This article is especially for those who
・Have a lot of similar inquiries from customers
・Perform troublesome support work
・Provide many support by telephone and e-mail
If you read this article, you can
・Automatically answer FAQ and reduce the trouble of handling
・Simplify support and will not be chased by miscellaneous tasks
・Reduce inquiries by phone and e-mail, and concentrate on other tasks
"AI (artificial intelligence)" In recent years, opportunities to hear this word have increased. As you know, "AI" is expected to be utilized in various industries and fields.
"AI" is spreading in lots of sectors and fields, and now, attention is paid to improving work efficiency by "AI".
Users dislike kept waiting
Especially, the combination of "AI" and "chat support" is attracting attention. "Chat support" in Europe and the United States began with technical support of customer service, and now it is adopted in a wide range of industries, mainly mail order and web service.
Looking at the survey of New in Contract Research, Inc. (USA), consumers in the United States emphasize chat support. This indicates that chat is getting established as an important tool just like email and phone.
In order of importance to consumers, the agent-supported channels include
・1-800 to live reps (81%)
・Online chat (67%)
・Apps for mobile devices (50%)
・SMS/Text message (46%)
・Social networking sites (39%)
・Online video chat (32%)
（2015年 New inContact Research社）
Furthermore, in the case of "chat support", the satisfaction level gets higher than the mail and telephone.
Why do customers like 'chat support'?
The biggest reason is "They don't have to wait". According to the survey result by BoldChat (USA), the most attractive point of "Chat support" is "being able to respond immediately in real time", very high figures have been announced.
Everyone wants to solve it on the spot if a question arises while using the website. It is natural that you do want to skip the time to get answers from people in charge by telephone and to search for information from the site.
When a question arises, if there is a chat on the site, customers can resolve it on the spot, leading to the purchase and use of service.
Cut labor cost with "AI"
You must understand that you will not make customers wait when you put chat on the site. However, it is quite difficult to respond to all chats manually.
Then, by incorporating "AI" into "chat support", we can answers automatically. If you are in contact with customers, you probably notice that many of the questions are asked over and over again.
In fact, in many industries, 70 to 90% of inquiries are routine questions. By incorporating "AI" automatic response into "chat support", the work will be improved considerably.
How to set automatic response by "AI"
Our "chat support tool", Chat Plus has "AI" function, so you can make automatic response to routine logic.
With "AI" in Chat Plus, you can quickly create an exact match type chat bot. Since you can prepare options (buttons) in advance and automatically respond with the designed text and URL,
・Customers can save the trouble of entering inquiry contents.
・The operator can save the trouble of answering frequently asked questions.
We will show you how to create the first logic using Chat Plus.
3 steps to create logics
1 To identify questions
First of all, we will identify patterns of common inquiries. Inquiries can be identified by the following method.
We will use it if you have history of inquiries so far, such as e-mail and telephone correspondence history.
When summarizing the history (even if you do not keep history), discuss in the company.
・Support know-how book
It is also good to purchase and utilize the support know-how book by industry.
2 Classify questions
Next, we classify the analyzed query pattern by visitor attribute. Visitor attributes are mainly four patterns.
It is an inquiries from the customers of your service and those who are likely to become customers.
It is an inquiries from those who are interested in your service but are not familiar yet.
It is inquiries from so-called rival companies.
It is inquiries from those who arrived from the search result etc although they do not have any interests in your service.
For inquiries from（3）Competitors（4）Passing, there is no need to create a chat bot, so let's remove it from the summarized inquiry pattern.
3 Make questions
Next, sort out common inquiries from (1) fans and (2) prospective customers, and arrange them in a tree style.
For example, for an EC site, there are many inquiries below.
・Confirm delivery date of goods
・Confirm shipping fee
・Confirm order contents
・Forgot the password
・Unable to log in
Then, options for one question should be limited to 3 to 4 choices. More than 5 choices will cause customers to be confused.
In this case,
・I would like to know the inventory and delivery date of goods
・I want to know shipping fee
・I want to check /cancel my order
・I can not log in (I forgot my password)
We can combine similar options, like above.
It is nice to prepare
・"chat with an operator" button
Start operation of chat bot
After preparing the chat bot in 3 steps, you will enter the actual operation phase.
Until we try it, we do not know whether chatbot we prepared is good or not. It is also possible that there are a lot of unexpected questions outside of your choice and that some buttons you prepared are not pressed.
1 Confirm utilization degree and effect of logic
Look at the chat history and consider choices to leave and remove.
・Leave well-pushed options
・Leave options that tend to lead to conversions
・Remove options that were not pressed too much
・Remove options that are hard to use (ex. options with which talks flowed in an unintended direction)
2 Design new logic based on inquiry history in chat
Review options by considering why the talk flowed in an unintended direction, how to lead to conversions, and so on.
Also, organize and add frequently asked questions other than options.
By repeating 1 and 2, it becomes logic of chat bot which is easy to use.
How was it?
I think that you understand about business improvement by combination of "AI" and "chat support".
With three steps of logic creation, it has become possible to automatically answer the fixed question which had been asked repeatedly by e-mail or telephone by "AI", and it is possible to reduce the trouble of correspondence.
With the 2 step of operation, it became possible to raise the effect of chatbot based on the accumulated chat history .
By automating the correspondence with the chat bot and repeating the operation review, you can increase the efficiency of the company as a whole.
Furthermore, we can raise customer's satisfaction by solving the doubt immediately.
Please refer to this article and incorporate "chat support" with "AI".