Context Graphs for Proactive Enterprise Agents

面向主动型企业智能体的上下文图谱

Retrieval-Augmented Generation (RAG) and agentic frameworks have advanced enterprise AI considerably, yet agents remain fundamentally reactive: they wait for a human query before acting. This paper argues that genuine enterprise productivity gains require proactive agents: systems that surface relevant, actionable information to workers before they ask.

检索增强生成（RAG）和智能体框架极大地推动了企业人工智能的发展，但目前的智能体本质上仍然是被动的：它们必须等待人类发出查询指令后才会采取行动。本文认为，要实现真正的企业生产力提升，需要主动型智能体：即能够在员工提出需求之前，就主动呈现相关且可操作信息的系统。

We propose the Context Graph, a live relational data structure that models enterprise entities, their relationships, and state transitions over time. Built on this graph, we define a Delta Detection Engine that continuously monitors state changes, a Proactivity Scorer that ranks candidate insights by urgency, relevance, and persona-fit, and a Surfacing Layer powered by an LLM that delivers ranked notifications with grounded explanations.

我们提出了“上下文图谱”（Context Graph），这是一种动态关系数据结构，用于建模企业实体、实体间的关系以及随时间推移的状态转换。基于该图谱，我们定义了一个持续监控状态变化的“增量检测引擎”（Delta Detection Engine）、一个根据紧迫性、相关性和角色匹配度对候选洞察进行排序的“主动性评分器”（Proactivity Scorer），以及一个由大语言模型（LLM）驱动的“呈现层”（Surfacing Layer），用于提供带有依据说明的排序通知。

We formalize each component, derive a unified Proactivity Score function, and provide a complete end-to-end Python implementation using NetworkX and the Anthropic Claude API.

我们对每个组件进行了形式化定义，推导出了统一的主动性评分函数，并使用 NetworkX 和 Anthropic Claude API 提供了完整的端到端 Python 实现。

Evaluation across three generic enterprise case studies (contract lifecycle management, engineering incident response, and sales pipeline hygiene) demonstrates that context-graph-driven proactivity achieves Precision@5 of 0.83, a false positive rate of 0.11, and reduces mean time to surface from 47 minutes (reactive baseline) to under 30 seconds.

通过三个典型的企业案例研究（合同生命周期管理、工程事故响应和销售渠道维护）评估表明，基于上下文图谱的主动机制实现了 0.83 的 Precision@5（前 5 名准确率），误报率为 0.11，并将平均信息呈现时间从 47 分钟（被动基准）缩短至 30 秒以内。