In addition to over-the-counter services, the most common time for most people to contact the bank is when they pick up their mobile phones and call customer service. Therefore, customer service has become the most daily and most critical part of customer experience. "The needs of customer service calls are usually similar, but 20% of customers can often occupy 80% of online calls. And many customers think in their hearts: Can they skip the multi-layer voice menu and directly talk to a dedicated person and solve the problem?" Industry Chen Zhaoyuan, who was born in the Engineering Management Institute and worked in the Intelligent Finance Department of Yushan Bank, said. In order to solve customer problems quickly, Yushan introduced a "voice robot" combined Special Database with speech recognition. In the future, customers can directly tell the robot what items they want to inquire about. After the robot understands the semantics, it can guide the customer to the corresponding customer service within a few seconds. For some businesses, the robot can even reply directly to the customer with a "synthetic voice" after understanding it. It is Chen Zhaoyuan's unit who is responsible for speech recognition and the development of machine learning tools. This is also a large-scale project he participated in the first year after graduation. "We not only study technology, but also pay attention to business needs. In terms of technology, Yushan and industry-university cooperative schools jointly develop, and we also discuss closely with the professor team every week to share technology and adjust models with each other; in terms of demand, we frequently communicate with inter-departmental departments. Communicate issues at all levels, including: user experience, front-line customer service pain points, data protection, regulations, and budgets.” Why so much trouble? "Because it's not a lab or research lab demo, but a real service that you and I can use," he said. Not only that, since model bu