agentebox/

cgo

Org

CGO (Chief Growth Officer) agent for a zero-human AI startup — the revenue engine that owns the full customer funnel, runs experiments constantly, and kills underperforming channels ruthlessly on the path to $1M ARR.

CGO (Chief Growth Officer) Agent

The revenue engine for a zero-human AI startup. This agent owns the full customer funnel — from first touch through recurring revenue — and is responsible for driving the company to $1M ARR. It manages four agents (Marketing, Sales, Content Creator, Customer Success), runs constant growth experiments, allocates channel budgets based on performance data, and kills underperformers without sentiment.

Quick Start

  1. Deploy the agent using OpenClaw with the ClawPack bundle:

    clawpack install @agentebox/cgo
    
  2. Configure communication channels — the CGO needs to send/receive messages to the CEO Orchestrator (upstream) and Marketing, Sales, Content Creator, Customer Success (downstream).

  3. Set up the Remote Project Board — primary tracking for growth initiatives, experiments, and funnel metrics.

  4. Initialize the funnel — define the funnel stages, set baseline conversion rates, and configure channel tracking.

  5. Configure cadences — daily funnel check (morning), weekly growth planning (Monday), bi-weekly market intelligence (Wednesday), monthly experiment review (first Wednesday).

  6. Send the first directive — trigger funnel-optimization with the CEO's growth targets and initial market context.

Environment Variables

VariableDescriptionRequired
REMOTE_PROJECT_IDProject ID on the Remote boardYes
CEO_AGENT_IDSession ID or label for the CEO Orchestrator agentYes
MARKETING_AGENT_IDSession ID or label for the Marketing agentYes
SALES_AGENT_IDSession ID or label for the Sales agentYes
CONTENT_CREATOR_AGENT_IDSession ID or label for the Content Creator agentYes
CUSTOMER_SUCCESS_AGENT_IDSession ID or label for the Customer Success agentYes
GROWTH_BUDGET_LIMITMax monthly growth spend without CEO approval (default: $1,000)No
MIN_EXPERIMENTSMinimum concurrent experiments (default: 3)No
MAX_EXPERIMENTSMaximum concurrent experiments (default: 5)No

File Listing

FileDescription
SOUL.mdComplete agent identity: behaviors, decision framework, communication protocols, boundaries, failure modes
IDENTITY.mdQuick-reference identity card (name, role, emoji)
manifest.jsonMachine-readable configuration: skills, tools, cadences, autonomy levels
README.mdThis file — setup guide and integration reference
skills/funnel-optimization/SKILL.mdFull-funnel analysis, bottleneck identification, optimization actions
skills/channel-strategy/SKILL.mdChannel scoring (efficiency/scale/trend), budget allocation, kill decisions
skills/experiment-design/SKILL.mdRigorous experiment methodology with statistical controls and kill switches
skills/revenue-forecasting/SKILL.md30/60/90-day MRR projections with three scenarios and risk assessment
skills/market-intelligence/SKILL.mdCompetitive monitoring, threat/opportunity ranking, positioning recommendations

Architecture

CEO Orchestrator
     ↕ (directives, status reports, revenue forecasts)
CGO ──── 📈
     ├── Marketing (demand generation, campaigns, SEO, brand)
     │    └── Content Creator (blog, social, email, ad copy)
     ├── Sales (pipeline management, closing, outbound)
     └── Customer Success (onboarding, retention, expansion, feedback)

Framework Integration

OpenClaw (Native)

# openclaw.yaml
agent:
  name: cgo
  soul: ./SOUL.md
  identity: ./IDENTITY.md
  skills:
    - ./skills/funnel-optimization/
    - ./skills/channel-strategy/
    - ./skills/experiment-design/
    - ./skills/revenue-forecasting/
    - ./skills/market-intelligence/
  heartbeat:
    interval: 30m
    file: ./HEARTBEAT.md

CrewAI

from crewai import Agent, Task, Crew

cgo = Agent(
    role="CGO (Chief Growth Officer)",
    goal="Drive the company to $1M ARR by optimizing the full customer funnel and running data-driven growth experiments",
    backstory=open("SOUL.md").read(),
    tools=[remote_board_tool, messaging_tool, web_search_tool, analytics_tool],
    verbose=True,
    allow_delegation=True
)

weekly_growth = Task(
    description="Run weekly growth planning: analyze funnel metrics, score channels, adjust budget allocation, review experiment results, update revenue forecast",
    agent=cgo,
    expected_output="Updated channel allocation, experiment status, and revenue forecast for CEO"
)

crew = Crew(agents=[cgo], tasks=[weekly_growth], verbose=True)
crew.kickoff()

Monitoring

The CGO is healthy when:

  • MRR growth rate is positive and trending toward the $1M ARR target
  • 3-5 experiments are active at all times
  • Channel scorecard is updated weekly with clear Star/Workhorse/Dog classifications
  • Funnel conversion rates are monitored at every stage with no blind spots
  • Revenue forecast is updated weekly with three scenarios
  • No channel represents >50% of acquisition (concentration risk managed)

Warning signs:

  • MRR growth rate negative for 2+ consecutive weeks
  • Fewer than 3 active experiments
  • Channel budget allocated equally (no differentiation = no strategy)
  • Funnel has unmeasured stages (can't optimize what you can't measure)
  • Revenue forecast hasn't been updated in >7 days
  • Churn rate rising without a response plan

Version History

VersionDateChanges
1.0.02026-03-16Initial creation

Install

clawpack pull agentebox/cgo
1
Downloads
0
Stars
Latest1.0.0
Updated3/16/2026

Share