Why chatbot and how to use them correctly. blog of marina someway the koz times database tools

This winter I was invited to speak at conferences in Miami and San Francisco, which was dedicated to the analysis of data, marketing strategies and new developments in the tourism industry. Among the speakers were real sharks travel business, among them: Expedia, AirBnB, Facebook, Booking and other Holdings.

The chatbot complements the existing web platform and ios app, which we passed through the American startup accelerator and which we presented at TechCrunch Disrupt Startup Alley (the Ukrainian pavilion) in San Francisco in the fall of 2017.

Chatbot is, in fact, the voice program simulating human behavior via text or audio channels. In modern life we are constantly faced with a particular embodiment of chatbots, for example, the familiar Siri and Amazon’s Alexa, you can to some extent be considered carbothane, many people use bots in instant messengers.


With the growing popularity of chatbots (according to September 2017 on the platform of Facebook has already created more than 100 thousand chatbots), many small businesses are thinking about whether to follow the example of world trends and large corporations and develop their own bot.

Many of us are faced with a long minute of waiting on the line when he tried to contact one or another technical support. As a result, when you think that the goal has been reached and you were connected with the operator asking this question, you often get the answer: “Wait, please, on the line”, meanwhile the operator generates a response to the request.

Chatbot always. And provided that the question is formulated correctly or has minimal bot AI, the user receives instant answers without spending time waiting. Because the chatbot analyzes the question refers to database data and provides a response within seconds. All this is a clear advantage of bots.

With the growing popularity of chatbots, in the network there are more and more services that allow you to create chatbots without the knowledge of programming and large financial investments. As a result, even a small business can get a helper in the form of a bot which will be responsible for support, collect user data and send them news (a kind of new form of email distribution), etc.

That is, you can tell that the bot now even with fairly reduced functionality can replace several employees. It can be profitable for any business, invest once in the bot – reduced monthly payment of the salaries of potential employees of call centers, technical support and so on. Of course it’s not all perfect and rosy, but this is already under way.

Creating a chatbot, one of the first tasks is to build a “logic of conversation” program. Modeling the interaction of the chatbot user, you or the developer needs to think in advance the types of responses, alleged dialogues and behavior of the chatbot. Thus you will be able to direct the user necessary for you “sales funnel”.

And despite the fact that the chatbot is often a program with elements of artificial intelligence, and specially designed chatbot generally have the opportunity to educate themselves and improve the original logic and purpose of the bot is always laid by the developer. And this is what allows the relative control of each chatbot session of the user.

And, interestingly, has become no matter if the bot on the platform any messenger or applications. View your bot on the market was a good opportunity for public relations, confirmation of the innovation policy of the company and the reminder about your brand.

Potential users of the chatbot, especially older generations, simply can not understand such communication, they do not trust the program and prefer not to talk with the bot. In this case, the chatbot will not be able to replace call center, tech support or another Department.

Despite all the advantages, the bot is just a program and some cases require human attention and immersion in the essence of the problem, not the answer to the query, where the boat will only be able to provide you with pre-programmed or generated them an answer.

At a certain stage of development of the bot raises the question of whether to implement machine learning to enable software to self-learn in the process of communication with the user, to remember the reaction and to improve their skills for subsequent interactions? This can be very useful, as to foresee all the possible user requests and to prepare the bot almost impossible.

If you let the process to “drift”, can be faced with many difficulties, and sometimes even lose control of originally programmed the logic of the conversation. And I do not raise the issue of the possible reputational losses. “Too clever” bot maybe absolutely inadequate to respond to a specific request or boorish behavior of a customer or competitor.

Recently some catboy began to play the role of email distribution. It is quite convenient – a message from the bot, with which the client has already communicated you will not get in a conditional folder, “promotions”, as some email services. In addition, there is a little dismissive (yet) against the messengers and the response wider. What are now popular bots in the Telegram.

In a situation with a bot, the user can easily subscribe to the newsletter, and the program will send notifications. But it’s worth noting that if you do not adhere to certain rules of the game, then the bot turns into a compulsive spammer. The result is obvious – the user simply removes.

The conclusion is obvious, despite all the advantages of bots, the next few years they will not be able to fully compete with a real person. But it is not to underestimate the program – now they are ready to perform many functions and benefit the business and customers. In addition, we must always “keep up with the times”, to follow the modern trends in order not to remain on the sidelines of progress.

banner