From Conversational Agents Back to Human Agents? Drivers and Inhibitors of Customer Switching Intention and Behavior

ABSTRACT In customer service, companies increasingly replace human agents (HAs) with AI‐powered conversational agents (CAs) because they provide efficiency gains, instant responses, and constant availability. Yet customer service still requires HAs, as customers continue to seek their assistance for different reasons. Against this background, we conducted a two‐wave longitudinal online study of European customers to examine which factors determine switching from CAs to HAs, capturing switching intention in study 1 ( N = 468) and self‐reported switching behavior 2 months later among recontacted participants in study 2 ( N = 237). The pull factor, driven by perceived usefulness of HAs, increases switching intention, whereas the mooring factor, driven by inertia and habit, inhibits both intention and behavior. The push factor reduces switching intention, suggesting that negative CA experiences may lead customers to disengage from customer service rather than switching to a HA. Task difficulty intensifies the push and pull factors while weakening the mooring factor. Switching intention positively influences subsequent behavior. Overall, the findings suggest that HAs are particularly valuable for difficult service tasks that CAs struggle to handle. Firms should therefore identify difficult or unresolved requests and provide customers with a visible and straightforward option to switch to a HA before they disengage from customer service.