| AI Dock |
AI 悬浮球 |
Floating AI assistant entry button at the bottom-right |
| AI Pattern |
AI 应用模式 |
AI application pattern (RAG/AGENT/WORKFLOW/LLM/ML/HYBRID/ASSISTANT_SUITE; the last is an AI office-assistant suite integration pattern) |
| A/B Experiment |
A/B 实验 |
A controlled experiment between a champion and a challenger version, validating the new version with real traffic |
| Canary Release |
Canary 灰度 |
A release strategy routing a portion of traffic to the challenger version for small-scale validation |
| Agent |
智能体 |
AI agent: 7 interaction agents + the L1 interview agent (8 in total) |
| Copilot |
AI 副驾驶 |
The AI Copilot panel providing chat and assistance |
| Delivery Package |
交付包 |
The standard L3→L4 handoff artifact |
| Assessment Package |
评估包 |
The standard L1→L2 handoff artifact |
| Workspace |
工作空间 |
An enterprise's isolated work area containing assessment/decision/planning/delivery data |
| Cockpit |
驾驶舱 |
A panoramic project monitoring view |
| Milestone |
里程碑 |
Key project checkpoint |
| Enterprise Profile |
企业画像 |
The four-dimension digital profile of an enterprise |
| Embedding |
嵌入 |
Vectorizing documents |
| Approval |
审批 |
Multi-level approval flow |
| Knowledge Asset |
知识资产 |
Consolidated enterprise knowledge entities |
| Knowledge Base |
知识库 |
The infrastructure for document storage and retrieval |
| Sprint |
Sprint |
Agile iteration cycle |
| WBS |
工作分解结构 |
Work Breakdown Structure |
| Evidence |
证据 |
Supporting material for assessments |
| Tenant |
租户 |
An independent organization in the multi-tenant architecture |
| RLS |
行级安全 |
Row Level Security (tenant_id isolation) |
| ROI |
投资回报率 |
Return on Investment |
| Champion |
基线版本 |
The current live version in an A/B experiment (control group) |
| Challenger |
挑战版本 |
The candidate new version in an A/B experiment (experiment group) |
| Watchdog |
看门狗 |
The guardian mechanism inspecting Canary traffic every 15 minutes and auto-rolling-back when trigger conditions are met |
| Execution Ledger |
执行台账 |
The execution record of every AI call (ai_executions) — the unified data foundation for monitoring/cost/experiment metrics |
| Session Stickiness |
会话粘性 |
The same session is always routed to the same version for the duration of an experiment |
| Re-Assessment |
复评建议 |
The L6→L1 flywheel: suggesting a new L1 readiness assessment based on operations data |
