Knowledge is a critical organizational resource. Yet organizational restructuring, project-based work, employee mobility, and demographic change threaten the loss of vital tacit knowledge. To retain this knowledge within organizations, it must be externalized into explicit forms that can be shared and reused. While Generative AI (GenAI) offers promising opportunities through its natural-language-based, situationally adaptive interactions, the Information Systems (IS) field lacks prescriptive guidance on designing artifacts for this purpose. Adopting the echeloned Design Science Research (eDSR) methodology, we design and evaluate the GenAI chatbot artifact externalAIze as an instrument for knowledge externalization. We derived four design principles: phased reflective questioning, curated input options with free-text fallback, adaptive motivational feedback, and human-in-the-loop knowledge validation. We evaluated the artifact via a quantitative study with 53 participants, comparing three interaction modalities. The evaluation demonstrates that curated input options with free-text fallback optimally balance users’ cognitive effort and efficiency with expressive freedom. Furthermore, phased reflective questioning effectively elicits deep expertise, adaptive motivational feedback supports user engagement and willingness to articulate knowledge, and human-in-the-loop knowledge validation successfully improves the overall quality of externalized knowledge and mitigates GenAI hallucination risks. Our results contribute to the literature by moving beyond the descriptive potential of GenAI in knowledge externalization to provide prescriptive design knowledge.