In the last century 30 years , Charlie · Chaplin in his own film 《 Modern times 》 Played an assembly line worker , The task is to tighten the hexagon nut . Because of the long working hours , Charlie had hallucinations , When you see the hexagonal buttons on women's skirts on the street, you have to screw them up .
In this scenario , We found that , People are driven by machines 「 Brainwashing 」 了 . The machine doesn't stop , People don't stop . Little by little , People also 「 become 」 The machine , There is no thought , Only mechanical repetition .
Such scenes are still repeated today . for example , In the customer service scenario , Many enterprises introduced intelligent customer service as early as a few years ago to complete some simple work . But the staff responded that , The application of these intelligent technologies has not made their work more efficient , Instead, it forms interference to some extent . because , In order to cooperate with or compete with these machines , Manual customer service must serve customers in a fixed mode as quickly as they do , Achieve business targets . In this work rhythm of blindly pursuing speed , It's hard for manual customer service to focus on service quality , Customers are also dissatisfied with this kind of mechanical service .
such 「 Only techno-enabled 」 Think technology can solve all problems , Ignore people's expressive and understanding ability to establish contact with the outside world , It is a mistake that many enterprises are prone to make in the early stage of introducing Intelligent Technology . After years of trial and error , They came to realize ： The reasonable construction of man-machine relationship is also an important knowledge .
Some time ago, Rong Lianyun held 「AI Intentional Intelligent decision-making 」 At the business conference , Rong Lianyun invited 《 attachment 》 The founding editor of the magazine 、 Technology bestsellers 《 Out of control 》 The author of —— Kevin · Kelly and others discussed this problem together . Kevin · Kelly thinks ,「AI Will be with us , Will work with us ,AI It won't replace us , This is a very important educational process , It is also the focus of human-computer cooperation .」 On how to collaborate , Rong Lianyun CPO Xiong Xiegang 、 Rong Lianyun AI Liu Jie, President of the Academy of Sciences, respectively from the product 、 In terms of technology, their practical experience was introduced .
Reshape the man-machine cooperation mode , Start with respecting people's values
Mentioned earlier , Although intelligent customer service has been in the customer service department for some years , But the layout of machines and people acting in their own way not only does not help enterprises solve too many problems , Instead, it brings trouble to manual customer service , It also reduces customer satisfaction . This can not help but make people reflect ： What is the original intention of the enterprise to set up the customer service department ？
「 We have communicated with a large number of enterprise customer service departments , Everyone has a common view ：
We set up a customer service department , The most important thing is to make the communication between customer service personnel and customers more valuable .」 Xiong Xie just talked about . obviously , If manual customer service only focuses on keeping up with the speed of the machine , This value cannot be realized .
In order to change this situation , Rong Lianyun changed the idea of product design , Instead, design intelligent customer service from the perspective of artificial customer service , Find a way to make intelligent customer service provide more help for manual customer service .
The first thing that the attendants do after entering the post , Is to create a virtual （ machine ） people , Let it grow with you .」 Xiong Xiegang said . When serving customers , Intelligent outbound robot can help manual customer service make calls , Find out the preliminary information of the customer ; Agent assisted robot can provide professional knowledge for manual customer service 、 Customer background 、 Applicable scripts and other reference information , So that the newcomers can quickly enter the State . In my spare time , Intelligent training robot can learn together with artificial customer service , Train each other by simulating user scenarios . If a robot trained by a customer service creates benefits in outbound calls and other scenes , This part of the revenue is also linked to customer service personnel , This increases the customer service staff's sense of gain and achievement .
This positive experience of their work value is very important for the work of customer service , It is directly related to whether customer service can maintain a positive mood when facing customers , And whether you can work in this position for a long time . Only the customer service experience is positive , The service they provide is positive , Communication with customers can be more valuable .
Stick to this route for some time , Xiong Xiegang also got some feedback from the customer service ,「 Many customer service said , They have been engaged in customer service for so long , For the first time, I felt that my work was respected .」「
We hope that through such changes , Really change the situation that human value is ignored in enterprises . That's our intention .」
The machine is enough 「 Understand people's heart 」, To really help people
Let the machine help people 、 Growing up with others has been proved to be a feasible path by Rong Lianyun . But to really help people , The machine should be enough 「 Understand people's heart 」. In Liu Jie's opinion , This means that the machine should complete the transformation from perceptual intelligence to cognitive intelligence , That is, not only to be able to obtain 、 Identify external information , Also have excellent understanding and thinking ability . These abilities can be summarized as ：
understand 、 empathy 、 Expression and strategy .
The improvement of machine understanding ability is inseparable from large-scale pre training model and knowledge map , Rong Lianyun adopts the mode of integrating the two ： First, we use the large-scale pre training model to achieve a very powerful AI The brain , Then on this basis, the ability of information extraction and discovery is added , And connect the extracted fine-grained information with the knowledge and ability behind it 、 linkage , In order to achieve better reasoning .
Empathy focuses on the feelings of users 、 mood 、 Insight into true intentions . In this respect , Rong Lianyun has designed some emotion analysis models and opinion mining models based on natural language understanding , It also puts forward a unique two-way reading comprehension framework . The framework adopts the way of algorithm self question and self answer , Give Way AI The model itself finds out which aspects of the product are commented in the text expressed by the user , And the attitude towards these aspects （ Positive or negative ）, The information that makes up the triplet , To judge the hidden intention of the user .
The efforts in expression and strategy are reflected in the construction of gold medal script . For intelligent customer service , It's not enough just to produce fluent language , Also integrate the understanding of business strategy and users , This requires the model to learn from massive customer service data , Summarize the excellent expression and strategy of artificial customer service , Better serve the needs of users .
Speaking of this , Liu Jie gave an example . If a customer presents a credit card to the bank 「 To raise 」 Request , But according to the rules of the system , This need cannot be met . here , Traditional intelligent customer service may directly give a refusal answer , Perform the simplest duties . But actually , This simple and crude approach shows that the model only has the most basic understanding ability , But empathy 、 Lack of expression and strategy choice , Therefore, the experience brought to customers is very poor .
Liu Jie thinks , The correct approach should be to base the model on AI The brain , Integrate customer background data and knowledge map , To restore the user's real situation and real demands , Then give the corresponding script . That's the only way , Intelligent customer service can really help manual customer service , And help customers , Deliver the information users need most when they need it most .
Take the above withdrawal request as an example . If the model can judge the customer's identity through various information （ For example, just graduated college students ）, Then it can make a reasonable prediction about this identity （ Need to rent ）, Then provide specific suggestions after the prediction is confirmed .
In such an interaction , We will make users feel ,AI The algorithm is not just to answer the current literally reflected question one by one , But can more accurately cut into the needs behind , Help him find a way to solve the problem , Tide over difficulties .」 Liu Jie explained .
Based on the above products and technical concepts , Rong Lianyun has built an air business hall 、 Intelligent voice robot 、 Intelligent Agent Assistant 、 A series of products such as intelligent accompany training , Applications span the Internet 、 Finance 、 manufacture 、 Energy and other industries .
From these products , We can see that ronglianyun attaches importance to people when building products . there 「 people 」 Including employees in the enterprise , It also includes the customers faced by the enterprise . For employees , A handy platform or tool saves you a lot of time and energy , You can focus on business issues that have core values , So as to achieve rapid growth . For customers , An intelligent or artificial customer service with temperature can help you solve more problems , Really get the services you need .
A bank customer service manager mentioned , They found in long-term practice , The greatest value of customer service communication with customers is to reflect people's inclusiveness ,「 Everyone may have only one face with us , But as a customer service , Never know what's behind his communication with you , What are you going through , Carrying something . Only understanding with an inclusive nature , You will really stand in each other's perspective at this moment , To solve his unspoken needs and difficulties . Helped him at the critical moment , He will remember the good of the enterprise .」
Liu Jie revealed in the interview , In fact, in addition to the application scenarios mentioned in the speech , These cognitive intelligence products have more uses , For example, in the after-sales scenario, solve the problems of customer use and fault repair , Realize conversational intelligent recommendation in intelligent marketing scenario . Just like the normal customer service scenario , These communications will be man-machine collaboration , And there is temperature 、 valuable .
At the scene of the event , Rong Lianyun CEO Sun Changxun concluded , This discussion is a good start ,「 We will continue to innovate , Combine efficiency with humanity , Committed to making technology more understanding of people , More intentional , So as to really improve the business efficiency of enterprises , Create an enterprise model of new business civilization in the next five years .」
author[Heart of machine],Please bring the original link to reprint, thank you.