There is an epidemic right now in the market of people selling AI to businesses. Most of them care more about selling to students than delivering to clients.
I have nothing against the info business. I come from that world.
But in 2026, leaders in AI should be in the season of doing, not the season of teaching.
Think about it. Why would someone with thousands of students to serve put real effort into a handful of business clients if those clients are just a side project for him?
He would not.
The trap you're in
On one side, you try to build AI in your business on your own and struggle to get results. That's not your fault. You don't have the time to be on top of AI news and also run your business.
On the other side, most people who run "AI agencies" are not really agencies at all. They are info businesses that sometimes serve clients on the side.
I did research on my own competition as an agency owner. The weakness of the case studies these so-called agencies show, when they show any at all, is shocking.
Almost nobody is putting their hands on real businesses. Teaching pays better in the short term, so that's what most of them do.
Who I am to say this
Before AI, I built and scaled a high ticket coaching business to $3 million. Today I build AI systems for seven and eight figure coaching and consulting companies.
So yes, I'm on the selling side too. Run everything I say here on me as well. That's the point. Do your due diligence.
Most of what you see out there is people selling courses to developers, and those developers then try to sell AI automation to businesses without ever having run one.
It's a messy market. You pay for incompetence at a very high stake.
An example of what "yes to everything" costs you
One of my clients wanted a copilot for his sales team, something he could ask anything about his sales calls and get an answer back.
I told him no. Not yet.
First you need a system that reviews and scores 100% of your calls. I call it the AI sales manager. Then you put the copilot on top of that if you want more granular answers.
If you're running hundreds or thousands of calls a month and ask a copilot "how did the calls go this month" with nothing structured underneath, it just gives you an answer at random.
That's not using AI. That's making a mess with extra steps.
A developer who started his business a few days ago cannot tell you that. He has never sat inside a sales process. He builds exactly what you asked for, and you find out three months later it was the wrong thing.
The one who says yes to everything is the one who costs you the most.
And it looks like good service. You ask for something, and the guy who wants to close the deal says "sure, I can build that."
In most of the requests I get, my answer is no. Or "not like that, like this instead."
Almost nobody does that, because saying no to a founder takes two things: the experience to know you're right, and the guts to say it.
I learned this the hard way. I spent hundreds of thousands of dollars on agencies when I ran my coaching business. Maybe 5% of agencies out there are actually worth working with.
The AI creator isn't your guy either
What about the AI creator with thousands of viewers, keeping up with the latest news?
Probably not him either. He doesn't have the time, and you were never his real customer. His business model is different.
So you're stuck between someone who doesn't know and someone who doesn't have the time.
What to actually check
1. History, not tech stack
Don't ask if they can use this tool or that tool. Ask if the team has run a business before. Have they solved business problems, with or without AI?
Hiring an AI agency that has never run a business is like hiring an ads agency that has never run its own ads.
Same reason I'd hire a full stack marketer over a pure media buyer any day. The media buyer knows how to turn campaigns on and off. That role is getting commoditized.
The full stack marketer tells you if the creative or the message is wrong before you burn your ad budget. That's worth more than deep knowledge of one tool and zero understanding of the whole picture.
2. Real case studies
Don't look at their audience or how polished their videos are.
Ask who they've done this for, what numbers they actually moved, what the ROI was, and what the reasoning was behind it. Do they speak your language, the terms you use every day?
3. Do they actually have the time
Will they be able to deliver? Are they clear on deadlines? Ask how fast you should expect a response on Slack, and watch their reaction.
4. Be careful with "we automate anything"
Technically they're right, and so am I. Today you can build almost anything in any industry.
But I take discovery calls with businesses outside coaching and consulting and I say no. Not because I couldn't. Because I know I wouldn't do my best work there.
It's not just moral, it's operational. If you build a voice agent for a coach today and a software for a restaurant tomorrow, you never systemize your own process. You start from zero every project.
And guess who pays for the learning curve.
Pick the specialist, not the generalist.
The one question that matters most
If I could only give you one of these, it's this one.
Next time you're on a call with someone who wants to sell you AI, ask them: "What would you not build for me, and why?"
If they don't have an answer, they're not going to build the right thing for you either.
None of this is magic. There's no shortcut. There is work, there is judgment, and there are people who have actually done it before.
Then there are people who will be very expensive and not useful at all.
That's the truth.
-Ruben
Grow your revenue or cut your costs in 90 days. Or you don't pay.
Done-for-you AI systems, the same ones I run in my own agency and build for 7 and 8-figure coaching and consulting businesses. Want them installed for you? Let's talk.
