I Quit My Job After Learning One Skill in 47 Days

Day 1: I had a job I did not hate. Day 47: I handed in my notice. Day 90: I earned more that month than in any single month at that job.

The skill I learned was AI automation — specifically building no-code workflows using Make.com and ChatGPT for small businesses.

I want to be careful with how I tell this story because the internet is full of "I quit my job" posts that are really just advertisements for a £997 course. This is not that. I will tell you exactly what I did, what it cost, how long it took, and what I wish I had known before I started.

Why I chose this particular skill

I spent about a week researching before committing to anything. I was looking for something that met three specific criteria:

1. I could learn it from free resources — I was not willing to pay for a course until I had seen real proof it worked
2. There was obvious, immediate demand from paying clients
3. It did not require existing credentials or a portfolio to get started

AI automation hit all three. Small businesses were — and still are — paying $500–$2,000 for basic workflow builds that take a few hours to set up. The tools (Make.com, Zapier, ChatGPT) all have free tiers and extensive free documentation. You do not need to be a developer. You need to understand business problems and know which tool solves them.

The 47 days, broken down

Days 1–14: Learning

I watched free YouTube tutorials on Make.com for about 1–2 hours every evening after work. By day 10 I could build a basic workflow connecting a Google Form to an email. By day 14 I had built something more complex: a lead intake system that read form submissions, scored them against simple criteria, and sent personalised responses.

I paid for nothing. Make.com's free tier was enough for learning. The only cost was the evenings.

Days 15–25: Building demos

I picked one industry to target: local solicitors. I had a family member in that profession and had heard enough complaints about admin work to know real problems existed.

I built two demonstration systems without any client:

• A client intake chatbot that collects initial case details and books a consultation
• An email triage system that reads incoming enquiries, categorises them by practice area, and routes them to the right solicitor

Neither was for a real client. I recorded short screen recordings of both, showing exactly what problem each solved and how much time it saved, and I wrote a simple one-page explanation for each.

Days 26–35: First outreach

I contacted 20 small law firms. My message was specific: "I noticed your contact form leads to a generic inbox. I have built a system that reads those enquiries, categorises them by practice area, routes them to the right person, and sends an immediate personalised acknowledgement. I have a working demo I can show you in 10 minutes."

Eight replied. Four took a call. One became my first client.

Days 36–47: First paid project

The client wanted the email triage system — the exact demo I had already built. We agreed on £950 for the build and £250/month for maintenance.

I spent about 12 hours building the real version, customised to their actual email structure and team layout. It worked. They were happy. I asked for a written testimonial immediately.

That testimonial became the foundation of everything that came next.

Day 47: I handed in my notice.

Was it reckless?

Slightly, yes. Here is the honest math at the point I resigned:

• One client paying £250/month recurring
• One strong lead who had verbally committed to a £700 project but had not signed yet
• Three months of personal savings as a runway

That is not a lot of certainty. What gave me confidence was the speed at which the first client appeared — fewer than 10 outreach messages to one paying client. The demand was real and I had seen it with my own eyes.

If I were advising someone else: do not quit until you have £2,000–£3,000 in confirmed monthly recurring income. I quit earlier than that and got lucky with the pipeline timing.

Month 3 income breakdown

By month three after leaving:

• Two retainer clients: £500/month combined
• New automation build for a restaurant booking system: £1,400
• New build for a property agency lead qualification system: £1,800
• Referral from first client for a second law firm: £500 discovery call, build pending

Total: £4,200 in one month. More than my old salary.

What the skill actually involves

I keep saying "AI automation" but let me be precise.

The skill is: understanding how a business currently does something manually, translating that into a series of logical steps, and connecting tools to perform those steps automatically.

You do not need to code. You need to ask the right questions about a process. When a customer fills in your contact form, what happens next? Who sees it? What do they do with it? How long does that take? What would happen if you could not do that today?

Then you translate those answers into a Make.com workflow.

That is it. That is the whole skill.

It takes about two weeks to learn at a basic level. About two months to be genuinely confident with it. The demand is real, the pricing is solid, and the barriers to entry are low enough that someone with no technical background can have paying clients within 45 days.

The 47 days were real. The income was real.

The question is not whether you could do the same thing. It is whether you are willing to spend 1–2 hours per evening for six weeks to find out.

The one thing that will stop you before you start

Most people research for three weeks, decide they need to know more before approaching clients, research for another two weeks, and never actually message anyone.

The skill is learned fastest when a real client brings you a real problem. The first client I landed did not care that I had only been learning for five weeks. They cared that I had a working demo of exactly what they needed.

Build demos before you have clients. Then go get clients before you feel ready.

The 47 days started on the day I decided to send the first message, not the day I first opened a Make.com tutorial.