1) What is a chatbot and why do you need one?

A chatbot is an application that simulates the correspondence between a user and a real person. Such bots can be found on most modern websites and platforms, and very often they act as a tool to attract users. For example, you visit some online store, browse various products there, and at some point a box pops up: “If you have any questions - contact us!”, but when you type your question, instead of an answer you are offered to leave a phone number for the store operator to call you back later.

Or, for example, in the chat widget you are offered to choose a certain topic by clicking on the button: Payment, Knowledge Base, Product Question or Connect Operator - the options can be very different. All this is done thanks to pre-configured algorithms (bots), which are used to improve the user experience.

A bot can recognize some user questions and give pre-prepared answers to them. The variants of using bots are limited only by your imagination:

• providing tech support;
• training;
• interactive;
• marketing;
• order placement.

A properly configured bot will help to increase your sales and customer loyalty, as well as facilitate the work of the support team. And setting up and using bots today is possible in almost all popular ways of communication: in chat widgets, social networks and messengers. 

2) How to create and customize a chatbot in DaoDesk?

The entire process takes place in the Bot Builder menu, which is presented as a visual builder and flowcharts:

Start with the main thing: define the goals and tasks that the bot should solve - they directly determine how the development will proceed, what functionality and structure will be used. Planning is a very important stage of creating a chatbot.

It will be able to ask your customers qualifying questions to further distribute them through the funnel:

• will pass the lead to the sales team;
• switch the dialog to a support person;
• start an email chain to warm up the potential client;
• answer the most popular questions.

The bot will free up the team's time for important personalized conversations, and with the visual chatbot builder, you'll assemble any scenario in minutes, even without a developer.

While creating and configuring the bot, you will have a UML diagram (a modeling language that will help illustrate the bot's structure) displayed:

In the upper left part of the screen is the Bot Name - this is the name that will be visible to the end user (client) when communicating in messenger. It will also be displayed in the “Assignee” field in the request, which with the help of reports will allow you to collect statistics and understand how many requests were processed by a particular bot for a period of time.

Next to the name there is a gear with additional settings:

• Phrases to call an agent - if any phrase from the list is written by the client, the bot will be stopped and the dispatcher's “Call an employee” event (condition) will be triggered. This will give an opportunity to customize the application further actions after the client decides to stop communicating with the bot.
• Bot response to agent's call - what the bot will respond to the client if the “Employee Call” event is called.
• Channels where bot will be automatically assigned - when a request is received from the selected channels, the bot will be automatically assigned to them:

By clicking on the + button in the center of the screen you will be able to customize the steps/actions for the bot:

• Add Reply. Performs the normal sending of a response to the client. All the same tags can be used here as when adding a response in the Dispatcher:

• Buttons.

The “Stop bot” and “Go to step” buttons are common for all channels.

For each set of buttons you can specify the name of a temporary variable {temporary_field}, which can be used after the bot finishes processing the request in the dispatcher. For example, the bot offers the user several buttons to choose from, and you want the value chosen by the client to be saved in the text field of the request. This example will be described in detail below.  

• Text menu.

A text menu is the same menu with options to choose from, but instead of clicking a button, the client is offered to enter the answer in the form of text from the proposed options. This option is recommended for channels that do not support buttons (for example, Whatsapp). The bot expects a certain text from the customer to proceed, and until the customer enters one of the suggested options, the bot will default to repeating the request over and over again.

• Request data input.

The bot requests data from the user in a certain format (to specify the client's phone number/email/tariff and any other information). And to make sure that the client entered exactly what he was asked to enter, you can use regular expressions - they allow you to specify a pattern that the string should correspond to. For example, if you ask a user for his phone number, his answer should not contain letters and the number itself should consist of only 10 digits. The more precise the pattern you set up, the more chances that the client will enter the correct data.
 

Features of the bot builder

1) When a user interacts with the bot (clicks on buttons, goes through steps - all actions until the bot is stopped), the bot is automatically assigned as the executor of the request.

If you, for example, want to know how many requests from users the bot has solved without the operator's involvement, you can see it with the help of reports.

2) When the bot is stopped, the executor of the request is automatically changed to “Unassigned”. At this point, the request must be intercepted by the operators - either manually or automatically using dispatcher autodistribution.

3) As long as the bot is listed as a performer, no dispatcher rules will work in this request. The rules will start working after the bot is stopped.

4) You can find out who, when and where made edits in the Bot Builder in the “global audit” menu.