Skip to content
Kevin Burke
Go back

Late nights, purchase orders, and how BoutiqueOS started

After a full day at my day job, I was back at our kitchen table copying wholesale orders into our boutique’s point-of-sale system. The information was already in the PDF in front of me. I was just moving it, one line at a time.

When my wife and I bought our Monkee’s franchise, we thought we knew what we were signing up for: buy great product, sell it, take care of customers. We were half right. Selling is the fun part. What nobody tells you is how much happens behind the counter before anything reaches the sales floor.

The work nobody sees

Every season, a boutique places dozens of wholesale orders with brands through platforms like JOOR, NuORDER, and Faire. Each order comes back as a PDF or a spreadsheet. Every line on every order (style, color, size, cost, retail price) has to get into the point-of-sale system before the boxes show up. For us, that system is Heartland Retail.

In our setup, there wasn’t a practical way to move those orders into Heartland automatically. So it was me, reading a PDF on one screen and typing it into Heartland on the other. Line by line. Order after order.

Then the product arrives and the next round starts: photos, descriptions, getting it online. More switching between systems. More copying the same information into a different box.

None of it is hard. It’s just slow, repetitive, and easy to get wrong. And every hour spent on it is an hour not spent on what actually grows a boutique: buying the right inventory, working the sales floor, and taking care of customers.

The first tool

After enough late nights, I did what an operations person does when a process is broken. I tried to fix it.

I’m not a career software developer. But AI had gotten good enough that I could describe the problem, work through it step by step, and build something real. The first tool was simple: take the order file a brand sends, read it, and turn it into something Heartland can import. What used to take an evening took a few minutes.

That changed how I looked at everything else in the business.

The pattern

Once order import worked, I started seeing the same shape everywhere:

If a task fit that pattern, I could usually build something to handle it. Turning vendor photos and copy into product descriptions and getting them onto our Shopify store. Pulling sales and inventory into one place so we could see which brands were actually earning their spot on the floor. Each one started as a fix for our own store.

Over time those fixes added up to BoutiqueOS. Today it helps independent boutiques running Heartland do the same thing. It has imported more than 6,000 wholesale orders, and the boutiques using it get back dozens of hours a month.

What it taught me

BoutiqueOS isn’t really the point of this post. The point is what building it taught me.

The biggest wins from AI in a small business usually aren’t flashy. They’re the boring, manual jobs sitting between your systems: the copying, the reconciling, the retyping. Those jobs eat hours every week, and most owners assume that’s just part of running a business.

It doesn’t have to be. Sometimes the problem isn’t that your business needs better software. It’s that the software you already use doesn’t work together. Fix that, and you get your time back for the work only you can do.


Share this post:

Next Post
Why I built my own AI agent instead of just asking ChatGPT