A Study On Agent Application Model In B2C E-Commerce Websites

Agent begins to emerge in more places with the development of network technology. However, most studies have been focused on agent itself by studying its own identity as well as the application of the role of purchasing guide, and rarely put them into task flow. With B2C e-commerce platform, this study is focused on what role the agents play in a modern e-commerce environment. It is the same role played by users and items. We will attempt to find a breakthrough from the users' interaction with the site and establish an application model about agent used in B2C websites. And also, it will improve user experience and provide the basis for the application of agents in the B2C e-commerce sites in future. There are five phases in the whole research process:Phase one:Literature reviewing. Understand the application and research situation of the agents currently, and sum up the research methods.Phase two:B2C websites analysis. Comparatively studying nine B2C websites with larger user groups based currently on the market, we obtain general shopping process of B2C websites,analysis and establish user model of B2C.Phase three:Establishing agent application model. Re-position the virtual character's role in the B2C e-commerce websites through the result above.Then manage paper prototype test and establish the agent application model.Phase four:Validation and refinement. To verify the application of agent, we manage user testing methods and find that agent has significant effect on error avoiding,satisfaction and information transportation in user experience.Phase five:Weight of the application model.We use questionnaire survey analysing statistical weight in the various parts of the application model,which could provide the necessary data to support the design of agent for the future.In this study, we achieve the planned targets by researching the application of agent in B2C e-commerce sites exploringly. Using usability engineering approach and user-centered design method, the author of this paper provides some guiding significance for researching and designing the agent in future.