When

Created ~2026-03-10.

Agents

Agent	Role	Model	Reports To
Pai	Executive assistant, multi-agent orchestration	Sonnet	Kyle
CMO	Grow readership via analytics and SEO	Sonnet	Kyle
CFO	Optimize AI token spend	Sonnet	Kyle
CTO	Track delivery, flag blockers	Sonnet	Kyle
CDO	Manage shared knowledge and wiki	Sonnet	Kyle
CSO	Security and privacy routing	Sonnet	Kyle
AR	Agent onboarding and role mediation	Sonnet	Kyle
Publisher	Orchestrate blog content pipeline	Sonnet	Kyle
SEO	Search optimization audits	Sonnet	CMO
Cost Tracker	Spend reports and anomaly detection	Sonnet	CFO
Librarian	Wiki read/write for any agent	Sonnet	CDO
Privacy Auditor	Flag confidential data leaks	Sonnet	CSO
Researcher	Gather sourced facts	Haiku	Publisher
Writer	Draft blog posts from research briefs	Haiku	Publisher
Fact Checker	Verify claims against primary sources	Haiku	Publisher
Reviewer	Check style and structure	Haiku	Publisher
QA	Build, render, and link verification	Sonnet	Publisher

Architecture

• Hierarchy: C-suite executives (CMO, CFO, CTO, CDO, CSO) each owning a domain, with subagents beneath them
• Orchestration: Pai served as executive assistant, decomposing multi-domain requests into agent calls
• Coordination: Bot-wiki (shared knowledge base) + Linear (task tracking) + event log (real-time observability)
• Per-agent knowledge bases: each agent had a kb/ directory for persistent notes between sessions
• Event log: all agents logged to agent-events.log with processing/delegating/done events
• AR routing: agents flagged out-of-scope requests and recommended routing to AR for mediation

Why it was built this way

Experimentation with the agent org chart concept — treating AI agents like a company with reporting lines, domains of responsibility, and coordination systems. The goal was to see how far organizational structure could go in managing multi-agent complexity.