Skylab · Build & maintain
An AI strategy is worthless without a working system behind it.
Skylab builds and maintains that system for the one-person businesses too small for the big firms and too important to trust to a freelancer — starting with a two-week Discovery Sprint, not a six-figure quote.
The problem — in your own words
Your whole business is in your head and one spreadsheet — and you’re terrified of both.
You’re brilliant at putting deals together. You’re drowning in the paperwork that comes after.
I spend all day on the phone making deals, then all night typing them into the sheet.
Every deal, I re-key it three times — into the sheet, into the confirmation, into QuickBooks.
I'm chasing signatures, grades, and trucks all day. That's not the job. The job is finding the match.
If I'm sick for a week, the business stops. Nobody else can run the sheet.
I probably leave deals on the table every week — I just can't get to them all.
One typo in the position sheet could cost me a whole margin.
Your income scales with deals closed. But most of your hours go to capture, confirmation, and chasing — admin that closes nothing. The spreadsheet is both a ceiling on your revenue and a single point of catastrophic failure. And nothing you’ve tried fixes it: more templates are still manual, an assistant just moves the knowledge into a second head, and enterprise software is built for Cargill’s desk and priced for Cargill’s budget.
Proof — the CUPE receipt
We’ve already done this — and it’s still running 10 years later.
Twenty years of MS Access and spreadsheets → a live web system, now maintained and extended by one person with AI.
- 20 yrs → 1Spreadsheets & MS Access → one live system
- 10 yrsStill maintained and running, a decade on
- 1 personBuilt, owned, and extended by one operator
CUPE BC ran their biggest operation on MS Access and spreadsheets for 20 years. We did the discovery and wrote the RFP — defined what actually needed building before a line of code existed. We guided it to a working, live web system that replaced the fragile stack. And 10 years on, we still own, maintain, and keep extending it. The client is still paying, because it still works.
The honest part: that build predates AI-native delivery — AI didn’t write it, and we’ll never pretend it did. What’s proven is the pattern: spreadsheets and one brain → a maintained production system run by one person. AI is what now makes that same pattern fast and cheap enough that a solo operator can finally afford it. CUPE took a team and a timeline you never could. AI collapses that into weeks and a price that fits.
The offer — three rungs, prices shown
One ladder. Each rung de-risks the next. No six-figure leap of faith.
Transparent, fixed pricing — because you shouldn’t have to book a call to find out what it costs.
Discovery Sprint
$5K–$10K
The front door. We map how your business actually runs and hand you a ranked list of where AI agents win, plus a concrete build plan for the #1 bleed. You walk away with a real operating map you own — whether or not you ever hire us for the build.
First Build
$25K–$60K
Your first working system, fixed price, fixed scope. For a broker that's capture → confirm → chase: agents that turn calls, emails, and texts into structured deals, draft the confirmations, and chase the signatures, grades, and trucks. Your old spreadsheet keeps running untouched until the new system has earned its place beside it. An order of magnitude below the $150K–$1M ERP world — because delivery is AI-native and lean.
Own & Extend
$1.5K–$4K
We host it, maintain it, and keep extending it — the next agent, the next workflow. Month-to-month. You own the code the whole way through; leave anytime and it's in your repo. Builds are lumpy; a maintained system compounds.
Your income scales with deals closed. Free up ~15 hours a week of admin and you close more. Priced against the deals you’re leaving on the table every week — not against overhead.
You own the code. Always.
Why Skylab
Everyone who can build won’t maintain it — or won’t touch a shop your size. Everyone your size who’ll show up can’t actually build.
Skylab is the one that discovers it, ships it, and stays.
Your real competitor isn’t another consultant. It’s doing nothing — the spreadsheet and your head. It already “works” — until the day it doesn’t, and then the business stops. That’s the risk no one else is pricing. Once you decide to move, here’s the whole field, and why every option but one dead-ends:
- Do nothing (spreadsheet + your head)It works, until it doesn't. Then everything stops.
- Hire an assistantNow you're a manager, and the knowledge is still trapped in a human head — just a second one you have to train.
- Off-the-shelf CRM / CTRM / ERPBuilt for Cargill's desk, priced for Cargill's budget — and you'd still do all the data entry. You bend your business to the tool.
- AI strategy consultantsThey hand you a deck and a bill. Nothing ships. You're where you started, minus the fee.
- AI automation / no-code agenciesThey'll wire up a bot that breaks in a month and can't build the real system underneath it.
- Big dev shop / SI$250K, nine months, and they won't take a deal your size seriously.
- Cheap freelancerBuilds you something and vanishes — no discovery, no maintenance, no one to call when it breaks.
Trust — the small bet that de-risks the big one
You never make a big bet. You make a small one that proves the next.
The three-rung ladder is the guarantee: $5K to prove we understand your business → a fixed-price build with your old system still running → month-to-month ownership you can leave anytime, code in your hands.
- “You're one person — what if you get hit by a bus?”
- You own the code, always — no lock-in, it lives in your repo. And the ladder proves us before you bet anything real. Our being small becomes your safety: you keep the asset no matter what.
- “My whole business is in that sheet. I can't have it break.”
- That's exactly why we start with a Discovery Sprint, not a rip-and-replace. Your old sheet runs untouched until the new system has earned its place beside it. Parallel run, never big-bang.
- “I've been burned — paid for something that never shipped, or died after launch.”
- The maintenance retainer is the whole model, not an upsell. CUPE's system still runs 10 years later, and the client's still paying — because it still works.
- “How do I know you can even build this?”
- CUPE is the receipt — and it's not just one client or one industry. Selected work includes BCAA and Nike. Twenty years of shipping, plus a UBC CS/AI background.
- “That's real money for me.”
- Priced against the deals you're leaving on the table every week, not against overhead. And Rung 1 is a $5K–$10K test, not a leap.
- “What if we're not a fit?”
- Then the Discovery Sprint pays for itself anyway — you walk away with a real operating map you own, whether or not you continue. That's the risk reversal: standalone value in the deliverable itself.
Selected workCUPE · BCAA · Nike— not just one client, not just one industry.
Insights
What we’re thinking about.
Notes on getting your business out of your head and into a system that runs while you sleep — from the LinkedIn feed and the Skylab journal.
AI Didn't Replace Search. It Just Hid It.
A friend of mine was chatting with ChatGPT last week and came away unimpressed. He'd asked about Mark Carney, and the model went on about Carney running for office.
How about an AI concierge for shopping online? What do you think of this…
I went looking for a pair of running shorts last week. Size medium, light enough for a July heatwave, with a pocket that actually holds a phone.
A chatbot waits. An assistant runs.
A few months ago I watched a PhD student on YouTube walk through his personal AI setup. What stuck with me was not the model he ran. It was that his assistant did things while he slept.
Start with a small bet
Book a Discovery Sprint. Walk away with a map you own.
Fixed price, one to two weeks. We map how your business actually runs and hand you a ranked plan for where AI wins first — whether or not you ever hire us for the build. No deck, no sales sequence.
Prefer email? hello@skylabconsulting.com
Questions
Straight answers.
- How long until something actually works?
- The Discovery Sprint is 1–2 weeks. Your first working build ships in weeks, not the quarters an ERP rollout takes — because delivery is AI-native and lean.
- Do I own what you build?
- Always. The code lives in your repository from day one. Month-to-month maintenance means you can leave anytime and walk away with the whole system.
- What if I already have a spreadsheet that works?
- Good — we keep it running. We never rip-and-replace. The new system runs in parallel until it has earned its place beside the old one.
- What do you need from me to start?
- A conversation. The Discovery Sprint maps how your business actually runs — you don't need to prepare anything or know what to build. That's our job.