Context
As a small marketing team with no plans to hire, scaling output meant scaling through AI — not headcount. Most teams treat AI as a copy-paste productivity boost. I took a different approach: rebuilding GravityKit’s core marketing operations around AI agents and automation pipelines from the ground up.
My role
I designed and built the AI-first marketing infrastructure, identifying the highest-leverage workflows across content, SEO, analytics, and research, then building the agents, MCP integrations, and reusable skills to execute them.
Actions taken
- Built 30+ reusable Claude skills spanning content creation, competitive research, SEO auditing, and reporting.
- Automated SEO monitoring (404 detection, orphan content identification, missing metadata and alt text), replacing periodic manual audits with continuous, agent-driven checks.
- Connected Metabase and GA4 via MCP, enabling AI-powered analysis that surfaced actionable insights on churn and retention, directly leading to new email automation workflows.
- Built a fully automated landing page pipeline: AI researches the topic, generates imagery and captures screenshots via Playwright, then publishes the page to WordPress via the Block MCP. End-to-end, no manual steps.
Results
- Cut implementation time for core marketing tasks by over 90%: workflows that took hours now complete in minutes.
- Scaled content and landing page output without adding headcount, directly increasing the team’s capacity for strategic work.
- Turned analytics from a periodic reporting task into a continuous feedback loop, with AI surfacing insights that drove measurable retention improvements.
- Established a reusable, skills-based operating model that compounds: every new skill permanently expands what the team can execute.