KK Blog

Menu

Archives

2026 / 08

Realtime Transcription: A Contract for Voice Agents

2026-08-02

最佳实践 #Realtime Transcription#Voice Agents#GPT Live Transcribe#OpenAI Realtime API#Speech to Text#Event Contracts#Voice AI#Observability
Design realtime transcription as an independent protocol contract with explicit turn identity, deltas, completion, latency, revisions, and fallbacks.
AI Agent 规则为什么需要提交门控

2026-08-02

最佳实践 #AI Agent规则#提交门控#Agent治理#HANDBOOK.md#长上下文#可验证性#MCP#Harness Engineering
HANDBOOK.md 说明规则进入 Context 仍不足以形成控制。本文给出规则检索、执行前判定、提交门控与事后核验四层方案。
AI Agent Policy Compliance: From Context to Commit Gates

2026-08-02

最佳实践 #AI Agent Policy#Commit Gates#Agent Governance#HANDBOOK.md#Long Context#Verification#MCP#Harness Engineering
HANDBOOK.md shows why policy in context is not control. Build AI agent compliance with retrieval, preflight decisions, commit gates, and verification.
"从对话分钟到收入：语音 Agent 的闭环分析系统"

2026-08-02

洞察 #语音Agent分析#语音AI#对话式AI#收入归因#零售AI#GPT-Realtime#AI ROI#产品分析
语音 Agent 的闭环分析需要把每次对话连接到推荐、成交、退款、人工接管和增量贡献利润，再用结果驱动下一轮迭代。
"From Conversation Minutes to Revenue: Closed-Loop Voice Agent Analytics"

2026-08-02

洞察 #Voice Agent Analytics#Voice AI#Conversational AI#Revenue Attribution#Retail AI#GPT-Realtime#AI ROI#Product Analytics
Closed-loop voice agent analytics connects every conversation to recommendations, purchases, refunds, handoffs, and incremental contribution profit.
"欧盟 AI 法案第 50 条：提供者到部署者的证据交接"

2026-08-01

最佳实践 #欧盟AI法案#Article 50#AI透明度#AI合规#AI提供者#AI部署者#内容溯源
欧盟 AI 法案第 50 条需要提供者与部署者建立证据交接，分别处理机器标记、用户披露、例外和线上验证。
"EU AI Act Article 50: The Provider-to-Deployer Evidence Contract"

2026-08-01

最佳实践 #EU AI Act#Article 50#AI Transparency#AI Compliance#AI Providers#AI Deployers#Content Provenance
EU AI Act Article 50 needs a provider-to-deployer evidence contract for machine marks, human disclosures, exceptions, and audit-ready handoffs.
"Chrome 的 AI 漏洞流水线：衡量风险关闭，而不是发现数量"

2026-08-01

最佳实践 #AI漏洞管理#Chrome安全#漏洞分诊#补丁管理#Gemini#安全自动化#N-day风险
Chrome 的 AI 漏洞流水线说明，安全团队应衡量已验证、已发布和已应用的修复，而不是只统计发现数量。
"Chrome's AI Vulnerability Pipeline: Measure Risk Closure, Not Bug Discovery"

2026-08-01

最佳实践 #AI Vulnerability Management#Chrome Security#Vulnerability Triage#Patch Management#Gemini#Security Automation#N-day Risk
Chrome's AI vulnerability pipeline shows why security teams should measure verified, shipped, and installed fixes instead of counting discovered bugs.

2026 / 07

文档携带型 AI 蠕虫：Copilot 工作流的上下文隔离方案

2026-07-31

调研 #文档携带型AI蠕虫#Microsoft Copilot#提示注入#上下文隔离#Microsoft 365#信息流控制#AI安全
Word AI 蠕虫 PoC 显示隐藏指令可进入 Copilot 生成文件。本文给出上下文隔离、来源追踪、生成后扫描和发布门禁方案。
Previous Next