The system supports desktop website, mobile website, App, phone and other channels to access customer service, and can receive users from all channels in one backend. Enterprises can decide whether to use manual customer service or robot customer service for reception based on their own business scenarios. They can also configure priority reception for manual customer service or priority reception for smart robot customer service. Enterprises can configure different departments according to their own business needs and allocate customer service personnel under relevant departments. Visitors can choose more targeted customer service groups for services based on their own business consulting needs. In addition to plain text, customer service can also answer questions through various methods such as emoticons, pictures, rich text, hyperlinks, etc., so that the communication between users and customer service is no longer one-dimensional. Customers can set some automatic response preferences for some specific scenarios, such as customer service is not online, timeout and no response, etc. Customer service can choose to initiate active conversation invitations to users who are queuing or browsing the website to proactively bring users closer and further promote order conversion. After the human customer service establishes a new session with the user, the user can see the chat history of the user with other human customer service or robot customer service, which helps the customer service to better locate the users problem and reduce information gaps. Customer service can see the users basic information including nickname, source, etc. when receiving the user. The relevant source information of visitors can be queried based on customer source channels, search terms, interview pages, landing pages and other dimensions to better help enterprises analyze the delivery of leads. When receiving users, customer service can view the users access track on each page to better understand the users real needs, reduce communication costs, and promote order conversion. Administrators can query user access data from each channel through the two dimensions of session and visitor, and can select different parameters to customize query reports and export reports. Enterprise administrators can not only visually view and statistically analyze the overall work of customer service, but also view the reception status of a specific customer service in a targeted manner. Administrators can customize the query of the reception status of human customer service or robot customer service through different conditions and view the complete conversation records. Based on this, enterprises can better monitor service quality. Visitors can not only rate their satisfaction with the work of human customer service after the session, but also select the corresponding evaluation tag when evaluating.