PRISM: Festina Lente Proactivity -- Risk-Sensitive, Uncertainty-Aware Deliberation for Proactive Agents

2026-02-02

AI Summary

This work addresses the challenge of balancing benefits and burdens in proactive agent interventions, where existing approaches either rely on fragile heuristics or employ undifferentiated lengthy reasoning, failing to precisely control intervention timing and necessity. The authors propose PRISM, a framework integrating decision-theoretic gating with a dual-process reasoning mechanism: interventions are triggered only when the user’s acceptance probability exceeds a threshold derived from asymmetric cost considerations, and a resource-intensive slow-mode reasoning is activated near decision boundaries for counterfactual verification. Inspired by “festina lente,” its risk-sensitive gating focuses on high-stakes ambiguous scenarios. Through gated alignment and structure-locked distillation, PRISM decouples response policy from the intervention gate, enabling tunable, controllable, and auditable behavior. Evaluated on ProactiveBench, PRISM reduces false positives by 22.78% and improves F1 score by 20.14%, significantly outperforming strong baselines while achieving superior accuracy, efficiency, and controllability.

Technology Category

• Reasoning under Uncertainty: Sequential Decision Making
• Multiagent Systems: Adversarial Agents
• Planning, Routing, and Scheduling: Model-Based Reasoning

Application Category

• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

Proactive agents must decide not only what to say but also whether and when to intervene. Many current systems rely on brittle heuristics or indiscriminate long reasoning, which offers little control over the benefit-burden tradeoff. We formulate the problem as cost-sensitive selective intervention and present PRISM, a novel framework that couples a decision-theoretic gate with a dual-process reasoning architecture. At inference time, the agent intervenes only when a calibrated probability of user acceptance exceeds a threshold derived from asymmetric costs of missed help and false alarms. Inspired by festina lente (Latin:"make haste slowly"), we gate by an acceptance-calibrated, cost-derived threshold and invoke a resource-intensive Slow mode with counterfactual checks only near the decision boundary, concentrating computation on ambiguous and high-stakes cases. Training uses gate-aligned, schema-locked distillation: a teacher running the full PRISM pipeline provides dense, executable supervision on unlabeled interaction traces, while the student learns a response policy that is explicitly decoupled from the intervention gate to enable tunable and auditable control. On ProactiveBench, PRISM reduces false alarms by 22.78% and improves F1 by 20.14% over strong baselines. These results show that principled decision-theoretic gating, paired with selective slow reasoning and aligned distillation, yields proactive agents that are precise, computationally efficient, and controllable. To facilitate reproducibility, we release our code, models, and resources at https://prism-festinalente.github.io/; all experiments use the open-source ProactiveBench benchmark.

Problem

• proactive agents
• selective intervention
• cost-sensitive decision making
• false alarms
• benefit-burden tradeoff

Innovation

• cost-sensitive intervention
• decision-theoretic gating
• dual-process reasoning
• aligned distillation
• proactive agents