This study aims to examine whether the AI-personalized version of the Medilepsy® app is more effective than the non-AI (standard version without AI personalization) can improve key outcomes, such as medical adherence and transition readiness, among underserved adolescents and young adults with epilepsy, ages 14-24, in Florida, USA.

Outcomes are organized into primary (effectiveness), secondary (usability), and exploratory (language experience) endpoints.

Inclusion Criteria:

• Diagnosed with epilepsy
• Currently prescribed anti-seizure medicine (ASMs)
• Identified as part of an underserved population, defined as meeting at least one of the following: member of a racial or ethnic group historically underrepresented in research or healthcare or low-income status (e.g., eligible for public assistance or government-subsidized health coverage)
• Are between 14-17 years old and able to provide assent, with a parent or LAR present to receive app notifications throughout the study
• Are between 18-24 years old and able to provide informed consent
• Able to read, speak, and write in English
• Resides in Florida
• Receive outpatient medical services

Exclusion Criteria:

• They do not own a mobile device with internet access.
• They have a history of severe intellectual disability.
• They are unable to operate a mobile device (keyboard or touchscreen).
• Reside in an in-patient setting.
• Adults unable to consent.
• Pregnant women (excluded because pregnancy is not relevant to the study focus).
• Prisoners.