CAD by Morning, Sponsorships by Night

Hat count, those years: three job titles, if anyone had bothered to write them down.

Nobody did. That wasn't how the place worked, and you already know that if you read Season 1. There were no job descriptions. There was a thing that needed doing and a person standing close enough to get handed it. By the stretch I'm about to tell you about, I'd been handed three things and never put any of them down.

Marketing, still. CAD and product design, picked up off a hard drive and never returned. And project management, because once you understand both the parts and the brand, you become the person who gets put between them.

So here's a normal day.

---

The Morning Side

Morning, I'm in CAD. Sheet metal. A bracket, a fixture, a revision to a part that didn't fit the way the print said it would. Quiet work. Single track. You load the whole problem into your head and you don't come out until it's solved, because the cost of coming out is loading it all back in again.

I knew what we actually built. Not what the catalog said we built — what came off the line. I'd learned that the hard way, with a mower frame rolled onto the shop floor arguing with the files. So in the morning I lived in the real product. The dimensions. The constraints. What the fab shop could and couldn't do by Friday.

---

The Evening Side

Evening, I'm somebody else.

I'm on the phone with a dealer. I'm scouting a location for a shoot. I'm negotiating a sponsorship with no leverage except a machine and my own conviction. I'm deciding which event gets a unit and which one gets a no. This is the opposite kind of work. Fast. Social. Reading the other person. Three moves ahead. You don't load one problem and solve it — you hold twelve loose ends and keep them all from dropping.

I knew what the brand promised the market, because I was the one making the brand show up in it. The identity itself — I pulled the right people into a room to settle the logo, the naming, the look. But the promise a customer actually felt wasn't a statement on a wall. It was every photo, every placement, every sponsorship, every machine staged on a dealer's lot. I was the one deciding, shot by shot, what the brand looked like out in the world. Nobody handed me that to go enforce. I was building it in the field, one decision at a time. Then I carried the whole picture around in my head and applied it everywhere it needed applying.

Two different brains. Same day. Same guy.

I'm not going to pretend the switching was free. It wasn't. The tax on doing both isn't the hours — everybody assumes it's the hours. It's that your head never gets to settle into either one. You're always half-loaded with the thing you just put down and half-bracing for the thing you're about to pick up. It followed me home. The people who love you learn the difference between you being in the room and you being in the room.

But that's not the part that matters here. The part that matters is what I only saw years later.

---

What I Actually Was

I was the only place those two worlds touched. And the cleanest proof is a thing nobody else in the building would have even recognized as a problem: the photoshoot date drove the development cycle.

Here's what I mean. Before a product could show up in the market — the catalog, the launch, the placements I'd already lined up — it had to be photographed looking the way it was actually going to look. Production-intent. The right design, the correct configuration, the changes landed. A shoot was a hard date on the brand calendar. Miss it and everything downstream slipped with it.

But getting a machine camera-ready meant getting design changes through the dev process first. Through engineering. Through fab. By the shoot date. Which quietly made the shoot date — a marketing date — a hard engineering deadline.

Nobody on the development side saw it that way. To them, marketing wanted pictures. Nobody on the marketing side saw the dev cycle as their problem. To them, the part should just be ready. The two sides didn't share a calendar and they didn't share a reason. The only place where "the shoot is on the 15th" and "this change has to clear fab by the 12th" were the same fact was inside my head. I set the shoot. I drove the changes through. I was the calendar and the reason both.

One wire ran between what we built and what we sold. The wire was a person. Not the most qualified person. The one person whose head had been in both rooms.

The company had no system that connected what we built to what we promised. None. There was no place where product reality and market promise got reconciled — except inside one person who happened to span both. The connection didn't live in a tool, a process, a shared record. It lived in me. And it worked, right up until you realize what that actually means.

It means the most important link in the operation was a human doing it from memory. It means the day I moved to the next role, that link didn't transfer — it just went dark, and somebody downstream tripped over the gap and got handed it, the same way I'd been handed the hard drive.

Same shape as the parts manual. Different room. The company didn't have the thing underneath that should have held the relationship, so a person became the relationship. I keep finding this exact shape in every shop I walk into now. The quote that only works because one guy knows both the customer and the cost. The handoff that only happens because somebody remembers to make it. The connection between two departments that is, when you actually look, a single overworked human carrying it in their head.

Tribal knowledge, personified. Indispensable is the word everybody uses for that person. It feels like a compliment. It's a diagnosis.

---

If I'd Had AI…

Here's where the easy version of this story shows up, and I have to be honest about why it's wrong.

The tool-vendor version goes like this: if Matt had had AI back then, he never has to wear three hats. Generative design takes the morning. Content and scheduling tools take the evening. Coordination software routes the requests. One person spans all of it without the burnout, and the whiplash just disappears.

That's true. That is literally what those tools do now. And it is exactly the wrong way to deploy them.

Because making one person able to wear three hats faster doesn't fix anything. The two worlds still wouldn't connect anywhere. You'd just have a more efficient version of the same gap — the morning silo and the evening silo, both running quicker, still meeting nowhere but inside one person's head. You'd have automated the symptom and left the disease.

The thing that made me useful later was never that I did both jobs. Plenty of people do two jobs. It was that I was the place they met. The cross-wiring — knowing in the morning what a fab constraint would cost a sponsor in the evening, knowing at night which design change was about to blow up a promise — that only formed because I was the integration layer, by hand, painfully, for years.

So the right way to point AI at this isn't to reduce the hat count. It's to build the thing the company never had: structure. The information was all there — I was capturing every bit of it, in my head, all day. Capture was never the problem. What was missing was the layer that holds the relationships between what you build and what you sell, so the connection lives in the operation instead of in one person who's going to move on, burn out, or walk. AI can be that layer. It can hold the structure the company never wrote down — the graph that, in my case, was just me.

But — and this is the part the demos skip — it still takes someone who's been the integration layer to know what's worth connecting. The tool can hold the relationships. It can't tell you which ones matter. That judgment is the residue of having done it the hard way. I didn't know what the company was missing until I'd spent years being the missing piece myself.

Which is the principle I build everything on now, so I'll just say it plainly. The tool was never the point. What gets multiplied is human judgment and human direction — the deciding, the noticing, the knowing-what-matters. AI doesn't replace that. It puts a lever under it. Multiply the judgment, don't hand it away. That is the entire game, and everybody selling you the opposite has it backwards.

That's the advantage of doing it the hard way first. And it's the whole reason I do this work now: most of what a client needs isn't another tool to make one person faster. It's the structure underneath, so the business stops depending on whoever happens to be standing in both rooms.

---

The Question

Who in your shop is the only place two parts of the business actually connect — and what happens the day they take another job?

That's the link you don't have a system for yet. Find it before it finds you.

---