When Should Users Check? Modeling Confirmation Frequency in Multi-Step Agentic AI Tasks
TL;DR
A decision-theoretic model is developed to determine time-efficient confirmation point placement and results show that 81% of participants preferred the intermediate confirmation approach over the confirm-at-end approach used by existing systems, and task completion time was reduced by 13.54%.
Abstract
Existing AI agents typically execute multi-step tasks autonomously and only allow user confirmation at the end. During execution, users have little control, making the confirm-at-end approach brittle: a single error can cascade and force a complete restart. Confirming every step avoids such failures, but imposes tedious overhead. Balancing excessive interruptions against costly rollbacks remains an open challenge. We address this problem by modeling confirmation as a minimum time scheduling problem. We conducted a formative study with eight participants, which revealed a recurring Confirmation–Diagnosis–Correction–Redo (CDCR) pattern in how users monitor errors. Based on this pattern, we developed a decision-theoretic model to determine time-efficient confirmation point placement. We then evaluated our approach using a within-subjects study where 48 participants monitored AI agents and repaired their mistakes while executing tasks. Results show that 81% of participants preferred our intermediate confirmation approach over the confirm-at-end approach used by existing systems, and task completion time was reduced by 13.54%.
