When Should the Agent Speak? A Survey of Intervention Timing for Always-On AI Assistants

Capable actors have arrived: agents complete multi-step tasks against verifiable goals, and egocentric perception resolves user intent with high accuracy. What remains unmodeled is the decision of when to act unprompted. We survey intervention timing for always-on assistants (smart glasses, MR headsets, ambient copilots) around a single decision rule, intervene iff the expected benefit of acting exceeds the expected cost of interrupting, and organize the literature into five layers: signals, decision, action, memory, and evaluation. We reconnect two lineages that currently do not cite each other: the 1999–2017 interruptibility literature, which formalized interruption cost rigorously but had no capable actor, and the 2024–2026 proactive-agent wave, which has actors but rediscovers the cost term only in fragments (false-alarm pricing, cognitive load, social violation, compute) that no work unifies. We argue that evaluation is the gating layer. Until intervention quality is measurable, a reinforcement-learning reward for proactivity cannot be defined. We therefore propose the design of a benchmark for open-world intervention timing with an explicit cost term. Six open problems close the survey.