Sorry or Solution? How AI and Humans Shape Justice in Complaints

Abstract

Service failure refers to situations where a firm fails to meet customer expectations, often leading to complaints. How companies respond to complaints strongly influences customers’ perceptions of justice, which are key predictors of recovery satisfaction. In turn, recovery satisfaction can improve word-of-mouth and strengthen brand loyalty. This study examines how different response framings during service recovery affect perceived justice, and whether it matters if the response comes from a human or an AI agent. A scenario-based experiment is conducted using a questionnaire, in which participants imagine a complaint situation and evaluate either a solution- or apology-framed response. The findings reveal that solutionframed responses lead to higher distributive justice, while apology-framed responses increase interactional justice. Apology responses from human agents tend to score higher on interactional justice than those from AI agents, although a larger sample is needed to confirm this statistically. The study contributes to literature by introducing framing into the service recovery context, exploring agent type as a boundary condition, and emphasizing the value of treating justice dimensions separately. Practical implications are offered on how firms can train human and AI agents. Limitations and suggestions for future research are discussed to support further development of this topic.