An international trial has confirmed that an AI-driven tool named PETRUSHKA significantly improves antidepressant treatment outcomes by tailoring medication choices to individual patients. Developed by the University of Oxford, PETRUSHKA represents the first clinical prediction tool in mental health to demonstrate clear effectiveness in a large-scale randomized study. The tool was tested in 2024 across 47 sites in the United Kingdom, Brazil, and Canada, involving over 500 adults diagnosed with major depressive disorder. The primary challenge in treating depression is the frequent trial-and-error approach, which often leads patients to stop medication early due to severe side effects or a lack of perceived benefit. PETRUSHKA addresses this by combining clinical evidence, demographic data, and patient preferences, particularly regarding side effects, to support shared decision-making between clinicians and patients. The algorithm generates personalized recommendations within three minutes, making it feasible for routine use in primary care settings. Results published in JAMA indicate that patients whose antidepressant selection was guided by PETRUSHKA were substantially more likely to continue their treatment. Specifically, the group using the tool was approximately 40% less likely to discontinue their medication within the first eight weeks compared to those receiving standard care. Fewer participants in the AI-guided group stopped treatment due to adverse effects. By the 24-week mark, those in the PETRUSHKA cohort reported greater improvements in both depressive and anxiety symptoms. Andrea Cipriani, a Professor of Psychiatry at the University of Oxford and the study's lead investigator, emphasized that mental health has historically lagged behind other medical fields. He noted that PETRUSHKA allows clinicians to personalize treatment from the outset by merging the best available evidence with what matters most to the patient. This approach aims to help more people in the National Health Service and similar systems stay on the medication that is right for them. Mike Lewis, Scientific Director for Innovation at the National Institute for Health and Care Research, highlighted the transformative potential of combining digital technology with personalized treatment. He stated that harnessing data and innovative tools enables care to be tailored precisely to each patient, reducing the personal and economic burden of depression, especially in settings with limited specialist psychiatric support. Henry Winchester, a 45-year-old freelance writer and study participant from Bristol, shared his positive experience. Previously skeptical of antidepressants due to severe side effects, he found that PETRUSHKA identified a medication with milder side effects that was life-changing for him. He believes this personalized approach could make prescribing easier for general practitioners and better for patients. The tool, an acronym for Personalizing antidEpressant Treatment foR Unipolar depreSsion combining individual cHoices, risKs and big datA, was co-produced with individuals who have lived experience of depression. While further research is needed to evaluate long-term outcomes and cost-effectiveness, the trial provides strong evidence that digital decision-support tools can play a crucial role in advancing precision psychiatry and improving mental health care at scale.