Effects of different service failure types and recovery strategies on the consumer response mechanism of chatbots

The increasing application of artificial intelligence to online retailing and the rapid deployment of online robots have made online robot service failures common. This study explores consumer response mechanisms to different types of service failures and recovery strategies of chatbots based on role congruity theory and psychological accounting theory. Questionnaires were used to collect data and test hypotheses. The study found the following. First, chatbot service failures can affect consumers' choice of different recovery strategies. After the functional failure of chatbots, consumers are more inclined for the chatbot to be involved in service recovery. After the nonfunctional failure of chatbots, consumers are more inclined toward human involvement in service recovery. Second, different service recovery strategies affect the level of perceived governance. Compared with humans involved in service recovery, robots have a higher level of perceived governance. Third, the level of perceived governance affects consumers’ willingness to forgive. Fourth, the level of robot intelligence plays a moderating role in the types of service failures affecting the choice of recovery strategies. The findings of this study can enrich the response mechanisms and boundary conditions of online robot service failure and provide important insights for online retail enterprises to effectively respond to robot service failure and make reasonable use of human-robot collaborative work.