Interactive Simulation

AI CMO Team Simulator

This dashboard models the CMO Harness system (`kai-cmo-harness`). Watch a raw campaign brief pass through specialized agent steps. Each step executes localized prompt logic and outputs raw updates, simulating real-world agent orchestration.

Why We Orchestrate Specialized Single-Purpose Agents

Context & Prompt Tension Isolation

Splitting copywriting, quality auditing, SEO optimization, and file packaging into distinct agent personas prevents prompt collapse. Monolithic agents fail when trying to satisfy conflicting goals simultaneously.

Independent Quality Assurance

A copywriter agent cannot grade its own homework. Decoupling the QA Guard ensures auditing remains objective, measuring outputs against strict, rule-based checklists without self-evaluation bias.

Structured Database Logging

Logging intermediate states (briefs, raw drafts, scorecards, optimized text) to a SQLite database (events.db) provides a reliable audit trail. Developers can diagnose pipeline failures without reading raw LLM trace logs.

0
Campaign Brief Ingest
Raw input is structured and logged to SQLite events.db.
Raw Brief
1
Copywriter Agent (Kai)
Writes punchy, opinionated drafts based on the brief.
Drafting
2
QA Guard Agent
Audit against the Four U's: Useful, Urgent, Unique, Ultra-specific.
Quality Gate
3
SEO Specialist Agent
Optimizes headings, anchors, and keyword density.
Optimization
4
Delivery Agent
Packages content and logs completion metrics.
Export
kai-cmo-harness console
Console initialized. Select a campaign template and click Start.
Workspace Output IDLE
Simulator workspace is empty. Start the simulation to watch workspace artifacts render dynamically.