In ‘79 I built my own computer — a soldering iron and a box of parts, two years before Microsoft had an operating system. Today I’m 73 and my entire staff is AI. My receptionist, my marketing team, my executive assistant: none of them are human, and none of them draw a salary. My goal is to take this company to seven figures — myself and my AI team.
The score right now is zero. This is episode one.
This series is the whole climb, documented as it happens. I’m going to share the wins, the numbers when I have them to share, and the mistakes — especially the mistakes. In part, you could say this is a tutorial on how not to build a business, because I’ve made my share. Nothing gets cleaned up in the retelling. If AI employees can carry a real business to seven figures, you’ll see exactly how. If they can’t, you’ll see that too.
Fifty years, briefly
I spent ten years in the Air Force, 1971 to 1981, working on computers — six years on ground-based systems, then the airborne computers of the F-15 right when it came into the inventory. That’s where I was in ‘79 when I ordered my Heathkit H89. It came as a kit; you built it yourself. Two 8-bit processors, eight KB of memory, no hard drive, and it ran CP/M — the operating system Microsoft DOS would be patterned on, two years before DOS existed.
People called machines like that toys. The businesses that took them seriously anyway owned the next twenty years. Remember that — it’s the whole reason this company exists.
After the service I spent eight years as an air traffic controller (and picked up an Amiga along the way — the machine that was ten years ahead of its time; some habits don’t quit). Then I quit the FAA and started a company remanufacturing printer cartridges, back when the LaserJet was new. That turned into fixing copiers, then selling them, and eventually a full-line Toshiba dealership — which I built, and sold. Then I crossed to the other side of the industry: I ran three service organizations myself, then spent ten years at Sharp and eight at Konica Minolta supporting the service departments of dealerships. Twenty years of watching hundreds of dealer owners wrestle with the same two problems I’d had: finding good people, and affording them.
The employee day camp
The lesson started the day I hired my first employee. I thought I was building a copier business. What I’d actually signed up for was scheduling, sick days, drama, turnover, refereeing — basically an employee day camp. A lot of my people were good people. But between the payroll taxes and the babysitting, my business had turned into something I didn’t want, and the expense took most of the fun out of owning the thing I’d built.
My wife and I had started that dealership with just the two of us, no industry experience, and about $20,000 in the bank. (That was the wrong way to do it.) One thing that hurt us most as we grew: we could never afford marketing, so the only marketing that got done was whatever we did ourselves. And we had no one to bounce ideas off — we’d talk something over, decide it sounded good, and go do it. That led to a lot of the mistakes.
Pretty much every dealer I met over the next twenty years was living some version of that same story.
2019: the second soldering-iron moment
Back in 2019, before the pandemic and right after I retired — and I use the word “retired” very loosely — I got my hands on something called GPT-2 and started hunting for uses for it. I’ll be straight with you: I knew it was cool, but I didn’t have a clue where it was headed or how fast things would change. Nobody did.
But I’d felt that exact feeling once before, in 1979, holding a soldering iron.
When AI got good enough to hold a role — not help with tasks, but own a job: answer the phone every single time, do the research before every call, run marketing every week — I knew precisely what I was looking at, because I’d spent fifty years waiting for it. Staff without a day camp. Employees without payroll. The machine I built in ‘79 had eight kilobytes of memory. Any one of my AI employees uses more than that to say hello.
The year I spent building (and rebuilding)
About a year and a half ago I went all in — and for those who don’t know me, all in means seven days a week, as many hours as I could sit in front of the computer.
I spent hundreds of hours building and rebuilding. I got the original security wrong, because I didn’t know what I wanted the end result to be — I teach “start with the end in mind” in my management class, and I didn’t do it. So I went back and started over: a new company, a marketing system built to scale, and the security I was actually going to need. Part of what I built is essentially a global platform that can provide virtual AI employees anywhere in the world, in any language, meeting local security needs.
And I wanted my virtual employees to sound like and behave like humans. I hate IVR. I hate telling the same story twice — if you tell somebody a story, that story should persist. So it wasn’t two or three versions of Suzy Q, my voice agent. It was probably scores of versions. My wife hates this analogy, but Edison was said to have taken nearly nine hundred attempts to build a working light bulb — and when asked how it felt to fail that many times, he said he hadn’t failed; he’d learned things that didn’t work. I can tell you a lot of things that don’t work now.
The team that runs my back office
Where I am today: I have a team that is my entire back office. Suzy Q answers my phone line. My Sales Prep agent researches every prospect before I meet them and hands me a briefing. The Agency does my marketing — a 30-day content calendar and suggested posts, the thing I could never afford at my dealership. And my Board of Advisors is built to be contrarian: they don’t like agreeing with each other, and they don’t like agreeing with me. When you’re building something, you need somebody who can pick holes in it. About the only thing they ever agreed on was that I needed to step up the security I provide my customers.
The scoreboard — Monday, July 13
Revenue from this project: zero. No paying customers yet. What I do have: a distributor set up in Australia, and the paperwork in hand for distributors in the UK and the EU. The distribution network is coming along — you could say it’s a unique piece of what I’m going to do.
On Wednesday I’ll show you some of the team in action, so you can see exactly what they do. Friday, we’ll see where the week ended.
Don’t want to wait? Call my office — (855) 553-0229. Suzy Q will pick up, day or night. Tell her you saw episode one. Or book a guided demo and we’ll set the same team up for your business.