Types of Chatbots

There are two main types of chatbots: non-AI chatbots and AI chatbots.

▶ Non-AI Chatbots (Scenario-type)

Scenario-type chatbots follow a scenario and interact with users in a one-way flow. The user selects a question from a number of options and is presented with a series of related questions. The questions identify the final question, and the response is displayed.

If you already have an existing FAQ scenario that can be used, it is relatively easy to create a flowchart-style FAQ, but some services allow you to import Excel data. There is also a method of transferring questions to operators if they are not answered by a branch of the scenario.

This is suited for operations that have questions that can be easily reduced to simple words and their relative answers. Examples include responding to date and time specifications for home delivery, or responding to inquiries about menus at restaurants by proposing menus that meet requirements while confirming preferences such as price range and allergy compatibility. Therefore, it is important to make advance preparations such as creating scenarios and system integration to ensure that the answers to questions are tailored to the content of the inquiry.


▶ AI Chatbot (Interactive)

AI chatbots extract keywords from input conversations. By using natural language processing, they can recognize questions even if they are asked in the language of each questioner, and can handle a wide range of expressions.

It can detect nuances in the user's questions and responses, allowing it to find relevance while guiding the user to a non-routine answer. Highly effective results can be expected by introducing the system to complex manual explanations, responses to individual inquiries, and services with a large number of FAQs.

However, periodic "tuning" is required to utilize the accumulated data and improve its accuracy. Tuning is the process of adjusting AI to improve accuracy. By adding and modifying "teacher data," which is a pair of questions and answers, the accuracy of the answers can be improved by eliminating the variability of the answers.

An example of an application is the use of AI chatbots in an internal portal. By placing AI chatbots in the company portal, they are used to distribute questions to each area, such as information systems, sales, purchasing, human resources, and legal, and to provide answers to more detailed questions.

AI chatbots can guide users to information they have but did not know how to use, such as how to open a shared schedule or view the progress of a project, making it easier to access information.

Some municipalities have also introduced AI chatbots to provide entertainment for tourist attractions, offering suggestions tailored to tourists' requests and itineraries as part of their sightseeing tour guidance. The interactive nature of the service has been well received as it is easy to understand and provides tourist information.

For more information on chatbot types, please refer to " Why Chatbot Operations Fail? Please also see

How AI Chatbots Work

for a brief explanation of how AI chatbots work.


?Elements of AI chatbots generating conversation

Next, we will look at the elements of AI chatbots generating conversation.

1: Keyword analysis

Extract and analyze important keywords from user inquiries. The more accurate the analysis of the keywords that users throw up, the more suitable answers can be derived.

2: Enrichment and accumulation of FAQs

Since chatbots select the most suitable answers from FAQs, the more complete the FAQs are, the more accurate the answers will be. Therefore, it is necessary for the operator to prepare and accumulate FAQs in advance.

Companies that can easily benefit from the introduction of chatbots

This section describes companies that are likely to benefit from the introduction of chatbots.


▶Examples of companies and industries that are likely to benefit from chatbots

・Companies operating mail-order websites
For companies operating mail-order websites, it is essential to provide enhanced customer services through customer support and product usage inquiries. However, with conventional methods, it is sometimes difficult to respond to all requests, such as "It is difficult to call me on the phone," or "I want to ask a question right away instead of sending an e-mail. Chatbots can respond to such requests 24 hours a day, 7 days a week, leading to increased customer satisfaction.

Manufacturing
In the manufacturing industry, chatbots are effective in providing guidance on handling instructions. This is because the content of the guidance is fixed and can be easily standardized, making the transition to chatbots relatively easy. Scenarios can be designed to cover the most frequently asked questions, and explanations can be made easier to understand by combining video and images.

Beauty/health-related
Chatbots are also useful for providing information on products and items tailored to an individual's physical constitution and health condition. It is also possible to branch out by using a selection method to respond to users' concerns and individual consultations. If there is a service that introduces the most suitable product by selecting conditions in this way, users will find it easier to purchase the product that suits them best.

To summarize the above example, if the number of inquiries is particularly large and the content of the inquiries is standardized and easy to answer, the use of a chatbot would be suitable. On the other hand, if the contents of the inquiries are different each time, if the contents of the guidance changes in a short period of time, or if the number of inquiries is small, the introduction of chatbots may not be effective.


▶ Chatbot operation for internal use

Chatbots are also effective for internal guidance.

It is thought that companies of a certain size, regardless of industry, will need chatbots for handling internal inquiries. Examples of how chatbots can be used for internal purposes include the following

Help desk: Inquiries about software, applications, and networks used in business
Internal portal: Guidance to a site that consolidates access to various information for internal use
Information system: Inquiries about operation of servers, networks, PCs, etc.
General administration and back office operations General: ordering and receiving supplies and equipment, various applications, inquiries about benefits and insurance, etc.

Chatbots can be applied to questions and inquiries related to the above areas, which can be expected to contribute to the efficiency of internal operations.

How to choose the right chatbot for your company

If the number of FAQs is not too large and the inquiries are routine, non-AI chatbots can be used. On the other hand, if the inquiries are complex and varied, AI chatbots would be suitable.

In the case of inquiries that are often subject to shaky notation, such as inquiries from an unspecified number of people in a variety of situations, an AI chatbot will not be able to provide a good answer unless it is equipped with an AI.

Also, if you need to use a chatbot as a portal that can be subdivided into various fields, an AI chatbot that can respond to a larger number of questions is recommended.

First of all, it is a good idea to identify what kind of business you want to implement chatbots within your company, and then select a chatbot with a mechanism that fits your business.

Knowing how chatbots work and the types of chatbots to choose the one that best contributes to your business

The type and scope of operations that can be handled differ depending on the type and mechanism of the chatbot. Non-AI chatbot scenario-type chatbots are suitable for a limited range of tasks, but they may not be able to respond to difficult or diverse inquiries. Select a chatbot that best fits your company's needs by carefully considering the business for which the chatbot will be introduced and the expected inquiries.

We provide "HUE Chatbot," an AI dialogue tool equipped with the largest dictionary in Japan. HUE Chatbot" has the advantage of being able to respond to "fluctuations" in the Japanese language. Flexible response is an important factor that directly affects user satisfaction. Please take a look at our introductory materials.