Persuasive conversational agents persuade people to change their attitudes or behaviors through conversation, and are expected to be applied as virtual sales clerks in e-shopping sites. As an approach to create such an agent, we have developed a learning agent with the Wizard of Oz method in which a person called Wizard talks to the user pretending to. be the agent. The agent observes the conversations between the Wizard and the user, and learns how to persuade people. In this method, the Wizard has to reply to most of the user's inputs at the beginning, but the burden gradually falls because the agent learns how to reply as the conversation model grows. Generally speaking, persuasion tactics is important to persuade people efficiently, but it is also useful to reduce the burden of the Wizard because it guides the Wizard to a way of persuasion. In this paper, we explicitly implement persuasion tactics into the persuasive conversation agent. Evaluation experiments show that the burden (the input ratio) of the Wizard was reduced from .55% (without tactics) to 3.3% (with tactics), although the success ratio of persuasion was little improved.
展开▼
