What can Artificial Intelligence Bring for Contact Centre?
Tuesday, 13 November 2018
The famous artificial intelligence named as Her in 2013 described a story of love between the assistant Samantha and the leading actor, a man. The dub from Scarlett Johansson given to Samantha a unique magic which triggered my thinking: What can Artificial Intelligence Bring for Contact Centre?
Before this, every time we conclude the characters of Contact Centre, I will always add the below sentence like pouring water off a steep roof: The Contact Centre is a created sector using limited resources to serve unlimited possibilities and interacting with uncertain customers, plus horizontal sustained resource management, procedure and result optimization. The limited resources include certain manual positions, certain IVR automatic answering devices and Chatbots. The unlimited possibilities and uncertainty reflect the channel varieties and discreteness of the interaction between the customers and the Contact Centre. The sustained resources include Recording, QA, WFM, monitoring, reporting etc. The procedure and result optimization are mainly for the improvement of the KPI and NPS. However, artificial intelligence appeared suddenly with an unstoppable way. Seems some changes have been made: First of all, let’s see the ChatBot. The traditional chatbot relies on the input of corpus and grammar. Most of them are offline corporate software and they belong to rough
intelligence. The learning capability of offline version is very limited. The processing ability of natural language for NLP is challenged by dialects and multi-language or combined language, like the combination of Chinese and English. That is also the main reason that many people don’t like chatting with ChatBot. The PhD thesis ONLINE makes me believes that to make better ChatBot, it is essential to do online instead of offline. With the help of online big data, the deep artificial intelligence is the key for chatting services. Meanwhile, there is special character for Contact Centre. For example, the ChatBot of the Contact Centre of an insurance company is very different from the one of a ticket booking company. And also, the online learning is totally different. Therefore, the online concept doesn’t have to be meaning that it must be connected to Internet. The effectiveness and accuracy of the big data from its sector is much more important. Second, the voice recognition is applied in a much broader way. As a Contact Centre agent, I saw lots of big companies who are doing voice recognition, like HUAWEI, GOOGLE etc. The Contact Centre sector, as the earliest one to use ASR (Automatic Speech Recognition), I must recognize that the voice recognition is not developed well. The old application of ASR provided an alternative for DTMF when IVR relies automatically. Due to the fact that the offline voice recognition can not reach the customers’ requirement of mature business of the Contact Centre, and also because of the complexity of the development of pronunciation and semantics, even though the voice recognition technology has appeared 10 years ago, it never develops fast and is used well. The complexity of pronunciation, the transferring speed of internet, dialects and accent make the accuracy of voice recognition very poor. The understanding of Semantics mainly concentrates on the complexity of oral expressions and the confidence of result recognition. In contrary, Siri applied on iPhone lets everyone of know: sorry, I don’t understand what you mean. This makes the customers very angry.
The Contact Centre companies thinks ahead. On top of voice recognition, they want to do Speech Analysis, which is used for quality control of recorded files. By converting the voice files into text files, and by searching key words to realize 100% quality control, it is much more accurate than before by random test, which is only 10%-20%. The introduction of artificial intelligence can boost the development of voice application. Nowadays the machine learning ability is still based on offline. The application of online intelligence and more advanced module recognition can not only improve the operating ability of Contact Centre, but also can play a “silent absence” role during the interaction between the customer and Contact Centre agent. The introduction of artificial intelligence can also help identify customers under Voice ID and Face ID, which differs from the traditional “please key in your account and password. Please press # to confirm.” A lot of chatting APP and video face recognition technology has been applied in banks. The artificial intelligence is not a big and unknown thing. It is a changing opportunity for software to provide more personalized and active service. But what exact changes it will bring for Contact Centre, it might be like Samantha in the film Her to make you try to stop but cannot.
By HONGLIN XU, Member