跳到正文
原文
教程AWS Machine Learning Blog当事人· Adam Maksimuk·· 2 小时前精选

Greenlight 如何在 Amazon Bedrock AgentCore 上构建自主 SOC

How Greenlight Financial built an autonomous SOC solution on Amazon Bedrock AgentCore

图片来自 AWS Machine Learning Blog
谁
Greenlight Financial
引文
This post explains how Greenlight Financial built an autonomous security operations center (SOC) solution on Amazon Bedrock AgentCore.

发生了什么

Greenlight Financial 介绍了如何在 Amazon Bedrock AgentCore 上构建 Luna——一个用于分流安全和金融风险案件的多智能体安全运营系统。该系统使用 30 多个专用智能体、分层监督器,以及用于从案件处理结果中学习的记忆机制。Greenlight 表示,该系统每月最多处理 4,000 起案件,其中超过 90% 无需人工干预即可解决。需要进一步判断的案件会转交给人工分析师,由他们决定是否采取补救措施。

推荐理由

Greenlight 的生产指标展示了分层监督器架构和记忆流程如何与自主解决率相关;其人工审批边界则说明了该设计如何处理受监管的安全和风险案件。

接下来读什么

查看这件事的完整进展Greenlight Describes Luna, Its Multi-Agent Security Operations System on Amazon Bedrock

这件事到目前为止

Greenlight Financial says its Luna system resolves more than 90 percent of up to 4,000 security and financial risk cases per month without human intervention. Built on Amazon Bedrock AgentCore, Luna uses more than 30 specialized agents, tiered supervisors, and memory that draws on past case resolutions. Cases that need additional judgment go to human analysts, who retain control over remediation decisions.

每天一封,跟上 AI 的重要变化

模型发布、价格变化、算力进展,附原始来源。免费,确认后开始发送,随时一键退订。 看一份往期日报 · 2026-11-06