Human conversations tend to follow a natural flow and have several characteristics in common depending on the type. This video introduces how to incorporate those characteristics into a chatbot conversation to make visitors more comfortable using the bot.

In addition, the type of bot that you are building will define the conversational flow. Lead generation bots, product sales bots, or engagement bots follow different conversational flows. A critical component of any conversation is the sequence in which points are discussed. In this video, we also talk about how to best determine the natural order of choices for your chatbot.

Designing a Chatbot’s Conversational Flow Starts with a Welcome Message

First impressions are critical. A welcome message is the first message that a chatbot displays. Your bot’s greeting is often the first thing users see when visiting your page.

The process of earning a potential customer’s trust begins from the moment they land on your site, and the right chatbot welcome message is the perfect first step to helping them feel valued.

The entire conversational flow depends on this first greeting content. The ideal welcome message is a short communication that helps a user connect with your business and encourages their engagement. A chatbot greeting message is a great way to achieve instant impressions and interest with new or recurring customers.

A best practice is to make sure to introduce your bot as what it is—a chatbot. This sets expectations with visitors to your site as to what your chatbot can and cannot do.

Navigational Flow and Chatbot Dialog Structure

From there on it is your job to design a conversational flow that will keep the visitor engaged with your bot. To achieve that, there must be a logical sequence so that the user knows at any point in time where they are and where they are going.

User Interaction Design and Information Architecture

Another dimension to designing the chatbot conversation is the interface and informational design. To define the information architecture of a chatbot, you need to know :

• What type of information do you want to present?
• In what sequence should the information be presented to keep the bot moving?
• Do you need to capture any user input?
• What media and tone is perfect for your target audience?

These first steps of defining the chatbot flow are about designing and structuring the conversation flow without getting into wordsmithing.

A Great Chatbot Earns Trust from the Start

People who have an unpleasant experience with a chatbot tend to lose any desire to interact with them in the future. Next time around, they will try their best to get to a human as quickly as possible. People that are insecure about what a chatbot offers will also abandon it quickly. For that reason, it is critical to earn the trust up front and keep the trust throughout the entire bot conversation. Here are some tips:

• Get to the goal or defined outcome with minimal choices
• Minimize jump-offs
• Offer “talk to human” option when appropriate
• Explain the rationale for non-obvious user choices
• Do not waste time by asking extraneous questions
• Minimize chit-chat in rule-based chatbots
  (Chit-chat is critical for an AI chatbot. However, if you are dealing with a rule-based bot you should minimize chit-chat.)

Defining the Optimal Sequence of Choices or Decision Points for the User

The best decision points for your chatbot depends on your business goal. What is the  purpose of your bot? For example, do you want:

• a lead generation chatbot (qualify the customer and harvest contact data)
• a product sales chatbot (direct customers to a product with minimal steps)
• an engagement bot (tantalize the visitor to continue engaging)

The business community can tell you what the true objectives or goals for the bot are. This is not a decision that should be left up to the chatbot designer.  

Structuring and Representing Conversation Flow

Representing the conversational flow of a chatbot is a non-trivial decision because it has to fit your way of thinking. It must help you organize your thoughts to be able to come up with the brilliant words and images that are going to win the customer over.

Many bot developers like use case diagrams as a way of depicting a conversation flow. Others prefer to explain the flow in the form of a table or a spreadsheet. My preferred method are decision trees. Decision trees are powerful because, at least in my mind, a structured conversation can be represented easily in a decision tree. In my chatbot course, I walk my students step by step through the conversational design of  “BAXBY” our company’s lead generation bot.