Jack couldn't code.
He had a spreadsheet. A colorful P&L tracker he'd built for his agency clients, with big numbers and conditional formatting.
He downloaded it as a CSV, uploaded it to Cursor, and typed "build it".
Five months later that spreadsheet is Profit AI, a Shopify app doing $30K MRR, plus almost $40K MRR in services on top. $147,000 collected since launch. Launched in December (Shopify approved it on Christmas Day), and by May it was paying him more than most senior dev salaries.
Pat Walls had him on Starter Story to break down exactly how he did it, and honestly, this is one of those episodes where the playbook is sitting right there in the open.
Here's the full story, with every number and every step.
What Profit AI actually is (and the real numbers)
Profit AI is a profit analytics and marketing automation app for Shopify stores.
The pitch is simple: see your profit in seconds. Ad spend per channel on the left, attributed sales in the middle, and true profit after every cost (including marketing) on the right. Normally that's hours of spreadsheet work. Now it's a dashboard.
Enterprise stores also get forecasting: the relationship between ad spend and revenue per platform. Plus an AI layer to launch ads, send email, generate a P&L, and deploy straight from Slack.
The numbers Jack shared on camera:
- → $147,000 total revenue since the December launch
- → $30K MRR from the app, just under $40K MRR from services
- → 122 users of the app, around 38 or 39 paying
- → 47 installs vs 43 uninstalls (yes, churn is rough, more on that below)
- → Zero paid marketing so far. All word of mouth. The "hard launch" hasn't happened yet.
And the pricing story is my favorite part.
The paid plan started at $5,000 a month. He got a couple of takers, all businesses he already worked with. So he dropped it to $800 a month. He also kept a free plan, against all the usual advice, because he wants everyone in the app: sign up free, get a 30-day trial of the full paid product.
$800/month with 38 paying customers. Do the math. That's the whole business.
The backstory: broke student to agency operator
Jack didn't start in SaaS.
In 2016 he launched an e-commerce brand selling software and sample packs to music producers (dubstep guys). Software margins, fast growth. He went from broke university student to "more money than I knew what to do with".
The growth engine was Facebook ads. Put $1,000 in, it just worked.
So he started consulting. First for mom-and-pop agencies, then bigger ones. His whole ethos: good decision-making comes from information. And his delivery mechanism was spreadsheets. Accessible P&Ls, LTV calculators, a ROAS calculator, an ad analyzer, a scenario planner.
The kind of scenario planner Deloitte charges tens of thousands for, except his was readable by a normal founder.
Here's the part I want you to sit with: he ran that spreadsheet business for years before the app existed. The product was validated every single week by real clients paying for the output. He just hadn't turned it into software yet.
The spreadsheet-to-app moment
Jack tried the traditional route first. He asked a developer friend (the guy who built Nookazon, the Animal Crossing marketplace) to build it.
The guy basically refused.
Then AI coding got good, and Jack got stuck in. His literal process:
- Download the spreadsheet as a CSV
- Upload it to Cursor
- Say "build it"
- Find everything the AI got wrong, and fix it one calculation at a time: "We need to fix this. This is the right calculation here."
That took him the best part of 3.5 to 4 months. Hundreds of hours in the terminal, as a guy who could not code when he started.
Pat's reaction on the show says it all: he took his own personal finance spreadsheet, dropped it into Claude Code, and had a working app in basically a day. That's how much the tooling has moved since Jack started.
The window Jack squeezed through over four months is now open wide enough to walk through in a weekend. I wrote a full guide on this in how to build apps without coding if you want the step-by-step.
The current stack (it's cheaper than your gym membership)
What runs a $30K MRR app in 2026:
- Claude Code on the max plan, about $200/month. The main dev tool.
- Codex in the terminal. Was $200/month, then they gave him a 100% discount, so currently free.
- Framer for the website (he might replace it with an AI-built site).
- Instantly for outbound.
- Google Sheets. Free. Still in the stack, because the spreadsheet never really dies.
Call it $250 a month of tooling to run a product doing $30K MRR.
No dev team, and no CTO co-founder taking half the company.
Why e-commerce tooling is the underrated goldmine
Pat dropped two data points in this episode that deserve their own article.
First: Eric Kos, a previous Starter Story guest with a Shopify plugin. A year ago he was doing $200K/month. Pat caught up with him recently. He just crossed $1M a month.
Second: an upcoming guest who does Klaviyo reporting. One report type, for one tool (Klaviyo), and it makes him $1M a year.
The pattern: Shopify brand owners drown in data and will pay real money for anyone who turns that data into decisions. Jack sees it daily. He just closed a large Miami agency as a customer, and their requests are endless: "send this data there, in this format, for this meeting." His Slack bot (which is just Claude with pipelines) handles it.
While everyone fights over AI wrapper ideas and creator tools, e-commerce tooling sits there quietly printing money. Less sexy, better customers, actual budgets.
The playbook you can steal this week
Strip the story down and the steps are almost embarrassing:
Step 1: Find the spreadsheet you already maintain. Not an idea you brainstormed. A thing you (or your clients) already use every week. Jack's app existed as a spreadsheet for years before it was software. The demand was pre-validated.
Step 2: Feed it to an AI coding tool. CSV into Claude Code or Cursor. "Build it." You'll get something 70% right.
Step 3: Fix the calculations one by one. This is the actual work. You're not writing code, you're supplying domain knowledge the AI doesn't have. This is also your moat: anyone can prompt "build a profit tracker", almost nobody knows which numbers are wrong.
Step 4: Price high, then correct. Jack started at $5,000/month and walked it down to $800. Most founders do the opposite (launch at $9, spend two years finding the courage to charge more). Starting high cost him nothing and taught him exactly where the market was.
Step 5: Let your existing network be the launch. His first users were brands he already knew, beta testing free. Word of mouth carried it to $30K MRR before any marketing push.
If you need idea inspiration beyond your own spreadsheets, I keep a running list in micro-SaaS ideas.
The honest caveats
I'm not going to pretend this is all upside.
Jack's churn is scary: 47 installs, 43 uninstalls. He's churning customers almost as fast as he acquires them. At $800/month price points, every churned logo hurts. He knows it, Pat called it out on camera, and the retention work is still ahead of him.
And roughly $40K of his MRR is services, not software. That's not a flaw (services revenue funded the whole thing), but if you're picturing pure passive SaaS income, that's not this story. Which matches his own advice, actually.
Asked what he'd tell someone moving from trading hours for dollars to software, he said: "Don't aim for laziness. If your goal is 'I want to make lots of money while doing nothing', you're going to get quickly replaced."
Then the tactical version: "Just ask Claude if it can do the thing." Connected to your tools, one prompt goes shockingly far.
FAQ
How much money does Profit AI make?
Profit AI has made $147,000 in total since launching in December, and currently does $30K MRR from the app plus just under $40K MRR from related services. It has around 122 users, with 38 or 39 on the $800/month paid plan.
Can you really build a SaaS without knowing how to code?
Jack did. He uploaded his client spreadsheet as a CSV to Cursor, prompted "build it", then spent 3.5 to 4 months correcting the calculations. Today he ships features with Claude Code (about $200/month) and Codex. The build takes real hours, but zero prior coding ability.
How do I find a spreadsheet worth turning into an app?
Look at work you already do repeatedly. Jack's P&L tracker was used by paying agency clients for years, so demand was proven before a line of code existed. Audit your last 20 spreadsheets (Pat suggests literally pointing Claude Code at your Google Drive) and pick the one other people keep asking you for.
Why is e-commerce tooling a good niche for micro-SaaS?
Shopify brand owners have real budgets and drowning-in-data problems. From the same episode: one Shopify plugin founder went from $200K/month to $1M/month in about a year, and another founder makes $1M a year on Klaviyo reporting alone. Narrow data and reporting tools keep working here.
How long did it take to go from spreadsheet to $30K MRR?
About 3.5 to 4 months to build the app, launch in December, then roughly five months from launch to $30K MRR, with no paid marketing. Growth came from brands Jack already knew and word of mouth.
Stories like Jack's are exactly what I dig into every week on the Profitable Founder Podcast: bootstrapped founders doing $100K to $10M a year, walking through the real numbers and the real decisions behind them.
