Effects of Conversational AI Assistance and Decision Stages on the Flow Experience of Older Users' of an e-Healthcare Decision Tool.

Many older users find it challenging to use Medicare Plan Finder (MPF), an interactive online decision support tool for Medicare plan selections without assistance due to the diversity of plan options and their low internet competency. This study implements an artificially intelligent decision assistant (AIDA) to support older users’ Medicare Part D plan selections on MPF and applies the flow theory to examine (1) the effects of AIDA on older users’ perceptions of online flow (control, concentration, and enjoyment) during a Medicare Part D plan selection on MPF and (2) the moderating role of the decision stage for these effects. Data were collected through an online experiment with a 2 (AIDA: present vs. absence) x2 (Decision Stage: need recognition vs. alternative evaluation) between-subjects design with a U.S. national sample of 420 older (ages of 65 +) Medicare beneficiaries. Structural equation modeling results revealed that AIDA improved the perception of flow in control, which in turn further enhanced the perceptions of flow in concentration and enjoyment. Further, the AIDA effect on flow was significantly stronger in the alternative evaluation stage (vs. the need recognition stage). The findings suggest positive effects of the AIDA intervention in driving older users’ superior online flow experience, especially at the stage closer to the final plan selection and provide significant theoretical and practical implications.