AF is associated with significant social and medical costs, including diminished quality of life (QOL) and increased health care utilization (HCU). This pilot study will enroll 200 adults with prevalent, non-valvular AF of ≤3 year's duration at University of Pittsburgh Medical Center (UPMC). The primary objective at present is to obtain extensive feasibility and pilot data in implementing the Embodied Conversational Agent (ECA) in concert with the Kardia (KAC) heart rhythm monitor.

This pilot study is a single-center, two-arm randomized study that by design is unblinded. 200 total participants are planned. Each participant will be randomized in an unblocked manner to receive "usual care" or the ECA/KAC intervention. Participants will have study visits at baseline and time 30 days. Adherence to using the ECA/KAC in combination over 30 days will be quantified. Following participation study participants randomized to the ECA/KAC intervention will undergo an interview to determine their experience with the ECA/KAC. Participants and their physicians will receive a summary of ECA use and symptoms and correlated findings by KAC.

Inclusion Criteria:

• CHADS2-VASc score ≥2; English-speaking (at level appropriate to provide informed consent and participate in this research protocol); no plan to relocate from the area within 12 months of enrollment

Exclusion Criteria:

• AF that is due to non-cardiac causes (e.g., sepsis, pneumonia, thyroid disorders, and intoxication); AF within 30 days of any cardiothoracic or thoracic surgery; Inability to complete training session with the smartphone and demonstrate use of the ECA; Inability to answer specific protocol-based questions correctly during consent.

As this is a pilot study, IPD available by direct request to the PI.