Papers
arxiv:2605.25971

Anticipate and Learn: Unleashing Idle-Time Compute in Proactive Agents

Published on May 25
· Submitted by
Haoyi Hu
on May 26
Authors:
,
,
,
,
,
,
,
,

Abstract

ProAct is a proactive agent architecture that uses idle-time computation to anticipate user needs and improve task completion efficiency and accuracy.

AI-generated summary

While AI agents demonstrate remarkable capabilities in reasoning and tool use, they remain fundamentally reactive: they compute responses only after explicit user prompts. This paradigm ignores a critical opportunity: the idle time between interactions is largely wasted, leaving agents unable to prepare for future user needs. To bridge this gap, we introduce ProAct, a proactive agent architecture that leverages idle-time compute to anticipate and fulfill likely upcoming user needs. By analyzing evolving dialogue history together with persistent memory, ProAct predicts upcoming needs and iteratively acquires information, allowing the agent to resolve knowledge gaps and prepare evidence before the user initiates a query.To rigorously evaluate proactive capabilities, we also introduce ProActEval, a comprehensive benchmark comprising 200 scenarios across 40 domains, featuring predictable need chains and diverse user cognitive profiles. Empirical results demonstrate significant advantages over reactive baselines. ProAct accelerates task completion by reducing required turns by 14.8%, decreases user effort by 11.7%, and cuts hallucination rates by 28.1% on ProActEval. Furthermore, MemBench evaluations confirm that ProAct achieves state-of-the-art reflective accuracy, underscoring its sustained and robust performance.

Community

Paper author Paper submitter

We are excited to share our work: “Anticipate and Learn: Unleashing Idle-Time Compute in Proactive Agents”.

Most AI agents today are purely reactive: they wait for the next user request before doing any useful work. In this paper, we study a different paradigm — proactive agents that use idle time between interactions to anticipate future needs, gather relevant evidence, update personalized memory, and prepare helpful actions before the user explicitly asks.

We introduce ProAct, a framework for scaling idle-time compute in proactive agents, and ProActEval, an evaluation benchmark for measuring whether agents can predict, prepare, and deliver useful assistance at the right time.

Key ideas:

  • using idle windows as a computational resource;
  • future-need prediction for proactive assistance;
  • background research and evidence preparation;
  • utility-aware delivery decisions;
  • personalized memory updates across interactions.

We hope this work encourages more discussion on how agents can move beyond passive response generation toward genuinely proactive assistance.

the most interesting bit for me is how proact binds a memory-backed policy to idle-time work, with provenance preserved so proactively prepared content stays auditable. that coupling plus a delivery-cost gate prevents pointless interruptions and makes idle compute genuinely purposeful. how would memory quality and decay influence long-horizon planning? btw arxivlens had a solid walkthrough covering these memory and evidence pieces, helpful for parsing the method: https://arxivlens.com/PaperView/Details/anticipate-and-learn-unleashing-idle-time-compute-in-proactive-agents-9201-70f488d5

Sign up or log in to comment

Get this paper in your agent:

hf papers read 2605.25971
Don't have the latest CLI?
curl -LsSf https://hf.co/cli/install.sh | bash

Models citing this paper 0

No model linking this paper

Cite arxiv.org/abs/2605.25971 in a model README.md to link it from this page.

Datasets citing this paper 0

No dataset linking this paper

Cite arxiv.org/abs/2605.25971 in a dataset README.md to link it from this page.

Spaces citing this paper 0

No Space linking this paper

Cite arxiv.org/abs/2605.25971 in a Space README.md to link it from this page.

Collections including this paper 0

No Collection including this paper

Add this paper to a collection to link it from this page.