Building an AI-Powered Lead Gen System with n8n

Building an AI-Powered Lead Gen System with n8n

Spent 2 hours today learning n8n by building an automated lead qualification workflow. Some hard-won lessons:

Technical:

• Output formats matter: Schema vs. Table vs. JSON views. Amazing how much this overlaps with IA work

• Expressions don't always evaluate in JSON context (learned this the hard way - hence the 2 hours)

• Code nodes are your escape hatch when built-in nodes hit limitations. Wish we didn’t have to get there, but we did.

• this.helpers.httpRequest() is how you call APIs from Code nodes

Strategic:

• Chose n8n over easier tools to maximize learning (no regrets)

• Built end-to-end first, optimization later

• Cost control: 10 articles/day = $3-15/month in Claude API calls

The Win: Got Claude API successfully qualifying leads automatically. The workflow runs daily, analyzes funding news, and scores companies 0-10 against an ICP.

The Reality: Took 2 hours to get what "should" have been a 20-minute setup. But I understand exactly how every piece works now.

Next session: Filtering, Google Sheets, and email digests.

#AIAutomation #n8n #LearningInPublic #BuildingInPublic

Rock Paper Scissors - Shoot in Javascript!

const rps = (p1, p2) => {
  // Check for draw
  // Check if p1 wins
  // Else p2 wins
  if (p1 === p2) {
    return "Draw!";
  }
  else if ((p1 === 'scissors' && p2 === 'paper') || (p1 === 'paper' && p2 === 'rock') || (p1 === 'rock' && p2 === 'scissors'))
  {
    return "Player 1 won!";
  }
  else {
    return "Player 2 won!";
  }
};

Counting sheep outloud in Javascript

Recently, I’ve decided that I need some ******** skillz so now I’ve gone through a process of doing some Kata right before bed time.

This one is for counting sheep. This is for my own development and journal so if you’re going at it own your own, BIG SPOILER ALERT below

When given a function(3), return a string of “1 sheep…2 sheep…3 sheep…”

var countSheep = function (num){

  //start with a counter i = 1
  // increment i
  // stop until i == num
  // "interpolation" to insert variables
  // return joint array
    let smallArr = "";
    for (let x = 1; x <= num; x++)
    {
       smallArr += `${x} sheep...`
    } 
    return smallArr;
};

Iron Mask Prototype - Face tracking filer in SparkAR

Imagine the man in the iron mask is stuck in a dungeon under water. This is what it would look like. Dark thoughts, I know but was the ‘mood’ at the time.

Created using SparkAR.

Assets

• Blarney Castle

• Black glowing lotus

Learnings

• SparkAR is actually very intuitive to use. The Patch Editor needs a lot of UX work to make useful but the currently state is logical enough that after a new trials and errors, it’s possible to create (o stumble) upon some cool effects

• Next steps: Create scaled animation and add materials to bubbles