Design And Implementation Of Intelligent Automatic Outbound Call System

With the development of AI technology,All walks of life are working in the direction of “AI +”.The traditional outbound call system only supports preview,prediction and other tasks that need to be participated in manually in the whole process,The outbound call industry is facing the problems of high labor cost,low efficiency,inaccurate data and high turnover rate.Therefore,enterprises need an intelligent,automatic,convenient and fast outbound call system.The emergence of AI technology provides a solution to this problem.The intelligent automatic outbound call system introduced in this thesis can configure specific business processes,It provides machines to imitate the way people speak,Answer the user's questions,Voice communication with users,Replace the manual agent to complete a single outbound call task.In order to solve the problems existing in the process of intelligent outbound call system,This thesis divides the intelligent automatic outbound call system into three modules for demand analysis,overall design,and database design.The basic function module of the outbound call system realizes the functions of loading the outbound call task,predicting the seat resources and transferring to the manual seat.The system can dial out the phone and record the call process.The automatic outbound call module based on fixed flow is relatively light,A node with multiple functions such as playback,number collection,manual agent transfer and query service is designed to meet the configuration of customer specific business process,In the process of system user interaction,Play the pre recorded voice file and record the user's key information,According to the jump node,the session flow between the user and the system is controlled,Complete a simple questionnaire and other automatic outbound tasks.The automatic calling function based on NLP is the core function of the system,Through the processing of FreeSwitch event,the processing of NLP response and the execution process of different nodes,the main business process is designed in detail.ASR and TTS engine are used to realize the voice interaction function.After freeswitch receives the user's voice flow,ASR technology is used to analyze and recognize the user's voice and convert the user's voice into text,then natural language processing technology is used to process the text content,understand the user's intention and analyze it,and get the answer to the user's question.TTS converts the answer to the question into voice,and finally plays it to the user through FreeSwitch.The whole process realizes the process of system analysis,processing and response to the user's voice without the participation of the human seat.After realizing the functions of each module of the system,several test cases are designed to test the functions of the system,and the expected results and actual results are given.the response time of the system to the user is tested and optimized,and the response time of the system to the user is controlled to one second left and right,so as to ensure the requirements of the system in terms of functions and performance.Finally,this thesis summarizes the work of the system and the future optimization and development of the outlook.