Emotional labor has long been studied in contexts where interactions occur face to face or over the phone. Yet an increasing share of customer service now takes place through chat or written messages. In an article published in the Journal of Applied Psychology, Arik Cheshin, Ella Glikson, Einat Lavee, and Allison S. Gabriel set out to understand how this shift changes the day-to-day work of customer service agents. Their qualitative research draws on observations and 48 interviews conducted with agents, managers, and digital service experts across several service organizations.
The first finding is somewhat counterintuitive. Text-based interactions appear to reduce the emotional load of the job. Agents describe a sense of distance from customers. Reading an angry message on a screen is less intrusive than being yelled at over the phone. The authors suggest that classic forms of emotional labor described by Hochschild, such as surface acting or deep acting, become less central in this context. Emotions are more subdued and interactions become more cognitive and technical.
However, this reduction in emotional strain comes with a different kind of effort. Agents often engage in what the authors call “robotic acting.” Responses are more scripted, revised before sending, and sometimes drawn from prewritten messages. The work becomes more mechanical. Paradoxically, agents must then invest additional effort to show that they are human. They personalize responses, explain that they are assisting several customers at once, or add small signals of empathy to avoid being perceived as a bot.
For managers, the challenges associated with text-based customer service differ from those of a traditional call centre. Agents must manage several conversations simultaneously, maintain a minimum level of relational quality, and navigate between scripts and personalization. Supporting them therefore has less to do with managing emotions and more to do with work design: realistic conversation loads, appropriate tools, and enough discretion to personalize responses.