AI Governance — Embedded SaaS + Agent
Cardinal #112 tangible-build · Pieces 106 + 107 + 108 + 109 consolidated · AI substrate beyond standalone tools
What this vessel does
AI is increasingly embedded in tools family members already use (Google Workspace, Notion, Slack), and agents can call multiple tools per user request. This vessel extends AI Governance beyond standalone AI tools (Pieces 100-105 + 110) to embedded SaaS AI monitoring + MCP/agent tracking + hallucination detection + per-family-member usage analytics.
Embedded SaaS AI monitoring (Google Workspace / Notion / Slack)
Piece 106
AI features embedded in existing SaaS apps — Google Workspace AI, Notion AI, Slack AI; monitored at network layer + per-app policy
MCP/agent tool call tracking
Piece 107
Model Context Protocol + agentic tool-call tracking — surfaces what AI agents are actually doing on family-member devices
Hallucination / inappropriate AI response detection
Piece 108
AI-output classifier detecting hallucinated halacha/facts + inappropriate responses; KolBo Approved rabbinic resource positive-redirect
Per-family-member AI usage analytics
Piece 109
Aggregate AI-tool-usage analytics per family member; composes with Piece 95 wellbeing summaries; chaver-sharable opt-in