AI adoption: should teams batch their requests?

Integrating an AI assistant makes it possible to consult it at every hesitation, notification or new suggestion. The management question is specific: should these prompts set the rhythm of work, or should teams organise consultation periods to preserve continuity of attention?

What the literature establishes

Sophie Leroy (2009, Organizational Behavior and Human Decision Processes) shows experimentally that some attention can remain attached to a previous task, particularly when it is unfinished, impairing performance on the next one. This “attention residue” means that leaving an activity is not enough to disengage from it mentally. Applied to AI, this finding suggests a hypothesis worth testing: some consultations fragment work more than they facilitate it.

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, influenced notably by rule complexity. The mechanism concerns the reconfiguration of cognitive control, rather than merely a subjective feeling of distraction. Moving from a case file to a conversation with AI could therefore introduce a cost separate from the time spent reading its answer.

The premature conclusion

One might conclude that all AI consultations should be confined to fixed time slots. These studies do not justify that prescription: a consultation may directly advance the current task or immediately resolve an obstacle. The useful distinction concerns continuity of activity, not the opposition between working “with” and “without” AI. (our executive and employee training programmes)

What the evidence cannot settle

These studies examine neither current generative assistants nor teams in Sion, so they do not quantify the local cost of an AI consultation. Experimental tasks also fail to reproduce all the constraints of a working day. Counting requests or measuring their duration is insufficient: teams need to observe how work resumes and distinguish avoidable interruptions from necessary requests.

A practical check in Sion

As part of SHR’s “Leading AI adoption in your team” programme, a team in Sion could compare similar work periods, first with unrestricted consultation and then with non-urgent requests batched together. The trial would cover document-based activities with no immediate safety implications: administrative summaries in hydropower, non-clinical documents in healthcare, presentation sheets in viticulture or visitor information in Alpine tourism, without entering sensitive data into an unauthorised tool. For each period, record the number of task switches, the reported time needed to resume the original activity and the deliverables accepted without correction against a predefined assessment grid. Batching would be retained only if continuity improves without reducing quality or delaying urgent requests, treating the result as a local indication rather than causal proof. To go further: explore the Leading AI Adoption in Your Team training in Sion in the canton of Valais, or browse our executive and employee training programmes in Switzerland.

In pictures: Leading AI Adoption in Your Team in Sion

• leading ai adoption in your team
• Sion
• research
• training sion