A Self-Triggered Agentic Push Recommendation System
2026-08-03 • Information Retrieval
Information Retrieval
AI summaryⓘ
The authors present STEPS, a new system for sending push notifications that decides both if and when to send messages to users, improving real-time adaptability. Unlike previous methods that either fixed sending times in advance or checked periodically, STEPS uses two AI agents to plan and execute notifications more efficiently. They also added a lightweight filter to reduce unnecessary computations and prevent bad decisions. Testing on the Douyin platform showed STEPS increased user engagement and lowered the rate of users disabling push notifications while greatly cutting computational costs.
push notificationrecommendation systemdecision transformerself-triggered agentclosed loop systemonline A/B testingcomputational overheaduser engagementresource constraints
Authors
Zhao-Yu Zhang, Qingying Chen, Chunyuan Zheng, Jing Zhou, Jian Sun, Siqi Chen, Leiying Chen, Chuan Zhou, Huiyou Jiang, Xin Tao, Haoxuan Li, Zhouchen Lin
Abstract
Push notification is a critical recommendation scenario on large-scale platforms, allowing the system to proactively reach users outside the application to improve long-term re-engagement. However, designing an optimal push system requires handling a complex action space for the "whether and when" delivery problem under strict system resource constraints. Existing solutions typically fall into two passive paradigms: pre-planned frequency methods that allocate delivery times via offline modeling, limiting real-time adaptability; and fixed-interval triggering methods that periodically poll the system, creating a strict dilemma between excessive computational overhead and diminished optimal timing capture. Furthermore, such multi-stage frameworks severely suffer from local optima. To overcome these limitations, in this paper, we propose STEPS, a proactive, Self-Triggered End-to-end Agentic Push Recommendation System, which is already fully deployed at Douyin with over 1 billion users. STEPS reformulates push recommendation as a self-triggered agentic process in which the system decides not only whether to send a push, but also when to invoke itself again, thereby forming a closed loop that balances real-time effectiveness and efficiency. Specifically, STEPS consists of two decision transformer-based agents: a planning agent that schedules the next system invocation using a gated ordinal regression method, and an execution agent that decides whether to send a push based on trajectory rewards. Furthermore, we introduce a lightweight filtering agent to both control computational overhead and act as a crucial safeguard against unreasonable planning behaviors. Online A/B testing demonstrates that STEPS significantly increases user active days by 0.2843% and reduces the push permission disablement rate by 1.9089%, while the filtering agent reduces computational overhead by 79.42%.