If we tried to boil chatbots down to their basic components, they would turn out to be little more than decision tree diagrams (flowcharts) equipped with a conversational user interface. To make great chatbots that will help to boost your business, you should be able to understand how they operate. Comprehensive knowledge of decision trees and flow maps is a must!
Don’t worry if you are not technically inclined. “Designing conversational flow maps” may sound intimidating, but it is just a mouthful way of saying “drawing flowcharts”.
Everyone can read a flow chart, and everyone can make one! They are very useful when you try to learn how to create chatbots.
What Is a Flowchart?
A flowchart is a diagram which represents actions and steps which follow in a logical and sequential order.
Within the flowchart, there are arrows and special nodes which affect the direction of the flow with specific conditions. For example, there may be a question with A and B answer options, and separate routes have to be taken in each scenario.
Check out this funny flowchart example:
Reading the diagrams like the one above is intuitive and you don’t need any experience to understand what’s going on immediately.
How to Make a Flowchart?
1. Start with pen and paper and draw your flowchart
Before you use a software tree diagram/decision tree maker like Microsoft Visio or any of the free diagramming apps available online, try to design your flowchart with traditional tools.
You can use sticky notes on a whiteboard or draw diagrams with a pencil in a notebook. This way you won’t have to struggle with mastering the user interface of a given tool and you will be able to focus on the design as such, by devoting your full attention to the problem you want to solve.
2. Mind the direction of the flow
Have you ever heard of Rube Goldberg machines? They are built with random objects that create a domino effect – a tennis ball lands on a lever, the lever triggers a spring which shoots a bowling ball, the bowling ball lands into a bucket which is attached to a string, the string pulls a handle, and so on and so forth.
A flowchart is somewhat similar to Rube Goldberg’s contraptions. You should be really careful about keeping track of what goes where and how it affects the overall course of action. One faulty element and the machine stops – the same goes for a flowchart.
3. Use consistent symbols throughout the flowchart
Flowcharts are frequently used in algorithm design and there are various sources you can use, such as The Algorithm Design Manual, to make them more universal and coherent. Some common practices include consistent use of the flowchart design elements, for instance, diamond-shaped nodes for questions, rectangles for actions, ellipses for the start/end nodes, etc.
You don’t have to stick to them, but it is important to come up with the system of symbols that work for you and use it consistently.
4. Avoid flowchart dead-ends
Make sure that you can move freely through your flowchart and eventually reach one of the finishing nodes (sometimes they are called terminators). It should be clear whether you should take another action, return to one of the previous steps within your diagram, or that it is, so to say, the end of the line. There shouldn’t be any orphaned, dead-end points. People who look at the decision tree shouldn’t be confused and they should always know what to do next.
5. Make the flowchart manageable
Sometimes flowcharts and decision trees can become very complex. If your flowchart appears overtly complicated, it may be wise to break it down into several subtasks – each with a separate flowchart.
A “simple” decision tree algorithm with just 7 Yes/No questions can easily produce as much as 128 different scenarios. You should remember to stick to the main “trunk” and the most important branches of your decision tree, without getting caught up in details.
If it becomes too convoluted create a separate flowchart.
Real-Life Flowchart and Tree Chart Examples
Flowcharts are all around us!
Kaplan’s Decision Tree is a good example of flowcharts/decision trees commonly used in real-life situations. It is applied in nursing and helps to determine the nature of a given medical problem.
Did you know?
Kaplan Decision Tree is also known as NCLEX Decision Tree (NCLEX exams are part of becoming a licensed nurse in the United States or Canada). It is a conceptual tool used in nursing practice.
Kaplan Decision Tree is a critical judgment strategy used to objectively evaluate medical intervention problems and make the best decisions by going over a set of steps and questions.
For instance, one of the steps regards the Maslow framework and focuses on determining whether symptoms are physiological or psychosocial.
Every chatbot has a backbone which looks very similar to a flowchart. Usually, they also use decision trees, which are flowcharts that focus on steps that involve questions and branch out to provide different end nodes.
What Is a Decision Tree?
A decision tree is a type of diagram or a flowchart that branches into multiple decision paths through different questions.
Decision trees, or tree diagrams/tree charts, are named for their look and structure – they are similar to upside-down trees with branches which grow into more branches that end with a leaf node.
How to Create a Decision Tree Diagram?
Decision tree diagrams are similar to flowcharts, but their structure is more straightforward. A decision tree begins with a single question. There are follow up questions based on the previous choices, and the structure ends with leaf nodes.
A leaf node, as opposed to a branch, usually offers a solution and ends the diagram structure, as there are no more elements to follow.
Decision Tree Examples
Look at these decision tree examples:
A decision tree is less common to use loops and circular movements. The direction of the flow is constant, and – usually – this type of diagram only goes downwards, or from left to right.
However, decision trees can also be detailed and overwhelming.
In the example below, you can see how a decision tree grows more and more branch nodes to finally produce sixteen leaf nodes (marked in green).
If we rotate the picture by 180 degrees, you can see clearly that the structure bears some resemblance to an actual tree (hence the name).
Circular Flow Chart
Some flowcharts are not supposed to end by definition. When the process is supposed to repeat itself indefinitely we can apply circular flowcharts. These flowcharts can be used to monitor different stages of a given cycle. They are commonly used in a business environment.
Take a look at the flowchart example below which depicts a development life cycle.
Chatbot conversation flows are more sophisticated and a well-designed chatbot shouldn’t repeat its messages and go in circles only. Therefore, circular flowcharts are rarely used in chatbot design (although some steps of chatbot flow maps involve going backwards or returning to previous states).
How to Create a Chatbot Flow Map or a Decision Tree Template?
Rule-based chatbots use conversation flows – carefully designed flow maps which take into account all possible conversation outcomes.
Chatbot flowcharts should be as simple as possible.
The more user decisions and steps they involve, the more chatbot dialogue options you have to predict and script – their number tends to grow exponentially.Check Chatbot Templates
Most chatbots combine flowcharts and solutions which resemble regular decision trees. See this example of an e-commerce chatbot – powered by Tidio – that handles product returns from customers:
Tidio’s editor can be used as a chatbot decision tree maker.
You can use one of the available bots to get the right flowchart templates and customize them. In the chatbot editor, you can also start from scratch to easily create a decision tree template and fill in the blank messages.
Let’s try to analyze an example chatbot flowchart.
As you can see, designing a chatbot decision tree diagram and turning the flowchart into a working chatbot is not that difficult!
Use Flowcharts and Decision Trees to Make Better Chatbots
Chatbots are an indispensable element of modern e-commerce customer service. Being able to design them is a huge asset to your business.
- Chatbots are just like interactive flowcharts, so the same design rules apply
- Nodes within your flowcharts should follow a logical sequence
- Your flowchart must be easy to read and not to complex (divide it into several separate flowcharts if you need to)
- Your flowcharts and decision trees should always end (however, you can influence what steps one has to take to reach the final node)
- You can map your design with pen and paper and use a flowchart maker of a chatbot editor later on
- You can use ready-made templates and adjust them according to your needs (Tidio comes equipped with 14 different conversation flowchart templates)
You should definitely give it a try and create your own chatbot flow maps from scratch or by modifying the existing ones!Try Chatbots For Free