New technologies, such as autonomous smart-home devices and consumer goods subscriptions, allow for the automation of consumers’ shopping. When using autonomous commerce, consumers practically outsource their decision making, whereby it is machines or firms making purchase decisions on consumers’ behalf. Algorithm aversion, however, might prevent consumers from using automated commerce and thereby stymie this promising technology. This research project applies principal-agent theory, a theory mostly used in B2B contexts, to the consumer behavior realm to analyze consumers’ willingness to accept service failure in automated retailing schemes. While consumers are less forgiving towards agents in situations with specific expectations (predefined automation), they are more forgiving and actually accepting of service failures in situations with vague expectations (surprise automation). This project, thus, seeks to identify ways for firms to increase consumers’ acceptance of service failures in agency situations and increase lifetime value in automated commerce.