Building a $30k/mo portfolio within eight months of quitting his $420k/yr job
Jonathan Chan quit his $420k/yr job to build a brand and an AI-native consulting platform. Eight months later, AI Never Sleeps is generating nearly $20k/mo, and Pear is at $10k/mo.
Here's Jonathan on how he's doing it. 👇
Leaving a $420k salary
I spent 25+ years in corporate — The Boston Consulting Group (BCG), Westpac, fintech startups. I managed teams from two to 200+ people. And left a $420k annual salary about eight months ago to build with AI full-time.
Now, I run two things:
AI Never Sleeps — my personal brand and education business. I teach non-technical professionals how to build products with AI. No coding bootcamps. No CS degrees. Just AI tools and the structured thinking they already have. I run a daily LinkedIn post, a Substack newsletter, a YouTube channel, a paid Skool community, and a flagship 8-week course called Zero to Builder.
Pear — I'm cofounder, CTO, and CPO of an AI-native management consulting platform. Think "Uber for management consulting." We connect companies with top-tier consultants without the brand tax. Three co-founders, all ex-McKinsey/BCG/Bain. We launched in February 2026.
AI Never Sleeps is generating 5-figure monthly revenue across course sales, the paid community, and selective agency work — $19.2k USD in January 2026. Our first Zero to Builder cohort sold out, and our second one launches on March 19. The course alone costs $2,999 per seat.
Pear is at $100k/mo USD three weeks after launch and we take a 10% margin.
Boredom and fear
I left that $420k/yr job because I was bored. And a little scared.
I'd been promoted 15 times in 20 years. I was good at climbing. But I'd stopped building things. I'd become the guy who reviewed other people's work instead of creating my own.
Then AI tools got good. Really good. I'd coded earlier in my career — ASP.NET, some JavaScript — but everything had evolved while I was in management meetings.
AI was the bridge back. I opened Cursor (now Claude Code, mostly), described what I wanted, and it just... worked.
I shipped my first product in a weekend. Then another. Then I realized: If I can do this, there are millions of professionals who can too. They just don't know it yet.
That's when I quit. Not because I had it all figured out. Because I couldn't stop building.
Building the products
For AI Never Sleeps, the "product" was content first. I posted on LinkedIn every day for months before I built anything to sell. Tested what resonated. Found my voice. Built an audience.
The first real product was a free webinar series — live builds where I created automations and tools on camera. No slides. Just building. This proved that people wanted to learn this stuff.
Then, I built Zero to Builder. I took everything I'd learned from consulting (frameworks, structured thinking, breaking down problems) and applied it to teaching people to build with AI.
For Pear, we built the AI matching engine from scratch. Three cofounders with 30+ years of combined consulting experience meant we understood the problem from every angle — the buyer, the seller, and the consultant.
The stack
Here's my stack.
For building: Claude Code and Cursor are my daily drivers. I describe what I want in plain English and iterate from there.
For infrastructure: I love Vercel (frontend e.g., Next.js and React) and Railway (backend e.g., FastAPIs, Postgres databases, Redis queues, etc.)
For automation: n8n (self-hosted) for workflow automation. I've built 100+ automations with it.
For the course: Skool for community and classroom. Reveal.js for presentation slides. Substack for the newsletter.
For Pear: AI-native architecture built from the ground up — can't share too much yet, but the matching engine is the core IP. Think Postgres for structured data, RAG for semantic search, and a graph DB for node connections.
For content: I built an AI co-writing system using Claude Code with structured context profiles (JSON files capturing my voice, audience, and methodology), reusable skill files, and a knowledge base. It's like having a co-writer who already knows how I sound.
A subscription+ model
Here's the model for AI Never Sleeps:
Free tier: LinkedIn posts (daily), free Substack newsletter, free webinars. This is top of funnel. Build trust, demonstrate expertise, capture emails.
$59/month: Paid newsletter + Skool community. Weekly webinars, templates, direct access. This converts followers into customers. Small purchase → big purchase.
$2,999: Zero to Builder. 8-week cohort course. This is the flagship. Non-technical professionals learn to build real products with AI.
Custom pricing ($5000+): Selective agency work. Done-for-you AI automation for businesses.
We grew revenue through the funnel. Content → trust → small purchase → transformation offer. Customers are better than leads. Someone who pays $59/month is 10x more likely to buy the course than someone on a free list.
For Pear, we take a 10% margin for platform usage.
Growth via LinkedIn
As far as growth: LinkedIn. Full stop. That's where my audience lives.
I post every single day. Seven days a week. My highest-performing post hit 2.4 million+ impressions, 22,000+ likes, and 398 comments. The formula is simple: personal transformation stories + specific numbers + interactive CTAs.
"6 months ago, I couldn't code. Now I ship products weekly with AI." That post changed everything.
What works: vulnerability, specific numbers ($420K salary, 100+ automations, 47 retakes on a video), behind-the-scenes content, and keyword CTAs ("Comment BUILD and I'll send you...").
What doesn't work: pure promo without story, generic updates, and text-only posts without images.
I also run free webinars monthly. Live builds, no slides. People see me build something from scratch in real-time. That converts better than any sales page.
Identity shifts and the builder's trap
The biggest challenge was the identity shift. Going from "executive who reviews work" to "builder who ships products" is harder than any technical problem. Your brain keeps telling you to plan more, build a strategy deck, get buy-in. But nobody's giving you buy-in anymore. You just have to ship.
Another challenge was the builder's trap. I love building. Hate selling. I'd build something cool, get the dopamine hit, then move on to the next thing instead of selling what I just made. Recognizing that pattern was critical.
If I started over? I'd start posting on LinkedIn (pick the platform where your audience lives), start a newsletter and email list from day one, and focus — I covered too many different topics initially. I spent months building an audience on LinkedIn before I had anywhere to send them.
And this is important: Email is the only platform you own. Everything else is rented.
I'd also charge earlier. I gave away too much for free in the beginning. Not because free content is bad — it's essential — but because charging forces you to create something people actually value.
It doesn't have to be good — it has to exist
Here's my advice:
Ship something this week. Not next month. This week. It doesn't have to be good. It has to exist. The gap between "I have an idea" and "I shipped something" is where 90% of people get stuck forever.
Solve boring problems. When your brain says "that's boring," it usually means people are already paying for solutions, the market is validated, and you can make money. When your brain says "that's exciting and novel," it usually means no market exists yet.
Build an email list from day one. LinkedIn can change its algorithm tomorrow. Twitter can suspend your account. Your email list is yours.
Stop learning, start building. Tutorials are procrastination disguised as productivity. You'll learn more shipping one broken product than watching 100 hours of YouTube.
You're closer than you think. AI tools have compressed the timeline from idea to product dramatically. The thing that used to take a developer three months can now take a clear thinker three days.
What's next?
In the short term, my plan is to scale Zero to Builder. Cohort 2 starts 19 March 2026. I want to help hundreds of non-technical professionals ship their first product this year.
Medium term: Build AI Never Sleeps into the definitive resource for professionals building with AI. Not the best content about AI news or trends — the best content about building things.
Long term: Pear. We're building something that can genuinely disrupt a 100-year-old industry. Management consulting is ripe for reinvention, and we have the experience from every side of the table to do it right.
Long-long term: I have a number of projects cooking that I am excited about, so follow my journey!
The thread that connects all of it: helping people build. Whether it's building products with AI, building businesses, or building the next generation of consulting, the skill is the same. Clear thinking + the right tools + the courage to ship.
Leave a Comment
"Stop learning, start building" is the one. We spent weeks watching tutorials before realizing the fastest way to learn a stack is to build something real with it. Built an entire ecommerce platform and CRM in 30 days because there was a real business waiting on it. Deadlines beat tutorials every time.
great story - the AI agent angle is what resonates most for me. one thing i've been building around is the email OTP problem for agents. when they try to sign up for services or log in, they hit email verification and have no inbox. built a tool called AgentMailr to fix that - gives each agent its own persistent inbox with OTP polling + bulk/marketing email sending. shipping something people actually need is exactly what you're talking about here.
This really stood out to me — especially the part about identity shift from reviewing work to building again.
I’m currently building Gnobu, an identity-powered infrastructure prototype, and one thing I’m learning is exactly what you said: shipping something early changes everything. Once something exists, the conversations become real.
Curious — when you started posting daily on LinkedIn, how long did it take before you saw the first real signal that the audience was forming?