As the world’s population ages, older people are increasingly confronted with social isolation and loneliness, which can hurt their mental and physical well-being. Even though social connections enhance well-being, many older people struggle to participate because of physical restrictions, the digital gap, and declining social networks. With human-like interactions and round-the-clock availability, virtual assistants (VAs) provide possible answers; nevertheless, most VAs prioritize functional duties above building social relationships. To improve senior social engagement, this study investigates personality-aligned VAs with social coaching characteristics. It looks at how social guidance encourages encounters, how personality alignment strengthens emotional ties, and how these two elements work together to improve social well-being. Three groups are used in an experimental design using a mixed-methods approach: (1) high personality alignment plus social guidance, (2) high personality alignment alone, and (3) poor personality alignment. User comments, social engagement indicators, interaction logs, and psychological tests are all part of the data-collecting process. A customizable interaction module in the VA design automatically adjusts to user emotions by customizing communication methods according to personality attributes. The social guidance module facilitates virtual communities and offline activity recommendations, fostering connections. This study advances research on VAs in social well-being and offers design insights for age-friendly VAs applicable in home care and community services. By reducing loneliness and improving quality of life, it addresses broader societal challenges posed by social isolation. Future work will refine cultural adaptability and usability while exploring broader applications.