Administrative Interruptions and Batching: What the Research Can Tell Teams

Published October 12, 2026

Administrative Interruptions and Batching: What the Research Can Tell Teams

Research question

Can research on interruption and task switching inform how teams route administrative requests through an assistant? The practical question is not whether all interruptions are bad. Some signals must arrive immediately. The design problem is deciding which requests deserve interruption and which can enter a scheduled queue.

This review considers laboratory task-switching findings, field observations of knowledge work, notification research, and worker well-being measures. These sources study different units and should not be collapsed into one numeric benchmark.

What task-switching studies show

Controlled studies commonly observe switching costs when participants change tasks or rules. The size depends on preparation, cueing, complexity, and practice. These experiments demonstrate a cognitive mechanism, but a laboratory switch is not equivalent to a founder reading an email during a business day.

Field research adds ecological detail: interruptions can increase reported stress and change work pace, yet some interruptions provide necessary coordination. Observational and experimental conditions, job types, and measurement windows vary.

Why batching may help

Batching creates predictable windows for requests that can safely wait. It can reduce the number of unplanned context changes and allows the assistant to assemble related decisions. A daily approval packet, for example, may be easier to process than scattered questions.

Batching also introduces delay. A payment fraud warning, client escalation, or deadline change may lose value if held for the next review. Therefore batching must be paired with an escalation rule, not applied universally.

Designing request classes

Teams can classify work by time sensitivity, reversibility, consequence, and dependency. Immediate alerts should be rare and observable. Same-day decisions can enter a scheduled window. Routine updates can appear in a daily or weekly digest.

The assistant needs authority to acknowledge, gather evidence, and take safe interim steps. Otherwise a queue simply hides unanswered work until the owner opens it.

A local evaluation design

Before changing the system, define interruption events and measure a baseline for two to four comparable cycles. Capture channel, request class, arrival time, response requirement, decision latency, deadline misses, and owner-reported disruption using a stable short scale.

After introducing request windows, compare the same measures and report volume. Segment urgent from routine work. Avoid attributing all change to batching if staffing, workload, or business conditions changed at the same time.

Human factors and preferences

People differ in role demands, attention patterns, and tolerance for queued communication. A rigid schedule may disadvantage collaborators in other time zones or people whose work depends on rapid decisions. Invite structured feedback and preserve accessibility accommodations.

Worker well-being instruments can help teams choose dimensions to observe, but they should not be used as clinical diagnostics or stripped from their intended administration guidance.

Interpretation boundaries

The literature supports the plausibility that unmanaged switching has costs and that notification design changes attention. It does not establish a universal number of interruptions, an ideal batch interval, or guaranteed productivity gains for every executive.

Teams should describe results as local operational evidence: what changed in this workflow, over this window, under these rules. Recheck after novelty effects fade.

Decision guidance

Start with one bounded queue, keep a narrow emergency channel, and publish service expectations. Ask the assistant to log exceptions that arrived in the wrong lane. Use those cases to refine triggers instead of adding more meetings.

The goal is not an interruption-free day. It is a communication system in which urgency reflects consequence, routine coordination has a reliable home, and neither the owner nor the assistant must constantly guess how fast a request should move.

Sources

• Mark, Gudith, and Klocke field experiment
• Monsell task switching review
• NIOSH worker well-being questionnaire
• Microsoft research on notifications and attention

Publication note

This evidence review was published on October 12, 2026. It distinguishes source findings from operational interpretation and does not claim that associations prove causation.