<![CDATA[Online complaint handling has become a critical aspect of digital service recovery as firms increasingly use artificial intelligence to manage customer communication. However, customers may evaluate AI-generated responses differently from humanized responses, particularly when service failures require empathy, sincerity, and relational repair. This study examines how response personalization and response type influence customer forgiveness in online complaint contexts and investigates whether perceived empathy mediates these relationships. A scenario-based experiment was conducted using a 2 × 2 between-subjects design that manipulated response personalization, consisting of humanized versus AI-generated responses, and response type, consisting of accommodative versus defensive responses. Data were collected from 342 participants and analyzed using partial least squares structural equation modeling. The results show that humanized responses generate higher perceived empathy than AI-generated responses, while accommodative responses produce higher perceived empathy than defensive responses. Perceived empathy significantly increases customer forgiveness and partially mediates the effects of response personalization and response type on forgiveness. These findings indicate that effective digital service recovery depends not only on what firms communicate but also on who is perceived to deliver the response. The study contributes to service recovery literature by demonstrating that empathy is a key psychological mechanism through which humanized and accommodative responses repair damaged customer relationships in increasingly automated service environments. The findings suggest that firms should balance the efficiency of AI-based complaint handling with the relational authenticity provided by human involvement.]]>
