Companies are increasingly employing text-based chatbots as a time and cost-efficient way to interact with customers. While companies begin to explore anthropomorphic chatbot designs by imbuing chatbots with human-like characteristics, the effectiveness of chatbot anthropomorphism remains unclear. We conducted three experiments to assess the effectiveness of chatbot anthropomorphism in customer–chatbot interactions. By equipping chatbots with human-like linguistic cues, we evoke different levels of chatbot anthropomorphism. Our results show significant positive effects of chatbot anthropomorphism on trust, purchase intention, word of mouth, and satisfaction with the shopping experience. More importantly, we identify social presence as the underlying mediating mechanism of these effects. These effects are robust and not contingent on different shopping contexts distinguished by hedonic versus utilitarian shopping motivations or the disclosure of (non-)sensitive information by customers. The present research derives managerial implications for companies that seek to effectively employ chatbots in customer interactions. Further, this study advances research on customers’ reactions towards anthropomorphized chatbots and demonstrates that social presence is a critical driver of successful customer-chatbot interactions.