Watch, 0:00 to 0:25

• Video link, n8n beginner intro: n8n Beginner Course, Introduction to Automation
• While watching, write 5 bullets on what n8n is used for.

Build, 0:25 to 1:35

• Create n8n account, or run self-hosted n8n.
• Create workflow W1S1 - Hello Workflow.
• Add nodes: Manual Trigger , Set.
• In Set node, set fields: message = Hello from n8n, createdAt = current time using expression {{new Date().toISOString()}}.
• Run workflow, inspect output, take a screenshot.

// Set node fields
message: "Hello from n8n"
createdAt: {{new Date().toISOString()}}

Watch, 0:00 to 0:25

• n8n Beginner Course, APIs and Webhooks
• Write definitions for API, endpoint, request, response, JSON.

Build, 0:25 to 1:35

• Create workflow W1S2 - Quote API.
• Add nodes: Manual Trigger , HTTP Request , Set.
• HTTP Request, Method GET, URL: https://api.quotable.io/random.
• Set node, output only these fields: quote, author, length, fetchedAt.
• Use expressions to map from HTTP output.

// Example Set node expressions
quote: {{$json["content"]}}
author: {{$json["author"]}}
length: {{$json["length"]}}
fetchedAt: {{new Date().toISOString()}}

Watch, 0:00 to 0:20

• n8n Beginner Course, What are nodes
• Skim n8n docs: Webhook node docs

Build, 0:20 to 1:40

• Create workflow W1S3 - Webhook Intake.
• Add nodes: Webhook , Set , Respond to Webhook.
• Webhook, Method GET, Path: /intake (or anything).
• Execute workflow, then open the test URL with query params: ?name=Sam&topic=n8n.
• Set node maps: name, topic, receivedAt, source = webhook.
• Respond to Webhook returns JSON { ok: true, name, topic }.

// Response JSON
{
  "ok": true,
  "name": "{{$json["name"]}}",
  "topic": "{{$json["topic"]}}"
}

Watch, 0:00 to 0:20

• n8n Beginner Course, How n8n handles data
• Write down what an item is, and what items[] means.

Build, 0:20 to 1:40

• Create workflow W1S4 - Multi Item Greetings.
• Manual Trigger , Function , Set.
• Function returns multiple items with names.
• Set node adds a greeting using expression.

// Function node
const names = ["Alice","Bob","Charlie","Dana"]; 
return names.map(n => ({ json: { name: n } }));

// Set node expression
message: {{`Hello, ${$json.name}!`}}
createdAt: {{new Date().toISOString()}}

Watch, 0:00 to 0:25

• n8n Beginner Course, Core workflow concepts
• Write 2 examples of routing logic, leads, support tickets.

Build, 0:25 to 1:35

• Create workflow W1S5 - Score Router.
• Manual Trigger , Set , If , Set (Pass) , Set (Fail).
• Set node sets score to 75.
• If checks score >= 60.
• Pass branch outputs result: pass, Fail branch outputs result: fail.

// Pass Set node
result: "pass"
score: {{$json.score}}

// Fail Set node
result: "fail"
score: {{$json.score}}

Watch, 0:00 to 0:25

• n8n Beginner Course, Useful nodes
• Write down 5 nodes that seem valuable, Cron, Wait, Merge, Split in Batches, Code.

Build, 0:25 to 1:35

• Create workflow W2S6 - Scheduled Quote + Time.
• Nodes: Cron , HTTP Request , Set (Time) , Merge , Set (Final).
• Cron, every 5 minutes for testing.
• HTTP, fetch quote from quotable API.
• Set (Time), create now fields: runAt, timezone.
• Merge, combine quote + time into one item.
• Final Set, return quote, author, runAt.

// Merge mode suggestion
Use "Merge By Position" if both branches output one item.

// Final output
quote: {{$json.content}}
author: {{$json.author}}
runAt: {{$json.runAt}}

Watch, 0:00 to 0:25

• n8n Beginner Course, Error handling
• Write a list of 5 things that can break in automations, credentials, API rate limits, bad input, timeouts, formatting.

Build, 0:25 to 1:35

• Create workflow W2S7 - Failure Sandbox.
• Manual Trigger , HTTP Request (bad URL) , Set (Fallback) , IF (error exists).
• HTTP Request URL: https://api.this-domain-should-not-exist-123.com.
• Enable Continue On Fail on HTTP node.
• If checks {{$json.error}} is not empty.
• True branch sets status: failed, includes errorMessage.
• False branch sets status: ok.

// If condition expression idea
{{$json.error !== undefined}}

// Fallback Set fields
status: "failed"
errorMessage: {{$json.error.message || "unknown"}}
failedAt: {{new Date().toISOString()}}

Watch, 0:00 to 0:20

• n8n Beginner Course, Debugging
• Write your own debugging checklist, data, config, credentials, outputs.

Build, 0:20 to 1:40

• Pick any previous workflow and introduce a bug on purpose, wrong field name or bad expression.
• Use node execution view to find the exact node where data changed.
• Use pin data, if available, to reuse output while adjusting downstream nodes.
• Create workflow W2S8 - Debug Notes and paste your debugging checklist into a Set node so it lives in n8n.

// Example Debug Checklist text to store
1) What is the trigger input
2) What does each node output
3) Are expressions referencing the right fields
4) Any credential errors
5) Any rate limit or timeout
6) Does final output match schema

Watch, 0:00 to 0:20

• n8n Beginner Course, Collaboration
• Write down what breaks when moving workflows across environments.

Build, 0:20 to 1:40

• Each learner chooses their best workflow from Week 1 or 2.
• Export workflow as JSON and swap.
• Import the other person’s workflow.
• Update any missing credentials and re-run to confirm it works.
• Create a standard README template in your notes doc, reuse it for every workflow.

# README template
Name:
Purpose:
Trigger:
Nodes:
Inputs:
Outputs:
Credentials needed:
How to test:
Failure modes:
Version notes:

Watch, 0:00 to 0:15

• Optional inspiration: Build your first workflow in n8n

Build, 0:15 to 1:45

• Create workflow W2S10 - Daily Quote Digest.
• Use Cron daily at a chosen time, or manual trigger for now.
• Fetch a quote via HTTP, format a message, then send to one output channel.
• Output options, Email node, Slack, or just a Respond to Webhook.
• Add basic error handling, if HTTP fails, send a different message.

// Message format example
subject: "Daily Quote"
body: "\"{{$json.content}}\"\n\n- {{$json.author}}\n\nSent: {{new Date().toISOString()}}"

Watch, 0:00 to 0:25

• OpenAI Node in n8n, Supercharge your automations
• Create an OpenAI API key if required: OpenAI platform

Build, 0:25 to 1:35

• Create workflow W3S11 - First AI Call.
• Manual Trigger , Set (Input) , OpenAI node , Set (Normalize).
• Set (Input) creates: question = a real question.
• OpenAI prompt references question and forces JSON response.
• Normalize Set extracts: answer, confidence, sourcesSuggested.

// Prompt template
You are a helpful assistant.
Return ONLY valid JSON with keys: answer, confidence (0-1), sourcesSuggested (array).
Question: {{$json.question}}

Watch, 0:00 to 0:20

• Inspiration: Use AI to find ideas with Reddit + n8n
• Skim: n8n AI docs: n8n Advanced AI

Build, 0:20 to 1:40

• Create workflow W3S12 - Support Triage.
• Webhook trigger receives message.
• OpenAI node returns JSON with category and priority.
• Switch node routes category, billing, tech, sales, other.
• Each route ends with a Respond to Webhook JSON acknowledging route.

// Classification prompt
Return ONLY JSON: {"category":"billing|tech|sales|other","priority":"low|medium|high","reason":"..."}
Message: {{$json.message}}

Watch, 0:00 to 0:15

• Automate content with AI, n8n + OpenAI

Build, 0:15 to 1:45

• Create workflow W3S13 - Text Summariser.
• Webhook receives text.
• OpenAI returns JSON with summary, keyPoints, actionItems.
• Respond to Webhook returns the JSON directly.

// Summariser prompt
Return ONLY JSON with keys: summary (max 60 words), keyPoints (array max 5), actionItems (array max 5).
Text: {{$json.text}}

Watch, 0:00 to 0:35

• Prompt Engineering for Beginners, AI Agents series
• Write 5 prompt rules, be specific, define format, give examples, constrain length, add refusal rules.

Build, 0:35 to 1:35

• Pick W3S12 or W3S13 and improve its prompt.
• Add a JSON schema example inside the prompt.
• Add a refusal rule, if the input is empty, return {"error":"empty input"}.
• Add a consistency rule, keep category values strict.

// Example schema snippet
Output example:
{"category":"tech","priority":"medium","reason":"password reset request"}
If message is empty, return: {"error":"empty input"}

Read, 0:00 to 0:20

• Best practices article: AI Workflow Builder best practices

Plan, 0:20 to 2:00

• Pick ONE capstone per learner. Options below.
• Create a blueprint in your notes doc with nodes, inputs, outputs, and test cases.
• List credentials you will need and create them now.
• Define a strict JSON output schema and a failure strategy.

// Capstone options
1) Daily digest, fetch RSS or API, summarise, email or slack
2) Lead qualifier, webhook intake, enrich, score, route
3) Support triage, classify, draft reply, send to review queue

Action plan, 0:00 to 2:00

• Create workflow named W4 Capstone - (your project name).
• Add Trigger, webhook or cron or email trigger, based on blueprint.
• Create Set node to normalise input schema, required fields only.
• Add logging step, use a Set node to create runId and startedAt.
• Implement first external call if needed, HTTP or integration node.
• Stop after the first major milestone runs end to end.

// Run id helper in a Code or Function node
const runId = `${Date.now()}-${Math.random().toString(16).slice(2)}`;
return [{ json: { ...$json, runId, startedAt: new Date().toISOString() } }];

Deliverable

• 1 workflow that triggers and produces a clean normalised JSON payload.
• README section, Purpose, Trigger, Input schema, How to test.

Action plan, 0:00 to 2:00

• Add OpenAI node for the main intelligence step.
• Force JSON only output and define schema explicitly.
• Add validation, if output missing required keys, set status failed and return fallback.
• Add routing with If or Switch based on AI output.
• Keep outputs small and predictable.

// Prompt skeleton
Return ONLY JSON matching this schema:
{"decision":"...","reason":"...","nextSteps":["..."]}
Input: {{JSON.stringify($json)}}

Deliverable

• AI step that outputs structured JSON reliably.
• At least 2 routes based on AI output.
• README, schema, 5 test cases with expected outputs.

Action plan, 0:00 to 2:00

• Add final output integration, email, slack, sheets, or webhook response.
• Format output for humans, keep a structured JSON version too.
• Add a summary text field that is easy to read.
• Run 10 tests, record outputs, fix formatting problems.

// Example human summary
summary: {{`Decision: ${$json.decision}\nReason: ${$json.reason}\nNext: ${($json.nextSteps||[]).join(", ")}`}}

Deliverable

• End-to-end workflow that produces a useful output externally.
• README updated with credentials and output format.

Action plan, 0:00 to 2:00

• Add error handling, continue on fail only when safe.
• Add an error route that returns a predictable JSON error object.
• Rate limit protection, use Wait or Split in Batches if calling AI multiple times.
• Cost guardrail, limit items processed per run.
• Add a logging output, store runId, status, and key fields to a sheet or database if possible.

// Error object format
{
  "ok": false,
  "runId": "{{$json.runId}}",
  "error": "...",
  "step": "openai",
  "timestamp": "{{new Date().toISOString()}}"
}

Deliverable

• Workflow survives bad inputs and external failures.
• README includes failure modes and how to recover.

Action plan, 0:00 to 2:00

• Prepare a 5 minute demo walk-through of your workflow.
• Show, trigger, data flow, AI step, output, error case.
• Export workflow JSON.
• Write a final README that another person could use to run it.
• Self assess using the final checklist below.

// Demo script outline
1) What problem this solves
2) Trigger and input schema
3) Key nodes and decisions
4) AI prompt and schema
5) Output and where it goes
6) Failure modes and safeguards

Deliverable

• Capstone workflow JSON export.
• README + test cases.
• 5 minute demo complete.