This presentation will showcase how Crisis Text Line leverages AI and machine learning (ML) to deliver high quality, human-delivered mental health and crisis response and prevention at scale; illuminate clinically relevant elements of crisis conversations for continuous quality improvement; and contribute solution-oriented insights and innovations to the broader field. Crisis Text Line’s volunteer-powered model and technology leverages AI as a tool not to replace human-to-human connection but to transform everyday people from all walks of life into a critical mass of trained volunteer responders offering life-changing support to those in need under live supervision of professional mental health staff, advancing mental health equity through improved access to realtime support. We use AI/ML to analyze our de-identified dataset of over 11 million conversations in the United States in English and Spanish to advance our efficiency and expand our impact. This evidence and insight generation helps us identify patterns, understand emerging needs, and respond more effectively in real time—especially when it comes to high-risk situations like mentions of suicide. Following an overview of Crisis Text Line’s platform and service delivery model, we will describe recent AI/ML clinical research science projects that focus on clinically relevant elements of crisis conversations designed to contribute solution-oriented innovations that both improve our service and benefit global society. Examples include such research efforts as: (1) identifying linguistic features of volunteer responders’ texts most strongly associated with perceived empathy or genuine concern; (2) detecting planned suicide methods in conversation by age and other demographics; and (3) identifying and classifying core issues and emotions that lead texters to reach out. We will also describe the results of recent work to expand AI/ML model building and analysis to conversations in Spanish, and the development of an AI driven simulated conversation environment that allows volunteers to develop their skills before engaging with live texters in crisis. Finally, we will share how Crisis Text Line’s data storytelling approach translates the insights generated by our research team into knowledge products and resources made available to wider audiences in an effort to shape policy, practice, and systems change globally, such as our public crisistrends.org dashboard. About the organization: Crisis Text Line is a leading nonprofit that provides free, 24/7, confidential mental health and crisis response through a volunteer-driven service delivery model and technology platform. With this approach, we have supported over 15 million crisis conversations in multiple languages across four countries including the United States, and we aim to expand access to 25% of the world within the next five years through collaborative partnerships, beginning with expansion in Spanish across the Latin America and Caribbean region in 2025.