RESEARCH QUESTION: What are the effects of using a fertility education chatbot, i.e. automatic conversation programme, on knowledge, intentions to improve preconception behaviour and anxiety? DESIGN: A three-armed, randomized controlled trial was conducted using an online social research panel. Participants included 927 women aged 20-34 years who were randomly allocated to one of three groups: a fertility education chatbot (intervention group), a document about fertility and preconception health (control group 1) or a document about an irrelevant topic (control group 2). Participants' scores on the Cardiff Fertility Knowledge Scale and the State-Trait Anxiety Inventory, their intentions to optimize preconception behaviours, e.g. taking folic acid, and the free-text feedback provided by chatbot users were assessed. RESULTS: A repeated-measures analysis of variance showed significant fertility knowledge gains after the intervention in the intervention group (+9.1 points) and control group 1 (+14.9 points) but no significant change in control group 2 (+1.1 points). Post-test increases in the intentions to optimize behaviours were significantly higher in the intervention group than in control group 2, and were similar to those in control group 1. Post-test state anxiety scores were significantly lower in the intervention group than in control group 1 and control group 2. User feedbacks about the chatbot suggested technical limitations, e.g. low comprehension of users' words, and pros and cons of using the chatbot, e.g. convenient versus coldness. CONCLUSIONS: Providing fertility education using a chatbot improved fertility knowledge and intentions to optimize preconception behaviour without increasing anxiety, but the improvement in knowledge was small. Further technical development and exploration of personal affinity for technology is required.