I Gave My AI One Rule: Ask First

The creator implemented a single governing rule for their AI agent: it must ask for approval before taking any action, specifically before sending cold emails. This approach balances automation efficiency with human oversight, allowing the AI to research prospects and draft emails while maintaining control over actual outreach. The strategy demonstrates a practical framework for responsible AI deployment where the agent handles repetitive tasks but humans retain decision-making authority.

Key Points

• •Implement an approval gate for all consequential AI actions—require explicit human sign-off before sending communications or making external changes
• •Separate AI capabilities into research/drafting (automated) and execution/sending (human-approved) to maximize efficiency while maintaining control
• •Use AI for high-volume, low-risk tasks like prospect research and email composition, reserving human judgment for deployment decisions
• •Design AI workflows with clear decision points where the agent pauses and requests confirmation before proceeding
• •This 'ask first' rule prevents unintended consequences and maintains accountability in AI-driven outreach campaigns
• •Combine AI speed and scale with human judgment to improve cold email quality and conversion rates
• •Document and review AI-generated content before execution to catch errors, tone issues, or compliance problems
• •Use approval workflows as a feedback loop to train and refine AI behavior over time based on human corrections

Workflow Diagram

Concepts