Building an AI-powered editorial system

Context

The content operation at Amplify Media produced articles, newsletters, emails, and social posts, but lacked a coherent strategy. Content pillars, audience profiles, and a consistent editorial framework were missing.
 

Work
 
I designed and implemented an editorial system in three layers:
 

• Strategy: Defined content pillars, audience profiles, and an editorial framework for each publication.

• Operations: Built AI-assisted workflows for information monitoring, relevance filtering, brief generation, drafting, and editorial review.

• Documentation: Established operating guidelines, quality standards, and human review requirements to make the system transferable.

The workflows integrated Claude, Notion, Slack, Codex, and Gmail.
 

Outcome
 
Production time per asset by ~85% and increased output capacity by 4x, sustaining a publish cadence of 7 pieces/week for over a year with a team of two writers.

Content production became more structured and consistent. Each piece served a defined editorial purpose, and the archive became a cumulative asset rather than a collection of disconnected publications.

AI handled research, briefing, and first drafts, while human oversight remained responsible for strategic and editorial decisions.

The entire system was documented so the team could operate and maintain it independently.