There are not enough health professionals or services to meet the growing demand for eating disorder treatment globally. As such, conversational artificial intelligence (AI) agents or “chatbots” may help to fill this gap. We co-designed with young people aged 13-18 years and parents/carers/supporters a rule-based chatbot called JEM™ which provides psychoeducation on body image and eating disorders, as well as brief evidence-based coping strategies. We have previously demonstrated preliminary acceptability and feasibility of JEM™. The aim of this study was to examine user evaluation feedback and analytics in naturalistic settings from a deployment of JEM™ hosted in Australia (Monash University website) and Canada (National Eating Disorder Information Centre [NEDIC] website) for the first 6 months. Chatbot users elected to anonymously complete online surveys throughout the chat to provide feedback. In both country deployments, JEM™ sent and received over 10,000 messages and the most common conversation topic chosen by users was brief coping skills. The online surveys showed that users were predominantly young women in their early to mid 20s, and LGBTIQA+ users represented more than 20% in both countries. User satisfaction rates were high, exceeding means of 75% in Australia and Canada, and more than 70% found what they were looking for from interacting with the chatbot in both countries. Furthermore, there were statistically significant increases (with small effects) in body image satisfaction and mood measures from pre- to post-completion of the coping skills in both countries. The major concern users had with JEM™ was the chatbot not understanding all user prompts owing to the simpler rule-based nature of the design. Overall, these results suggested that JEM™ is supporting a diverse range of users internationally with psychoeducation and brief coping skills. However, use of more sophisticated AI technology would likely increase user satisfaction with JEM™ further.