Learn / Track 3 of 5

AI automation in marketing

For marketers ready to automate a weekly job. By the end you have one automation running and a number that says whether it pays.

Lessons, in order

3.1 AI transformation for a social team: workflows first A plan, not tested yet, Redesign the workflow before you buy the tool.

3.2 Your first automation: a weekly research brief I've run this, The four parts, side by side in n8n and Make.

3.3 How a form became a one-page handover PDF I've run this, The Appointment Handover Sheet tool. Coming.

3.4 Build a landing page with Claude Code, no engineer I've run this, Coming.

3.5 A support-triage agent with a human in the loop A plan, not tested yet, A plan for a first triage agent.

3.6 How to tell if an automation pays for itself A plan, not tested yet, Why most agent pilots never reach production.

3.7 What these tools really cost A plan, not tested yet, What Claude Code costs. Check the date; prices change.

Do it this week

Pick one weekly job. Write the steps down in a doc. Do not automate anything until the steps read clearly.

The proof

DemandNexus: n8n, Make, Claude and written-down SOPs

Next track: AI in performance marketing