Eight Years After Siri Was Released, Why Are Chatbots Still so Popular?
Although there are many chatbots, their functions are not ideal at present. How to create a good chatbot? Should we expect technological disruption or consider more comprehensive design? Let's hear what senior experts have to say.
As early as 2011, Apple released the intelligent voice assistant Siri. Apple, which had great expectations for this product, has to face the current embarrassing situation of Siri. Siri either doesn't understand what the user is saying, or replies to the user with a bunch of confusing content. Now Siri's existence is more about teasing or being teased by users.
There are many chatbots on the market, but few of them are popular. "Mental retardation"and "Useless"place.
For example, on the Facebook Messenger app, the voice assistant M was taken offline after three years of operation, and the most commonly used function of Microsoft Xiaoice, which has been constantly upgraded, remains the meaningless "awkward chat".
The reason for this phenomenon is that, on the one hand, people have different expectations for Chatbots, and on the other hand, even with the most advanced technology, it is still difficult to fully understand the human dialogue system. The different emphases in design concepts have led to the final divergence.
So how can we make a satisfactory Chatbot? A senior person summarized a set of "USED" framework.
The best chatbot is the one that is used
Shin graduated from MIT and worked for Microsoft and Standard Chartered Bank (SCB). He is a veteran in the field of financial technology.
He was awarded the "Top 50 Chinese Corporate Innovators" award by CBN Weekly in recognition of his many innovations in SCB digital marketing in China. He currently runs the technology company he founded, Pand.ai, which is committed to building enterprise-level intelligent Chatbots.
Shin Wee in an article " A good chatbot is a USED chatbot", which talks about how to build an intelligent Chatbot. The following is a translation of the article.
Since building an intelligent chatbot for financial institutions and founding Pand.ai with my co-founder a year ago, one question we’ve often been asked is: How good is the chatbot you’ve developed?
To answer this question, we need to explain how we useDeep Learning Natural Language Processing (Deep NLP)To extract the semantics of user input, so that Chatbot can better Understanding the problemand provide more accurate answers.
Here, I will discuss the thought process behind building a chatbot that delights customers. We are now in the process of integrating these insights into a product framework that will form our core design principles, and hopefully will be helpful to anyone who wants to build or use an AI chatbot.
This frame consists of 4 letters:USED , where USED means:
U ：It Understand me
To build a good chat, you first need to make the robotUnderstand the conversation,Otherwise, it fails to respond as expected.
Most Chatbots rely onPattern MatchingThis technology is effective for "understanding" common sentences, such as "How are you?", "What's your name?", etc., but it has difficulty processing complex sentences or uncommon sentences. This is also one of the reasons why some "mentally retarded" customer service exists.
To build a good chatbot,NLP (Natural Language Processing) EngineThere are already some such engines, whichCloud Servicesmethod to analyze and process, and can be easily used in a variety of chat tools.
In addition, there are companies that provide more specialized NLP engines for specific vertical industries or markets, which can distinguish the subtle differences in language, such as the inclusion of jargon or slang, etc.
If you don’t have much background or don’t want to spend time building it yourself, the best way is to choose a suitable and powerful NLP engine so that Chatbot has no obstacles in understanding people.
S：It Serves me
A good Chatbot needs to perform certain functions in specific scenarios. Therefore, the most important thing is to make the robotBe familiar with the details of this field.
For example, if you want to deploy a Chatbot for a sales team to help them work better, you must prepare all relevant products and content to create an ideal conversation in a structured format.
In addition to basic Q&A, you should also consider adding a quiz component to your chatbot so that it can help salespeople refresh their understanding of the product.
E：It Engages me
In the initial stage of chatbot promotion, it is a challenging task to enable users to interact with robots. After all, it is not easy to get users to change their habits and accept new things.
Therefore, it is possible to formulateEffective push strategy,To help users get used to using new tools.
Of course, push doesn’t mean you have to send boring product descriptions, but you have to do it with your heart. For example, a simple Father’s Day greeting message sent to the fathers in the group will be a very heartwarming way.
D：It Delights me
Currently, most chatbots are moving towardsPlayful and cuteAfter all, products that bring happiness are often more attractive. However, with the emergence of more and more chatbots, it may no longer be realistic to expect users to be happy all the time just by saying something clever.
Find ways to make chatbots fun, but be careful not to make them awkward in serious situations.
Thankfully, language isn’t the only weapon a chatbot has. For example, an Easter egg buried in a conversation could be more surprising than any parody-like witticism.
The core of the Golden Rule
This "USED" framework can be used as a guide for building enterprise-level chatbots, but in fact, its core isTake its practicality into full consideration.
There are many KPIs that can measure the quality of a chatbot, such as response time and accuracy, but the most important point is whether it helps achieve business goals.
Whether you want to use chatbots to improve customer service and sales efficiency, or to generate potential customers, as long as it is a method that can achieve business goals, it can be used. For example, the following KPI:
- Number of conversations (not number of messages);
- The percentage of monthly active users (MAU, also known as DAU “daily active” and WAU “weekly active”);
- Number of sessions per active user
The development of Chatbots does not only rely on technology
Developing a chatbot, even an AI-driven one, is not difficult, but if you want to build a good, intelligent, and widely used chatbot, you need to consider:It's not just about technology.
After a period of enthusiasm, the development of Chatbot has gradually become more rational. The development of NLP has never been as fast as that of computer vision or speech recognition. But if you want to make a good Chatbot, perhaps the most important thing to consider is the design concept that can make it useful.
Because according to the current technological development, in the short term,The barriers to Chatbot products are data and design, not technology.
And maybe when building a Chatbot, you canPrioritize service,Then comes the dialogue system - just like humans, who first have ideas in their heads and then express them through dialogue.
There is such a view: a product is composed of many technologies, and only by having a correct understanding of each type of technology can a good product be made. "After all, we are still far from true artificial intelligence, and only those that can be used are valuable."