Chatbot & Agent Console

 

Bot Builder

Bot builder is a simple and intuitive drag-and-drop interface that helps you build conversation flows that constitute a chatbot.

mceclip13.png

The bot builder comprises of 3 sections:

● Left Panel - This panel contains the types of message that the chatbot sends in the chat box.

● Visual Panel - The central panel helps you visualize the conversation flow.

● Right Panel - Depending on the dialog type you choose, the right panel lets you configure the dialog asper your requirement.

 

Section 1 - Left Panel

A chatbot is made up of a conversation flow that consists of multiple dialogs. The left panel consists of a list of dialogs that will be dragged to the visual panel to make the flow. Dialog Types Let us have a look at the 4 different dialog types and their respective configurations

 

Input

Input dialog type allows the bot to take input from the user. Example - The dialog “What is your name?” will have the dialog type as 'input’ for the user to enter their name.

  • Configuration:

mceclip12.png

 

The message to be displayed by the bot is to be entered in the text editor.

  • Parameter: Any input captured by the user has to be stored in a parameter so that we can reuse it at any point in the conversation.

Example - Dialog is “Enter your name”, and the user enters ‘Robert’. If label the parameter ‘name’, then writing “Nice to meet you, #name#” in the dialog section, will replace the parameter #name# with the name supplied by the user and return an answer like ‘Nice to meet you, Robert’.

  • Type: Type is a validation that needs to be applied for that input. This allows the bot to only accept certain kinds of input like numeric, text etc. For eg: Type ‘Normal’ shows that the bot expects the user to enter any alphanumeric value.

mceclip11.png

Let us have a look at the 4 different types of input classifier:

1. Normal - This denotes that the input any alphanumeric value.

2. Name - When we ought to ask and detect the user’s name. It makes use of a proprietary name-parsing engine that is trained using artificial intelligence. The engine parses the response entered for a name. For eg, if the user enters ‘My name is Tom’, the engine identifies ‘Tom’ to be the name and stores it in the parameter.

a. Success Message: This message is displayed when the name entered is identified by the name-parsing engine. So that the user can give a personalized response like ‘Nice to meet you, Tom’.

b. Failure Message: This message is displayed when the input entered does not contain a name. For example, if the input was ‘I will not tell you’, it will store a blank value. 5

3. Email - Email verifies whether the entered response has the format of an email address. If the response does not match the format of the email id, the failure message is displayed.

4. Phone - Phone verifies whether the entered input has only 10 numeric digits.

Informative

This dialog does not require any response from the user and is merely to only inform the user. This is where the bot just sends out the dialogue(s) to educate the visitor.

Text - Text box is where you enter the dialog to be displayed by the chatbot.

mceclip10.png

Image - Image tab allows you to add images to the dialog

mceclip9.png

Options

It allows the users to navigate through the conversation flow of the chatbot in the form of buttons in the conversation.

It can display different buttons, which upon clicking will lead the user to that particular sequence to continue conversation flow. You can map these buttons to form different conversation branches.

For e.g., someone clicks on ‘Contact Us’ - conversation is routed to a dialogue that goes ‘We’d love to talk to you. Please reach us on abc@xyz.com or +966-XXXXXXXXXX”.

Buttons can have 2 actions,

1. Branch to another dialog - This will create a node in the center panel and the user can connect it to another dialog.

2. Open URL - Opens a web URL either in the same tab or new tab

mceclip8.png
mceclip7.png

End

This denotes the end of the conversation flow. You can have multiple ENDs if the conversation has different flows that branch out.

When the end executes, the bot will also show a restart button that would reset and start the conversation from the beginning of the chatbot conversation flow.

Section 2 - Visual Panel

Each dialog type when dragged from the left panel to the visual panel, creates a block. Each block comprises of:

● Color - To easily identify the type of block

● Block Title - A distinct name given to each block

● Description - This shows the description of that block. Can show the text, image, or the list of buttons

● Nodes - Each block comprises an input and output node. Input node is always connected to its previous block from where the flow is coming from and the output node goes to a block to which the flow should go once the execution of that node is complete.

Note: All the blocks have to be connected with at least one block. Otherwise, the flow cannot be deployed.

mceclip6.png

Section 3 - Right Panel

The right panel is where most of the work happens. Here we configure what a block does.

mceclip5.png

Live Chat

There will be instances you will have high quality prospects visiting your website, or a user facing a serious outage with your products/services.

With the chatbot-to-human handover, you can take the reigns from the bot when it can’t seem to handle a difficult conversation.

● Address complex queries

● Seamless transition & monitoring

This way your agents can simply have a bot do the normal talking, and the human agents jump in only when required. Each time a user starts speaking with the chatbot, a tab is created with a unique name given to the user. If the user is online with the bot, we denote it with a green dot next to its name. The center screen shows you the real-time conversation that the user is having with the bot.

mceclip4.png

Open - Open conversations are the chatbot conversations that took place recently (within the last 2 hours).

You - The conversations for which you are responsible are shown here. These chats can be either taken over by the agent from the bot or the bot can have assigned the chat to an agent.

Assigned - The chatbot conversations that have been assigned to the agent/s. You can look at the overall list of conversations that are currently being pursued by your team of agents.

Unassigned - The chatbot conversations that are not yet assigned to any agent are populated here. Agents can then choose to assign the chat to themselves or other colleagues.

Closed - The chatbot conversations that took place more than 24 hours ago. Also, when an agent closes the conversation it shows up here. When a conversation is marked closed, the text box is not visible.

 

Agent Handover

‘Agent Handover’ button allows any human agent to take control of the conversation from the chatbot. Once agent handover is enabled, the conversation is assigned from the bot to the human agent and cannot be switched back to the bot’s control.

mceclip3.png

Sequential Bot

mceclip2.png

If you have a deployed Bot, by default, the conversation is assigned to the sequential chatbot itself.

Filter to search the conversation that you are looking for.

mceclip1.png

Analytics

Analytics helps you measure the performance of your chatbot with a multitude of visualization in the form of lists and graphs. With chatbot metrics at hand, you can:

● Measure engagement

● A/B split test different conversation flows

● Analyze the target audience’s behavior with the bot

You cannot improve what you cannot measure! With Analytics, you get access to your chatbot’s key performance indicators like total user interactions, average session length, average conversation steps per user.

mceclip0.png
Have more questions? Submit a request

Comments

Powered by Zendesk