Building an AI marketing department

Building Marketing Workflows with n8n

Yaron Been

Marketing Automation Architect

Meet your instructor

$$

Yaron Been headshot

$$

Yaron Been: Marketer, automation architect, and founder of AI-driven marketing tools
Course goal: n8n automations you can deploy, adapt, and improve in your marketing stack

The FlowPilot challenge

recraft: half: Business professional in modern office looking at charts and strategy documents

$$

You joined FlowPilot (B2B startup) as a CMO
❌ No unified marketing strategy
❌ Landing pages convert poorly
❌ Sales team finds leads manually

Three workflows, one mission

recraft: half: Professional at laptop with glowing workflow diagrams and connected nodes floating around

$$

1. Multi-agent marketing system → Complete marketing plan (parallel agents)
2. Conversion rate optimizer → Suggestions for improved landing page (sequential chain + feedback loop)
3. Lead scraper → Automated Google Maps prospecting (batch processing + API)

The parallel agent architecture

Parallel agent architecture

Campaign Brief: business context, objective, target audience, and budget
Brand, Content, and Paid Ads Agents: work independently to keep them focused
Merge Strategies: waits for parallel branches, and Combine Outputs combines them
CMO Summary Agent: produces an executive brief

Collecting and reshaping outputs

Combine and synthesize

Merge outputs 3 separate items downstream
Code node reshapes the list into one combined object with brand, content, and paid fields

Structuring the final output

$$

CMO Summary Agent reads the combined object and generates one executive brief
Structured Output Parser enforces a JSON schema on every run
JSON output can be chained into a dashboard, Slack, or another automation

$$

{
  "executive_brief": "90-day integrated strategy...",
  "priorities": [
    "Brand campaign",
    "A/B test pages",
    "Lead pipeline"
  ],
  "action_plan": {
    "day_30": "...",
    "day_60": "...",
    "day_90": "..."
  }
}

What you'll learn

$$

🔀 Use parallel branches to run independent AI agents from a single input
🛠 Enforce JSON schemas with Structured Output Parsers
🧰 Merge and reshape parallel outputs with a Code node
🤖 Synthesize specialist strategies into one structured brief

Let's practice!

Building Marketing Workflows with n8n

Preparing Video For Download...