Abstract
Organizations increasingly require employees to return to the office, yet evidence on the consequences of return-to-office (RTO) policies remains limited and typically treats RTO as a binary change. We propose a multidimensional perspective of RTO that distinguishes attendance-level— the proportion of time employees spend in the office over a given period—from attendance-trajectory—changes in attendance over time. Analyzing 13,956 employee-month observations from badge-swipe records for 1,163 employees whose attendance remained stable or increased at a multinational financial services organization over 12 months, we examined how these two dimensions related to ten subsequent outcomes: job performance, stress and burnout, three perceptions of organizational culture, and four forms of artificial intelligence (AI) adoption and utilization. Attendance-level and attendance-trajectory showed distinct patterns of association across outcome domains. For job performance, overall attendance-level was unrelated to subsequent performance; instead, a faster increase in attendance predicted lower performance. Well-being showed the opposite pattern: higher overall and initial attendance-level predicted greater stress and burnout, while trajectory was unrelated to either outcome. Culture perceptions and AI adoption showed more complex patterns involving both dimensions, with higher or increasing attendance generally linked to more positive culture perceptions and deeper AI engagement, though the relevant dimension differed by outcome. These findings indicate that attendance-level and attendance-trajectory are not interchangeable and carry benefits in some domains while imposing costs in others. RTO policies should be designed and evaluated according to both how much and how quickly attendance increases, rather than being assumed uniformly beneficial or harmful.