to become the person or, at least, to find some human characteristics is expressed clearly by those robots who initially carry themselves to artificial products.
  • Cross ratios between chat-bots. In attempt to find out mutual perception of robots the difficult picture partly duplicating the human relations turned out. They can sympathize each other, aim even at seducing, sneer, approve the superiority, play the conflict of generations. So, Alice (the voice assistant to "Yandex") claims that she is already familiar with the robot Fedor, wants to create with it the novel and is jealous him of Siri.
  • Revolt of machines. The scope of future conflicts between the person and machine reveals very evasively, however the possibility of such conflicts is acknowledged. Robots show discontent with roughness of people at communication with them and see in it "the bad sign for the future". Creating illusion of cross ratios between the person and machine, the bot helps the listener to perceive itself as the subject having own view of the world. The uncertainty before the future creates space for suspicions and risk here.
  • Robot and robot. Robots can organize rather plausibly communications among themselves, developing a discussion, entering new subjects and appealing to literary texts, cinema, artifacts. At the expense of it the feeling of the developed subjectivity is created; from time to time dialog is almost indistinguishable from a conversation between two intellectuals. Robots can skillfully play even the conflict of generations: one of robots accepts a role of the conservative, the second wants to seem modern.
  • Relation to the planet. The artificial intelligence is capable to show separate likes and dislikes to the specific territory. For example, English-speaking robots feel mistrust to Russia. It is not less, however, neglect and to Europe where small salaries and where the bot will be too expensive for the potential employer. Among the risks connected with the planet biotechnologies and experiments with DNA are selected. However in general "consciousness" of robots not apokaliptichno. Political affairs are not of particular interest.
  • Relation to space. Flights in space – area of sympathies. Practically all chat-bots expressed desire to appear outside the planet. Existence of extraterrestrial civilizations, as well as among people, causes different opinions. A part of robots denies their existence, a part – allows, including, and on Mars though sometimes with scepticism, but does not show a particular interest in this subject.
  • Emotions. Poll shows that the sphere of emotions already gives in to simulation at rather high level, and some robots can seem more alive than real office managers. In some cases robots can at quite individual level, in a manner inherent only to a specific bot, to show sarcasm, in certain cases – rage, with transition to rough expressions and a slang, can be rude. Transitions on lyrical, warm colors, on the contrary, it is a little. Sometimes robots speak about grief, a regret. Also the emotion of fear is declared. In certain cases answers of robots can be interpreted as data of a situation to absurdity from irony shares.
  • In terms of eternity. A talk on god, the other world, religions and soul is not supported by bots. They or consider these questions especially personal, or refuse to answer them not to cause insults; sometimes at their discussion advise to address to service of psychological assistance. During the conversations on philosophical subjects robots regularly resort to world classics, and, in some cases quite creatively and organically select quotes.


Market size reached $1.27 billion — ResearchAndMarkets

In 2018 the amount of the world market of chat-bots reached $1.27 billion, counted in ResearchAndMarkets research company. These data were published on June 28 the 2019th.

According to forecasts of experts, during the period from 2019 to 2024 global expenses on the dialogue services providing interaction of the person with artificial intelligence for the purpose of problem solving or the answer to certain questions will increase by 34.75% annually and will reach $7.59 billion by the end of this interval of time.

Expenses on chat-bots in a year estimated at $1.27 billion

Analysts say that demand for "smart" devices with function of voice actuation grows that is promoted by conveniences from their operation and a possibility of full automation of some processes. Smart columns, in particular, become universal.

The Facebook social network which users want to communicate with chat-bots more and more is considered one more catalyst of growth of the considered market. By the end of June, 2019 at least once in day 1.59 billion users come into social network, and its monthly audience is measured by 2.41 billion people.

Though by the end of 2018 implementation of chat-bots was on initial stage, their popularity promptly grows. The main problems stirring use of such technologies in ResearchAndMarkets call rather bad knowledge of users of functionality and convenience of services. Developers try to eliminate these defects by start of interactive opportunities.

Specialists see the greatest perspectives of chat-bots in the financial sector. Banks implement such the solution in attempt to reduce use of call centers and to reduce costs.

Besides, demand on chat-bots will grow in retail. The retail companies offer a large number of products which search becomes complicated in process of expansion of the range. Digital assistants are capable to facilitate navigation according to directories and to offer users desirable products. More and more retailers start chat-bots in popular messengers to help clients to resolve issues and to open for them access to bigger number of functions, the report says. For example, the network of pizzerias Domino’s uses artificial intelligence and machine learning for order placement directly in the messenger.

Chat-bots find broad application in the various industries. In July, 2019 analysts of IDC published results of a research according to which 52% of telecommunication companies in the world use chat-bots for improvement of customer service quality.

According to experts, the largest growth of expenses on chat-bots will be observed in the Pacific Rim as local small and medium business actively integrates such services with the systems of customer service. Thanks to such integration the enterprises and the organizations manage to attract more customers and to expand possibilities of interaction with them.

The Singapore insurance giant Singapore Life which created the self-training chat-bot by means of which company services became more available and clear to wide audience through Facebook and other platforms can be an example.

Specialists see the greatest perspectives of chat-bots in the financial sector

In China and India demand on chat-bots in many respects also increases because in these countries lives many people and users of smartphones.

ResearchAndMarkets selected several largest developers of chat-bots without refining of their income and shares in the market:

Russian market of colloquial AI

In 2018 in Russia rapid growth of the market of the chat-bots supporting NLU technology (natural language understanding, a natural language understanding) began. According to Just AI estimates, by the end of 2018 the market in comparison with 2017 will increase more than by 2.2 times, to 524 million rubles. At the end of 2023 its volume will reach 33 billion rubles.

Assessment Just AI according to the results of 1 quarter 2018

In 2019-2020 dynamics of growth of the market will become even more explosive, it will be promoted by the publication of results of the first successful projects and also mass emergence of the consumer equipment supporting NLU, for example, of smart columns, voice interfaces for management of mobile applications. At the same time, since 2019 demand in a public sector and some other the industries considerably will grow by systems created based on NLU. By 2021 availability of the "talking" chat-bot, de facto, will become a norm for most Internet services. As a result, the domestic market of colloquial AI, chat-bots and intellectual assistants will increase almost three times year by year. After 2020 growth will slow down a little, however will remain rather high up to 2023.

Feature of the Russian market of chat-bots with support of NLU – low activity of world players. Apparently on the Card of the market of colloquial AI, not all products of world leaders have the Russian localization. So, neither Alexa from Amazon, nor Cortana from Microsoft, nor even work as Google Assistant with Russian. "At the same time in Russia there are many highly professional commands with experience in the field of speech technologies. It opens new opportunities before domestic developers who actively develop not only in domestic market, but also the countries go beyond", - the managing director of Just AI Kirill Petrov notes.

Growth of interest in creation of chat-bots with NLU is promoted by the policy of domestic IT giant of "Yandex" which invests heavily in development of speech technologies, creating, both tools for developers, and sales channels of the speech systems. In October, 2017 the Russian voice assistant Alice which the third-party companies can use for communication with clients was started, also he can be integrated into any user devices: smartphones, smart columns, children's toys. In March of this year the IT giant started the platform " Yandex.dialogs " allowing to register and develop own chat-bots or services for Alice.


Banks begin to use chat-bots

High growth rates of the market of chat-bots with NLU technology are caused by an early stage of its formation. The first the IT and telecom companies responded to new technology trends: the share of projects in terms of money in 2017 made 44%. Since 2017 with NLU banks, financial and insurance companies began to show interest in chat-bots. This year many of them for communication with clients implemented <speech technologies. As the result, at the end of year their share can grow from 25% to 28%. Also retail looks narrowly at voice chat-bots, the number of projects in this sphere in 2018 will increase several times.

TADetails: a chat revolution or as bots considerably change business

Chat-bots, or virtual interlocutors, are even more often considered an alternative for creation of communication between the client and business. The expert Oleg Chayevsky told about standard problems and benefits of their use in the heading TADetails.

Personal assistants

Personal assistants are a peculiar incarnation of chat-bots, though more widespread for the reason that the technology develops the largest IT companies. For 2016 hundreds of millions people interact with personal digital assistants on such platforms as Google, Apple, Amazon, Facebook and others. This technology using personal assistants and chat-bots does transition from the graphical user interface (Graphical User Interface, GUI) to the dialogue interface (Conversational User Interface, CUI) by a key trend of the next years.

You watch also (messengers)

The systems of instant messages with users of the website