推荐序一
Chatbots are redefining“self-service”for the customers of thousands of companies around the world because they make it possible for people to use their own words to explain what they want to do or what question they want answered.Growth in their use will rely on constant improvement in their ability to understand and respond to what people want，which means that their future relies on developers who are well versed in what chatbots are capable of doing and，more importantly，what they ought to be doing.
As you will learn by reading this excellent resource，advances in Natural Language Understanding make it possible for a computer system to capture what a person says or types into a smartphone，understand the individual’s intent and provide the right answers or recommended actions with great consistency.Next will come the ability to carry on longer conversations，covering several topics or“intents”in the same session.
As the authors of this excellent book so ably point out，it all starts with rapid recognition of the purpose of each conversation and the intent of the individual that initiated.There is a lot of math and data crunching involved，but the real difference is the human factor.The most loved chatbots use their training data and analytical engines to anticipate the purpose of a contact.They rely on human input to create conversation models that are pleasant，as well as effective.And they are the product of an ongoing feedback loop that enables subject matter experts to step in to improve answers when necessary and determine the best course to take when the automated agent finds itself in unfamiliar territory.
As chatbots become more efficient，we expect a majority of individuals to prefer to carry on business with them rather than people.As long as the chatbot has been provided with information from the proper knowledge bases，is trained on appropriate conversation models and continues to refine its answers based on supervised machine learning that makes it possible for them to continuously improve responses based on understanding new words and terminology.
This book is a fantastic starting point for people to learn what they need to build the best chatbots possible.You will have a chance to understand the basic math and statistics that make natural language understanding possible.But that is just a starting point.You will learn how to build dialogs between people and chatbots that are tailored to specific industries and use cases.And you will also become familiar with ways for your chatbots to recognize and accommodate a person’s emotional state or to engage in a multi-turn conversation that addresses multiple topics or objectives.
There is a long way to go for chatbots to fulfill the challenge of carrying on human-like conversations for multiple use cases.There is no better book for developers，like yourself，to help chatbots fulfill on the promise of true Conversational Commerce.
Dan Miller founded Opus Research in 1985 and has made the company a global provider of market intelligence and analysis of software and services that support multimodal customer care and improved customer experiences.Opus Research is focused on“Conversational Commerce”，the merging of intelligent assistant technologies，conversational intelligence，intelligent authentication，enterprise collaboration and digital commerce.
Dan Miller Lead Analyst/Founder Opus Research
dmiller@opusresearch.net