Under 50¢
Per usable AI-generated ad asset, down from about $10 with a designer.
I am Robert Sanberg, and I built AIIMN because I needed it.
I run an apparel business called WindRider, wholesale and direct to consumer. AIIMN is the room I went looking for when I was trying to get AI to do real work inside it, and could not find.
For most of WindRider’s life I wore too many hats.
I did all the Amazon work, I oversaw the marketing, I did all the finance, I had my hand in Ops. There was me, at nights, on weekends trying to get everything done.
For a few years that was manageable, and then the math started turning against me. Tariffs pushed up the cost of nearly everything we sell. Meta got more expensive at the same time, so the same results cost a little more every quarter. And while the costs climbed, the job itself kept growing, because somewhere along the way "be everywhere" became the standard, more channels and more formats, with the same hours, the same budget, and the same me.
AI was supposed to be the relief. Instead it was just one more thing on my plate. Every tool that promised to save time arrived with its own learning curve, its own tab, and its own decisions to make, and the work grew faster than the help did.
The cost of all this never showed up in a spreadsheet. It showed up at home, on my waistline, my hairline. The laptop started reopening after dinner. I’d be at the table but not really at the table, half-there, having to constantly ask my kids to repeat their questions as I was thinking about all the things I had to do or answering one more Slack message. None of it felt like a crisis on any single day, which is exactly why I let it go on for so long. It just kept taking a little more.
If you work inside a small business, you probably recognize the shape of this. Costs climb, the channel list grows, the team doesn’t, and the thing that was supposed to help becomes one more thing to manage.
I could not hire my way out of it.
There was no budget to add salaries against costs that were already climbing, so the only lever left was to change how the work got done. That is what pushed me at AI, not curiosity about the technology.
I watched YouTube videos, read blog posts, followed new tools, and tested constantly. Most of it I figured out myself, sorting through what sounded exciting, what worked in a demo, and what could survive contact with an actual business. I was not trying to vibe-code the latest app that popped into my head. I was trying to increase revenue, decrease expense, create more output, and improve how the company ran. Those are very different standards, and most of what is published online is written against the easier one.
The way through turned out to be unglamorous. Document the work I was already doing, in enough detail that something else could do it. Hand a piece of it to an agent. Read every output until I trusted it, and keep the approval on the parts where being wrong would actually cost me.
Creative stopped being the bottleneck.
Ads were the obvious place to start, because Meta wanted more fresh creative than we could produce and every week we came up short. The designer was never really the constraint. I was. The concepts, the hooks, the reviewing, the uploading, the copying and pasting of copy and URLs into the ads manager, all of that ended with me, a few hours a week, every week.
Once the production line was built, a usable ad asset cost us under 50 cents. It had been running about $10 with a designer. In a peak week one operator moved 100 to 150 ads through it. The money mattered, but the money was never the point. The time was.
The same pattern kept working elsewhere. Weekly reporting used to mean compiling numbers by hand, comparing Shopify against Google Analytics against Facebook, reconciling the places they disagreed, and only then actually thinking about what the numbers meant. Now the compiling and the reconciling arrive done, with every figure attached to the source it came from, and I start at the thinking.
What that looked like at WindRider
100 to 150 ads
In a peak week, with one operator across the production line.
Plenty that failed
Broken uploads, weak outputs, and systems I rebuilt after they met reality.
The failure worth naming is the one that changed how I build. An agent pulled revenue from the wrong source, concluded our product pages were broken, and queued a set of fixes. An approval gate caught the whole cascade before any of it went live. That is the reason I will not hand over a decision that costs money without a person in the middle of it, and it is the reason every workflow I teach has a gate in it.
The unlocks came from other owners, not from more content.
A handful of conversations with people using AI inside their own businesses got me through problems that had me stuck for weeks. Nobody handed me a magic prompt. What they gave me was a different question, a missing assumption, or a path I had not thought to look down.
That is a completely different thing from hearing what a tool can do in theory. A business owner can tell you whether it moved revenue, whether it cut an expense, how much time it actually gave back, what risk it introduced, and whether it just created another thing to manage. A win tells you something worked. A failure tells you why it failed, what it cost, and what to watch for next time, and failures are the part nobody publishes.
The difference was not more AI content. It was being in the room with people solving real business problems.
I wanted a room for owners solving business problems with AI.
AIIMN came out of that, and I won’t pretend the motivation isn’t selfish. I want to share what I have worked out at WindRider, and I also want to be surrounded by other owners doing the same thing, people willing to show what really works, what does not, and what they are trying next.
Here is the part I did not expect. When you are figuring all of this out alone, the hardest thing is not the question you cannot answer. It is the question you did not know to ask. The assumptions, the risks, and the possibilities you cannot see from inside your own business. Being in a room with owners working on different problems is what exposes those.
None of us has the finished map. We are drawing it as we go, all of us, and we get a lot further comparing real work than we do sorting through another thousand videos and product launches on our own. That is the room I needed. It is why AIIMN exists, and it is why I am in it too.
Two ways in.
Membership is $50 a month or $480 a year, and every application is read by a person. If you would rather have the system built into your business step by step, Built to Run is the six week guided build and six months of AIIMN are included, with no automatic paid continuation after that.