ArXiv TLDR

PASK: Toward Intent-Aware Proactive Agents with Long-Term Memory

🐦 Tweet
2604.08000

Zhifei Xie, Zongzheng Hu, Fangda Ye, Xin Zhang, Haobo Chai + 8 more

cs.AIcs.CLcs.CVcs.HCcs.MA

TLDR

PASK develops an intent-aware proactive AI agent with long-term memory, using a novel paradigm and a real-world benchmark for complex user needs.

Key contributions

  • Proposes DD-MM-PAS, a general paradigm for streaming proactive AI agents.
  • Instantiates PASK with IntentFlow for demand detection and a hybrid long-term memory.
  • Introduces LatentNeeds-Bench, a real-world benchmark for evaluating proactive agents.
  • IntentFlow matches Gemini3-Flash while identifying deeper user intent under latency constraints.

Why it matters

This paper addresses a critical gap in real-world proactive AI, moving beyond lab settings to tackle complex, ambiguous user needs. By introducing a new paradigm, a robust agent, and a real-world benchmark, PASK significantly advances the development and evaluation of truly intelligent, intent-aware agents.

Original Abstract

Proactivity is a core expectation for AGI. Prior work remains largely confined to laboratory settings, leaving a clear gap in real-world proactive agent: depth, complexity, ambiguity, precision and real-time constraints. We study this setting, where useful intervention requires inferring latent needs from ongoing context and grounding actions in evolving user memory under latency and long-horizon constraints. We first propose DD-MM-PAS (Demand Detection, Memory Modeling, Proactive Agent System) as a general paradigm for streaming proactive AI agent. We instantiate this paradigm in Pask, with streaming IntentFlow model for DD, a hybrid memory (workspace, user, global) for long-term MM, PAS infra framework and introduce how these components form a closed loop. We also introduce LatentNeeds-Bench, a real-world benchmark built from user-consented data and refined through thousands of rounds of human editing. Experiments show that IntentFlow matches leading Gemini3-Flash models under latency constraints, while identifying deeper user intent.

📬 Weekly AI Paper Digest

Get the top 10 AI/ML arXiv papers from the week — summarized, scored, and delivered to your inbox every Monday.