To briefly explore the future of Chatbot, we must first understand the current development of Chatbot.
To briefly explore the future development of Chatbot, we must first understand the current development situation of Chatbot. In my case, Chatbot is currently divided into two major formations, namely the FAQ Chatbot in the form of Click Flow, and the claim to be completely in Natural Language Processing. NLP Chatbot.
FAQ Chatbot
FAQ Chatbot first summarizes the established customer FAQs, then clicks on the answer or enters the answer form to guide the user to get answers. It is like the telephone service hotline of the last century, which is changed from voice to text, 0–9 Become a different kind of button. Since it is not an open-end question, you don't need to spend a lot of time doing NLP Training. But the shortcomings are inelastic in comparison. For example, the following sentence:
"I want a Tall Size Hot Cappuccino, no cream"
If you want to use the FAQ Chatbot, you need to split into 4 questions:
"Want to drink?"
"What size?"
"Heat setting?"
"Do you want Cream to face me?"
Instead, such problems can be solved by NLP Chatbot in one single question and answer.
NLP Chatbot
NLP Chatbot is the most popular term in the Chatbot world. It is one of the branches of AI. It allows the machine to learn how to distinguish the human language and understand the meaning of the language. So just use the NLP to know that the guests need a piece of cake. To Tall Size Hot Cappuccino for cream. Of course, in the face of the ever-changing expression of human beings, and the characteristics of Hong Kong people speaking Cantonese and Chinese and English, NLP training takes much longer than other Western languages. If NLP training is insufficient, this Chatbot will look like Siri often appear
Sorry, I don’t know if you talk about it tightly, why don’t I help you on the Internet and intimidate, “I want a Tall Size Hot Cappuccino, don’t you want cream”?
Since the two have their own advantages and disadvantages, how will Chatbot develop in Hong Kong?
Subjective analysis – Chatbot's development in Hong Kong
As a Marketing person, I personally never pursue the most advanced technology, I am more in pursuit of the most problem-solving technology. As I said earlier, the Cantonese NLP Chatbot still has considerable restrictions in Hong Kong. I am not denying the problem of understanding the ability of Cantonese NLP Chatbot in Hong Kong. I know that some Cantonese NLP Chatbots can understand most of Guangdong. Words and grammar between Chinese and English. I think that the current Cantonese NLP Chatbot is still in two major problems – the way NLP Chatbot answers and the acceptance of Hong Kong people's response to Chatbot.
The way Sales / CS answers is very much influencing whether the guests will continue to communicate with you. NLP Chatbot can understand the Cantonese grammar. I don't doubt it, but at this stage, I can judge the tone of the guests, and then respond naturally, so that the guests will continue to communicate with Chatbot. This is doubtful. When guests start to feel annoyed, they are simply not worth the candle.
In addition, it is very important for a new thing to be accepted by the market. If the customer knows that he is answering Chatbot, will he still want to continue to communicate? I think the highest level of Chatbot should be that the guests can 't detect that they are answering with Chatbot, but to develop to this extent, how much development cost is needed to optimize and fine tune at this stage?
So at this stage, I think that the FAQ Chatbot, which is a semi-automatic way to alleviate some of the staff, is cost-effective. The basic common questions are answered by the FAQ Chatbot. When you encounter a problem that Hatbot can 't answer, you will notify the customer service staff. Takeover, providing a highly elastic response.
But I believe that in the near future, NLP Chatbot will gradually replace the role of customer service staff due to advances in technology, and the company will gradually reduce the position of customer service staff until it is completely replaced. How many years does it take ? I believe this will be a matter of 10 years.
References: https://medium.com/rubot-ai/對於-chatbot-在香港的發展-da8dd8145203
Source: https://manychat.com/blog/