Highlights
- Sky Lubreski bought Scratch Golf, an e-commerce business that sells custom golf simulators, and inherited a dated, disorganized back office cobbled together from about 10 different SaaS tools.
- The disjointed system made it hard to manage the volume of customers coming in, plus the necessary admin every business runs on.
- Sky started building replacements for each tool on Bolt.new. Now Scratch Golf runs on his own system, the Command Center, and he's canceled nearly $3,000 a month in software and services.
- The proof landed in July: with the back office rebuilt and a new storefront live, Scratch Golf posted $96,000 in revenue, its best month since Sky and his partner bought the business.
When you buy a business, you inherit all its potential. And all its problems.
When Sky Lubreski and his business partner purchased ShopScratchGolf.com, they were thrilled with the wealth of customers interested in customized golf simulators. They were much less enthusiastic about the cobbled-together tech stack that made up the back office.
The business was running on about 10 tools: chat, email marketing, bookkeeping, calling software, a stack of plugins, and Shopify as the e-commerce layer. No workflow or information exchange existed between the tools. Everything had to be shared across each part of the system manually. It was inefficient, and it couldn't handle the volume of customers the website was pulling in.
Sky knew he had to streamline the back office to set his business up for future success. He looked at EDI providers, and one quoted him $12,000 up front and six months of build time just to connect his order flow to his vendors. Nothing was a great fit. He'd already been working on a web app for another project in Bolt.new, and he chose to build his own replacements rather than rent new ones.
He started working on a new system, eliminating his business's reliance on external tools one by one until he had a fully operational backend. Sky said:
"It probably took me maybe one or two months to get it to the point where we were actually using it, full stop. We eventually canceled every single subscription that we had. So Tidio, Quo, QuickBooks, Klaviyo, everything. We canceled it all. Now everything is in-house, in what I call the Command Center."
What does it cost to run a small e-commerce business on SaaS?
For Scratch Golf, as of 2026, the real invoices looked like this: about $390 a month across the core subscriptions, a Shopify bill of $105 that swelled past $500 a month after platform fees, and a Google Ads agency at $2,500 a month on top. Every function was siloed and invoiced independently.
Here's where each subscription went:
| Tool he rented | Monthly cost | What it did | Where it lives now |
|---|---|---|---|
| Tidio | $118 | Customer chat | Chat widget with live analytics, built into the Command Center |
| QuickBooks | $115 | Bookkeeping | A full ledger inside the same app |
| Quo | $69 | Phone and texting | Browser-based calls and texts on a $10 to $20 a month phone line |
| Klaviyo | $60 | Email marketing | Email flows running off the store's own database |
| Four Shopify plugins | $28 | One feature each | Features of the Command Center, not separate bills |
| Google Ads agency | $2,500 | Ad management | A marketing agent Sky is building into the Command Center |
| Shopify | $105 ($500+ with fees) | E-commerce layer | Checkout only, until his own payment handling ships |
Eliminating this messy tool stack didn't cut all operating costs, but those funds were reallocated to a more sustainable way of working that can support Scratch Golf as it scales.
His running costs now are the database (Supabase), Bolt.new, the phone line, and AI credits, plus the payment processing and workspace tools any store keeps. He and his partner reinvested the subscription fees in better tech. On top of that, Sky and his partner got back all the time they'd been spending on disjointed admin.
How do you replace a SaaS stack with one custom app?
One workflow at a time.
Sky didn't switch systems over a weekend. He ran the replacement like a project plan: describe a job to Bolt.new, build it, live with it for a while, then cancel the subscription it retired. Then the next one. Each cancellation made the next one feel less like a risk and more like routine.
Eventually his build evolved into a complete backend. The Command Center is the operational spine of Scratch Golf. The only external tool still running inside Sky's custom system is Shopify, woven in through the Shopify API. Otherwise, everything is a natively built function with visibility and interconnectivity across the whole system: vendor and margin tracking, calls and texts from the browser, every customer touch logged, the ledger, the email flows, and an AI answering customer questions straight from the product catalog.
It's a much easier business to manage, and it's a better experience for customers too.
"When we first engage with a customer via chat, then we email them a draft order. Then they place the order. Then we call them on the phone to confirm. All of that is logged sequentially in our custom CRM. Everything is visible and tied together," said Sky.
Sky is a mechanical engineer with a computer science minor, and he spent his years at Accenture as a consultant: writing requirements, testing systems, translating what a business needed into specs a developer could build from. "Not actually that great at coding itself," he said, "which almost primes me perfectly."
His experience was an ideal primer for building with AI. He knew how to translate complexity into practical systems that work the way they should, and that's exactly what he built.
Once the backend was sorted, Sky turned his attention to the storefront. He built a new website, ShopScratchGolf.com, launched it in June, and spent a few weeks shaking out the problems. In July, with the catalog fixed and the new storefront fully live, Scratch Golf did $96,000 in revenue, its best month yet by a wide margin. The full frontend experience is hardwired into the customer journey he'd already mapped out in the back office.
What does an AI-first small business look like?
The software industry is a massive, established goliath. Every SaaS product now has an AI assistant or the ability to build for-purpose agents, but that still isn't enough to keep up with the pace of AI building. No SaaS product built by a team of professional developers will ever really compete with Sky's Command Center, because that level of granular customization isn't possible at a mass-market scale.
The Command Center fits the needs of a niche industry. It supports the business model Sky and his partner envisioned. It's tailored to the expectations of their ideal customer. It works so well because Sky built it: it has everything his business needs and exactly nothing it doesn't. This is enterprise-grade workplace automation at small-business scale, designed by the person who knows the business better than any outside team ever could.
Scratch Golf is an AI-first small business. Its infrastructure was built on AI, and the Command Center itself runs agentic workflows to make the business faster and easier to run.
Every morning, an AI reads the database, the Google Ads account, the site analytics, live traffic, and competitor promotions, then writes up a whole-business report for Sky to review. The $2,500-a-month Google Ads agency is already gone. "Not only were they expensive, their results were not even close to worth their cost," Sky said. The marketing side of the Command Center runs the ads now. The step after that, the one he's building toward, is an agent that drafts orders straight from customer chats.
"If you have everything about your business right together, the AI can just do everything," Sky said. "I gave it full access to our Google Ads, our Search Console, Merchant Center, all that. More or less, if a person can do it, the agent can do it, as long as it has the tools and access to the right information. So that is the end game."
The pattern is bigger than one golf store. Among teams surveyed in late 2025, 35% had already replaced at least one SaaS tool with a custom build, and 78% expect to build more in 2026 (Retool, 2026).
Somewhere along the way, the Command Center stopped being just an internal tool. Sky onboarded a part-time employee who works in it every day. Friends who run their own businesses asked about it, so he spun up branched instances for them. Now he's weighing a launch as a product in its own right, and it already has an address: TheCommandCenter.pro. "Basically a Shopify competitor, but then some," as he put it.
What's hard about replacing your SaaS stack?
Three things, in Sky's telling.
- Code slop: the AI sometimes writes what he calls code slop. Asked to fix a broken function, it writes a new one that works backward instead.
- API guesswork: connecting outside services takes trial and error, because the AI will guess at an API endpoint rather than admit it doesn't know.
- Compounding complexity: a store that ran on 10 rented tools is now a real codebase. Sky either holds the whole map in his head, or plays whack-a-mole with bugs.
Bespoke software for the win
When Sky and his partner bought Scratch Golf, it ran on a tangled mess of disparate tools that bottlenecked customers. Sky started out wanting to fix a broken system and shed some subscriptions. He ended up building a modern AI-first operational backbone that's evolving into a product of its own. The proof landed in July: $96,000 in revenue, the store's best month since they bought it.
The one subscription still standing is the Shopify checkout, and even that isn't long for this world. Sky already has his own payment handling in the works. When that launches, the Command Center will be a complete, end-to-end e-commerce solution.
Sky blazed a trail for entrepreneurs who are tired of paying top dollar for bottom-of-the-barrel tech.
"Instead of having a one-size-fits-all, you can make it fit for you," Sky said. "Your one size."
Start building, free: Bolt.new
