IA on the job: does replying faster mean staying available?
An assistant can make a professional reply easier to prepare in the evening, even when no response is needed before the following day. The management question is precise: how can easier drafting be prevented from becoming an obligation to remain available outside agreed working hours?
What the literature establishes
Larissa Barber and Alecia Santuzzi (2015, Journal of Occupational Health Psychology) examine workplace telepressure: the preoccupation with and urge to respond quickly to work-related messages. Their findings associate this pressure with difficulties recovering from work, particularly psychological detachment. They encourage a distinction between the technical ability to reply and the pressure felt to do so; their research does not concern generative AI.
The mechanism: leaving work without leaving its concerns behind
Sabine Sonnentag and Charlotte Fritz (2007, Journal of Occupational Health Psychology) distinguish four recovery experiences: psychological detachment, relaxation, mastery and control over leisure time. Their validation study shows, among other things, that detachment is a distinct dimension, rather than simply the absence of work activity. Applied to AI, this framework suggests a hypothesis to test: a reply that has become easy to draft may nevertheless keep work mentally present.
The premature conclusion
One might conclude that all AI use outside regular working hours should be prohibited. These studies do not support that conclusion: freely preparing a draft and feeling obliged to reply immediately are not the same situation. The management challenge is to avoid turning reduced drafting effort into a new norm of responsiveness. (our executive and employee training programmes)
What the evidence cannot settle
These articles measure neither the effect of a generative assistant on recovery nor the effect of a delayed-send rule. Much of the evidence relies on individual reports, and the observed associations are insufficient to establish causality. Nor does a sending time alone reveal whether the recipient feels obliged to reply: expectations need to be documented rather than activity merely monitored.
A practical check in Lugano
As part of SHR’s “IA on the job”, a team at a bank, a fashion or trading company, or an Italian-speaking family SME in Lugano could compare two working weeks, first defining emergencies and each person’s working hours. The first week would provide observations; the second would introduce an explicit rule in Italian: except for an agreed emergency, an AI-drafted message requires no reply before the recipient resumes work. Each week, an anonymous tally would record, among non-urgent work messages received outside working hours, those for which recipients felt a reply was expected before their return, allowing that proportion to be calculated. This comparison would provide a local indicator of perceived pressure, without collecting message content or claiming to demonstrate a causal effect. To go further: explore the AI on the Job training in Lugano, or browse our executive and employee training programmes in Switzerland.
In pictures: AI on the Job in Lugano
- ai on the job
- Lugano
- research
- training lugano