AI: should every alert be delivered immediately?
AI tools can flag an anomaly, suggest a correction or request approval at any time. The managerial question is specific: which requests should interrupt ongoing work, and which can wait for a scheduled review?
What the literature establishes
Sophie Leroy (2009, Organizational Behavior and Human Decision Processes) shows that switching between tasks can leave some attention attached to the previous task, particularly when it remains unfinished. This attention residue can impair performance on the next task. For managers, this suggests that a notification’s brevity should not be mistaken for a negligible cost.
Switching tasks also means switching rules
Joshua S. Rubinstein, David E. Meyer and Jeffrey E. Evans (2001, Journal of Experimental Psychology: Human Perception and Performance) identify time costs associated with task switching in controlled experiments. Their findings support an executive control mechanism involving goal shifting and the activation of relevant rules. Moving from tolerance analysis to an AI-generated approval request could therefore involve more than simply looking elsewhere.
The premature conclusion
One might conclude that all alerts should be batched and uninterrupted work protected in every instance. That would overlook the fact that an interruption can prevent a faulty operation from continuing or enable urgent intervention. The trade-off is therefore between the cost of waiting and the cost of interrupting, rather than a general preference for silence. (our executive and employee training programmes)
What these findings cannot determine
These studies do not examine AI alerts in precision manufacturing and do not establish an optimal notification schedule. Nor do they allow local productivity gains to be calculated solely from the number of interruptions removed. Yet a common practice remains questionable: allowing a tool’s default settings to decide implicitly what deserves an interruption.
A practical check in Neuchâtel
In a Neuchâtel team working in microtechnology, watchmaking or precision microelectronics, compare two comparable working periods: first with the usual notifications, then with only non-urgent requests batched at agreed review points, leaving safety and critical quality alerts unchanged. Record each request, any task switch it triggers and any failure to meet a predefined response deadline; then calculate task switches per equivalent working time and the proportion of requests handled late. A reduction in the first indicator without an increase in the second would be an encouraging signal, not general proof of effectiveness. Within SHR — Swiss Human Resources’ programme « Manager à l'ère de l'intelligence artificielle », this protocol offers a practical exercise in balancing availability with continuity of work. To go further: explore the Managing in the Age of Artificial Intelligence training in Neuchâtel in the canton of Neuchâtel, or browse our executive and employee training programmes in Switzerland.
In pictures: Managing in the Age of Artificial Intelligence in Neuchâtel
- managing in the age of artificial intelligence
- Neuchâtel
- research
- training neuchâtel