AI on My Shoulder: Supporting Emotional Labor in Front-Office Roles with an LLM-based Empathetic Coworker

TL;DR

This work tackles the emotional labor burden borne by front-office CSRs and proposes Care-Pilot, an LLM-powered empathetic coworker designed to enable on-task emotional regulation through cognitive reframing. By synthetic-generation of uncivil incidents and a dual-phase evaluation (linguistic/psycholinguistic analyses and a CSR-focused user study), the authors show that Care-Pilot’s empathetic messages are perceived as more sincere and actionable than human-coworker messages, and that Emo-Reframe can help CSRs avoid negative thinking and recenter problem solving. The user study reveals both the potential and the practical challenges of deploying AI-assisted emotional support, including concerns about overreliance, reduced social connectedness, and integration into existing workflows. Overall, the paper provides empirical, design, and socio-organizational insights into how AI could scaffold emotional labor for workers, with implications for future AI-assisted wellbeing tools in workplace settings.

Abstract

Client-Service Representatives (CSRs) are vital to organizations. Frequent interactions with disgruntled clients, however, disrupt their mental well-being. To help CSRs regulate their emotions while interacting with uncivil clients, we designed Care-Pilot, an LLM-powered assistant, and evaluated its efficacy, perception, and use. Our comparative analyses between 665 human and Care-Pilot-generated support messages highlight Care-Pilot's ability to adapt to and demonstrate empathy in various incivility incidents. Additionally, 143 CSRs assessed Care-Pilot's empathy as more sincere and actionable than human messages. Finally, we interviewed 20 CSRs who interacted with Care-Pilot in a simulation exercise. They reported that Care-Pilot helped them avoid negative thinking, recenter thoughts, and humanize clients; showing potential for bridging gaps in coworker support. Yet, they also noted deployment challenges and emphasized the indispensability of shared experiences. We discuss future designs and societal implications of AI-mediated emotional labor, underscoring empathy as a critical function for AI assistants for worker mental health.

Figures (22)

• Figure 1: Schematic figures showing an overview of our study design.
• Figure 2: The Client-Agent learns different types of complaints through examples. The initial complaint is generated based on specified complaint 'category' and organizational 'domain'
• Figure 3: Abridged version of the prompt for the uncivil Client-Agent describes the details of the role needed to simulate an challenging client.
• Figure 4: Task interface for the user evaluation. The client names were randomly assigned. Appendix \ref{['si:simulation-interface']} contains dedicated figures of the major components of this interface for easier reading.
• Figure 5: The user evaluation findings sections correspond to 3 major top-level themes describing Care-Pilot. These themes were grounded in the normative patterns of client incivility and coworker support that were raised by our participants.