Table of Contents
- What is the chatbot scenario?
- How to create a chatbot scenario
- Step 1. Clarify the purpose of introduction and your company’s issues
- Step2. Set targets and personas
- Step3. Identify expected questions and answers
- Step4. Design the framework of the scenario
- Step5. Create questions and answers for the scenario
- Step6. Register scenario in chatbot
- Step7. Conduct test operation and make corrections
- Points to note when creating a chatbot scenario
- Set questions and answers based on actual inquiries
- Don’t give too many choices
- Make sure the conversation flows naturally
- Regular analysis and improvement after operation
- Chatbot quality depends on the scenario What is a chatbot scenario?
A chatbot that responds to customer questions and inquiries in a conversational manner can automatically present appropriate answers instead of operators. There are various types of chatbots, such as those that allow you to freely customize the content of your answers, and it is important to create scenarios to display appropriate answers. How should you create scenarios to improve the effectiveness of your chatbot implementation and increase customer satisfaction? This time, we will introduce the types of chatbots, how to design scenarios, and points to note when creating them.
What is the chatbot scenario?
First, let me give you an overview of what a chatbot scenario is, including its role.
There are two main types of chatbots: scenario type chatbots and AI type chatbots. In the case of a scenario-type chatbot, the conversation progresses according to the scenario set in advance, and the information requested by the customer is provided. It can be said that it is compatible with services where the content of the question is fixed to some extent.
Scenarios play a role in determining how the chatbot should respond to user questions, and can be said to be an important part of customer satisfaction. Therefore, it is necessary to design and set a scenario suitable for your company at the initial setting stage of the chatbot. For example, if there are many users asking for product information, prepare information such as product price and purchase method.
On the other hand, AI-type chatbots use AI (artificial intelligence) to analyze questions and generate appropriate answers. With tools equipped with machine learning functions, the more questions you answer, the more accurate your answers will be, and you will be able to provide appropriate answers to user requests.
How to create a chatbot scenario
How should a scenario be constructed when a company introduces a scenario-based chatbot? Below are the seven steps.
Step 1. Clarify the purpose of introduction and your company’s issues
First, clarify the purpose of introducing a chatbot, the problem you want to solve, and the scope of support. This is necessary work to set the direction of the scenario and to design it properly.
We will clarify what kind of problems you want to solve with chatbots and what kind of results you want to achieve by comparing them with the internal situation. For example, the purpose is to reduce the call volume to the call center, improve the operational efficiency of customer support, and reduce the burden on operators by reducing the need for manned support.
Step2. Set targets and personas
Next, set the target audience and personas you expect to use your chatbot. This is because there are differences in the information to be provided and appropriate methods of expression depending on the target and persona of the main user. The target is assumed to be the age and attributes of the chatbot user, and the persona is a detailed expression of the target character, such as personality and reasons for using the chatbot.
As with Step 1, setting these in detail makes it easier to design scenarios appropriately.
Step3. Identify expected questions and answers
Based on the perspective of the set persona, identify the questions and answers that the customer would like to know. At first, it is recommended to list as many as possible based on what you can think of and past inquiries. For example, questions about products/services and support content can be considered.
Also, if you classify the questions and answers that you have identified by category, you will be able to design scenarios efficiently.
Step4. Design the framework of the scenario
Design the skeleton of the scenario from the identified questions and answers. Visualize the flow of dialogue with the user using a flowchart.
Anticipate questions that may lead to an answer to one question and clarify the flow. For example, users who ask about product features are more likely to ask about price or how to buy next.
Step5. Create questions and answers for the scenario
Create specific questions and answers based on the framework you designed in Step 4. The point is to be conscious of the language that matches the target and persona.
Also, make sure the information you provide is correct. Providing incorrect information can even affect a company’s credibility.
Step6. Register scenario in chatbot
Register a scenario in the chatbot you want to use. Importing from the dashboard and manual input are common, but be careful of setting mistakes. Chatbots should choose tools that are easy to set up and do not require specialized knowledge.
Step7. Conduct test operation and make corrections
Before publishing, we conduct test operation internally to see if the scenario works as expected, and make corrections as necessary. At first, you may not be able to give an appropriate answer to an unexpected question. Analyze the cause and fix the scenario. At the same time, the accuracy of the chatbot is improved by checking for any oddities in the connection of the scenarios.
Points to note when creating a chatbot scenario
Chatbot scenario design is a factor that affects customer satisfaction. What should we pay attention to when creating a scenario to make the chatbot easy to understand?
Set questions and answers based on actual inquiries
When creating scenarios, set questions based on actual inquiries, not imaginary ones. The point is to list past inquiries and frequently asked questions and post the answers that customers want to know ahead of time. At this time, it is important to cover all the questions so that there are no omissions, so use the flow chart effectively.
Don’t give too many choices
Too many branches (choices) in a flowchart will confuse users and cause them to leave. To avoid overwhelming customers who use chatbots, we recommend sticking to five or so options.
The number of steps (hierarchies) in the conversation should also be as shallow as possible. Prepare simple, clear answers that help users get to the information they want with minimal effort.
Make sure the conversation flows naturally
Long, mechanical answers can be difficult for users to understand. In practice, it is rare for questions and answers to be in a question-and-answer format, so please create a natural dialogue by including elements such as language and emotions in your answers.
Regular analysis and improvement after operation
Scenarios are not finished once set. It is basic to revise periodically based on actual chat data. User habits will change over time, and new questions will arise. In addition, there is a risk of continuing to provide incorrect information when product/service information changes, such as shipping charges on e-commerce sites.
It is important to always maintain an attitude of improvement and increase the usefulness for users.
Chatbot quality depends on scenario
Chatbots can respond to questions and inquiries from customers in a conversational style and automatically present appropriate answers. Scenario making is an important task to display appropriate answers.
The procedure for creating a chatbot scenario is to first clarify the purpose and issues, then set the target and persona. After that, we will sort out the questions and answers and design the skeleton of the scenario. Create specific questions and answers, and finally register the scenario with the chatbot. Carry out test operation and make necessary corrections.
As a point to note when creating a scenario, it is important to set the questions and answers based on the actual inquiry content. Also, it is desirable to moderately limit the number of options without increasing them too much. Additionally, you need to make sure that the conversation flows naturally. It is also important to conduct regular analysis and improvement after operation.
In any case, the important thing is to improve the cost-effectiveness of the chatbot while achieving the initial set purpose of introduction. Let’s aim to achieve goals such as operational efficiency by modifying the scenario with improvements after operation.