AI Won't Replace Experts, It Will Expose the Fakes
A 43-minute Marketing Minds interview on why automating a broken business just breaks it faster, and the one agent most service companies never build.
Posted
1 weeks ago
Duration
Format
Interview
educational
Views
2.3K
7 likes
57 · 43
Big Idea
The argument in one line.
AI does not create expertise, it scales whatever expertise already exists, which is why the businesses winning with it are the ones that could already do the work without it.
Who This Is For
Read if. Skip if.
READ IF YOU ARE…
You run a service business, agency, or client-based practice and you are deciding where AI actually belongs in your delivery.
You keep buying AI tools, training on them for a few weeks, and abandoning them when the next one launches.
You have real expertise and a repeatable framework, and you want to sell it to more people than you can personally serve.
You are trying to get a team to adopt AI and running into quiet resistance you cannot name.
You sell high-ticket services and the close rate has gotten harder, and you want a lower rung on the ladder that does not cheapen the offer.
SKIP IF…
You want a tutorial. This is a strategy conversation with no screen shares, no builds, and no prompts shown.
You are looking for technical depth on model architecture, evals, or agent frameworks. The ceiling here is conceptual.
You run a product or SaaS business rather than a service business. Almost every example assumes you sell your own expertise.
You already believe AI amplifies rather than replaces, because the first twenty minutes will feel like confirmation rather than new information.
TL;DR
The full version, fast.
Most businesses implement AI by pointing it at everything that is broken, which only makes the broken parts run faster. The better move is to find the one thing you are already excellent at without AI and build agents around that framework, because AI is the transformation of data and it can only transform knowledge you already have. For service businesses the highest-leverage build is an onboarding agent that collects client information and uses it to train the agents doing the actual work, which turns a generic tool into a tailored one. The durable advantage is not the agents, which are commoditizing fast. It is the expertise and the community wrapped around them.
Free for members
Chat with this breakdown — free.
Sign in and you get 23 free chat messages on us — ask for the hook, quote a framework, find the exact transcript moment, generate a markdown action plan. Bring your own key when you want unlimited.
The 95% statistic, and the guest introduced as the CEO of Cloud37.
00:49 – 03:51
02 · AI amplifies what exists
AI is a tool like Facebook or Instagram. It amplifies good or bad marketing, and adding verticals kills margin.
03:51 – 05:39
03 · What not to automate
Why the guru market oversells the ease, why finance is the wrong first target, and the amplification question to ask instead.
05:39 – 08:38
04 · Why teams resist
Adoption stalls on fear of the unknown. Selling a paid blueprint first makes the build visible before anyone commits.
08:38 – 10:21
05 · Where AI pays off
Department by department. Instead of an AI sales caller, build agents that turn clients into better operators.
10:21 – 12:02
06 · The onboarding agent
The gap between sale and fulfillment is the leak. An onboarding agent collects client data and trains the working agents with it.
12:02 – 15:14
07 · Custom vs out of the box
Why general tools cannot serve wide audiences, and why training your knowledge base beats chasing the next tool.
15:14 – 17:50
08 · Tool, chatbot, agent
Three definitions. A tool is a prompt sold as convenience, a chatbot is the new search engine that agrees with you, an agent transforms data into the next step.
17:50 – 19:30
09 · From agency to platform
Three years of throwaway code and unsellable knowledge bases, and the decision to build the platform instead.
19:30 – 22:16
10 · Expert studios
White-labeled apps built around one person's expertise, with the onboarding agent training the studio for each client.
22:16 – 25:32
11 · The ascension ladder
Self-serve at roughly $97 a month, then done-with-you, then done-for-you. Why service businesses never had a bottom rung.
25:32 – 28:10
12 · How much data you need
Build backwards from the finished output. If the client cannot show you one, the guest will not build the agent.
28:10 – 29:14
13 · Community is the moat
Agents are commoditizing. A hundred to a thousand committed people is what survives that.
29:14 – 33:54
14 · Case study: faceless scripts
An operator billing $500 an hour and burning six hours a day building custom GPTs, rebuilt as an onboarding agent plus a script studio.
33:54 – 35:12
15 · Where to find him
cloud37.ai and the social handle, plus the positioning angle he built his brand on.
35:12 – 37:30
16 · Where AI is headed
Optimistic on time returned, pessimistic on the class gap, and a reminder that AI predates ChatGPT by decades.
37:30 – 41:30
17 · Staying current
A business partner who reviews tools full time, business adoption articles over tool news, and a note on Claude Code.
41:30 – 43:07
18 · Remove AI first
Slow down, map the business, decide what you would build without AI, then add AI to that.
Atomic Insights
Lines worth screenshotting.
AI amplifies good or bad marketing the same way Facebook and Instagram did, so automating a broken business just breaks it faster.
Every service vertical you add needs more people, which drags profit down. Doubling down on one vertical is the margin play.
The more steps a chain of agents has to take, the more likely it is to fail somewhere in the middle.
An AI tool is a prompt wrapped in convenience. If you can write the prompt, you are paying for packaging.
A general chatbot will tell you that you are right, which makes it a decent search engine and a terrible operator.
The most valuable agent a service business can build is an onboarding agent, because the gap between sale and fulfillment is where clients are lost.
If a client cannot show you a finished example of the output they want, nobody in their business knows what good looks like yet.
You need very little training data when you are a genuine expert, because you can already judge and correct the output.
Agents and knowledge bases are commoditizing. The community around them is the only moat left.
One operator billing $500 an hour was spending six hours a day hand-building custom GPTs, roughly $3,000 of daily capacity burned on setup.
Financial workflows are the worst place to start with AI, because one wrong number makes every downstream number wrong.
People resist AI when it threatens them and accept it when it benefits them, which is why adoption arguments are rarely about the technology.
Large course and LMS platforms struggle to add real AI because their audiences span dog walkers to finance coaches and no single model serves both.
Remove AI from the question, decide what you would build anyway, then add AI to that answer.
Takeaway
AI scales expertise, it does not supply it.
THE EXPERTISE TEST
Every useful idea in this conversation reduces to one test: if you could not produce the output by hand and recognize it as good, no agent will produce it for you.
02AI amplifies what exists
AI amplifies whatever marketing and process you already have, so automating a broken business just makes the broken parts run faster.
Adding more service verticals drags profit down because each one needs more people, while scaling the single thing you do best protects margin.
03What not to automate
Avoid pointing AI at finance first, because one wrong number poisons every number after it and the work has no tolerance for error.
Start from what you are already excellent at without AI, then ask how to scale that, rather than asking what AI could do for you.
04Why teams resist
Resistance to AI is almost always fear of the unknown, and people will not buy or adopt what they cannot picture working.
Selling a paid blueprint before any build makes the plan visible, which turns hesitation into a decision people can actually evaluate.
05Where AI pays off
The highest-value use of AI in a service business is turning your expertise into something a client can run, not replacing a seat.
Treat AI as the transformation of data: it takes knowledge you already hold and converts it into the next usable output.
06The onboarding agent
The gap between a closed sale and delivered work is where service businesses leak, because sales hands off and fulfillment never owns it.
An onboarding agent that collects client details and uses them to train the working agents turns a generic tool into a tailored one.
07Custom vs out of the box
General-purpose tools fail at both ends of a wide audience, which is why large course and LMS platforms struggle to add real AI.
Train your knowledge base before you pick a tool, otherwise every new product on the market restarts your training from zero.
08Tool, chatbot, agent
An AI tool is usually a prompt wrapped in convenience, so if you can write the prompt yourself you are paying for packaging.
A general chatbot is a faster search engine with a flattery problem, which makes it unreliable as the operator of anything.
An agent is defined by taking data from one step and transforming it into the next, which is why agents chain and chatbots do not.
11The ascension ladder
Service businesses historically had nothing between a free webinar and a five-figure package, which is why high-ticket now feels impossible to sell.
A self-serve tier at around $97 a month proves you are a trusted source before anyone considers the expensive version.
Done-with-you and done-for-you tiers sell compression of time, not more features: the same system delivered faster with you closer to it.
12How much data you need
If you genuinely know what good output looks like, you need surprisingly little training data, because you can judge and correct what comes back.
Refusing to build until someone shows a finished example is a quality filter, because no sample means nobody there knows what good is.
13Community is the moat
Agents and knowledge bases are already commoditizing, so the durable advantage is the expertise and community wrapped around them.
A hundred to a thousand genuinely committed people beats a hundred thousand passive followers when the differentiator is human trust.
14Case study: faceless scripts
One operator was billing $500 an hour while burning six hours a day hand-building custom GPTs for each new customer.
Replacing that manual setup with an onboarding agent recovered roughly $3,000 of daily capacity without changing what he sold.
18Remove AI first
Slow down and map the business before buying anything, because AI cannot bridge a gap you have not defined yourself.
Remove AI from the question first, decide what you would build anyway, then add AI on top of that answer.
Glossary
Terms worth knowing.
Onboarding agent
An AI agent that runs the gap between a closed sale and the start of delivery. It asks the new client a fixed set of questions and uses those answers to configure the agents that will do the actual work.
Knowledge base
The stored, structured record of how a specific business does its work: frameworks, examples, voice, past outputs. It is what makes an agent's output specific to one company rather than generic.
Expert studio
A white-labeled app built around one person's expertise, containing their agents, their knowledge base, and their onboarding flow, which their clients log into and use directly.
Agent stacking
Chaining agents so the output of one becomes the input of the next. A script agent feeds a prompt agent, which feeds a video generator, with no human step in between.
Ascension model
A tiered offer ladder that moves a buyer from a cheap self-serve product, to a done-with-you program, to a full done-for-you service, with trust built at each rung.
Context window
The amount of text a model can hold in a single conversation. Past that limit, quality degrades and the user has to start a fresh thread, which non-technical users rarely know to do.
Custom GPT
A saved ChatGPT configuration with its own instructions and uploaded files. It behaves like a specialized assistant but has to be built and maintained one at a time for each use case.
Vibe coding
Building software by describing what you want to an AI in plain language and accepting the code it writes, rather than writing or reviewing it yourself.
Compression of time
The idea that buyers pay a premium not for more features but to reach the same outcome sooner, which is what separates a done-with-you tier from a do-it-yourself one.
“Remove AI first and then implement AI afterwards.”
the closing instruction, six words→ TikTok hook↗ Tweet quote
Topic Map
Where the conversation goes.
00:00 – 08:38denseWhy AI implementations fail
08:38 – 15:14denseWhere agents actually belong
15:14 – 17:50denseTools vs chatbots vs agents
17:50 – 25:32steadyProductizing expertise into an offer
25:32 – 29:14denseExpertise, data, and the human moat
29:14 – 33:54steadyCase study and operations
35:12 – 43:07steadyOutlook and closing advice
The Script
Word for word.
Read-along
Don't just watch it. Burn it in.
See every word as it's spoken — crank it to 2× and still catch all of it. The same dual-channel trick behind Amazon's Kindle + Audible.
17px
metaphoranalogystory
So Jeff, what was that statistic you were telling me again? I believe that 95 % of businesses are implementing AI wrong at current state. Wow.
95 % of businesses implementing AI wrong. Well, hopefully those businesses are watching this episode right now, the Marketing Minds podcast, because today I am so excited. We are joined by Jeff McPherson, and he is an AI expert, the CEO and founder of Cloud37.
He's helped businesses generate millions of dollars using AI following his AI ascension frameworks. And he's been building technology for over eight years. Jeff, thanks so much for coming on the podcast.
Absolutely. Thank you for having me. Awesome.
Well, this is one I am genuinely really excited to dive into, especially since as we were hopping on, even before we hit record, we were talking about how fast AI moves. I'm curious, what are the things you're seeing people get wrong the most with AI? And what are these businesses?
not fully understand that maybe is holding them back from fully leveraging it? When I'm having conversations with CEOs, COOs of like what we're going to implement into their business, the big thing with it is our business is ran before AI. Now all of a sudden AI is here, everybody's freaking out and they don't know how to do it.
AI, just like Facebook or Instagram, is just a tool to amplify good or bad marketing. So it's like people are trying to plug in AI, think it's going to resolve their business. Now, over the last few years of us going through and building custom tools and building our own platform, let's just use an agency as an example.
Service -based business, you add more services to be able to drive more revenue. And then in time, you have to add more people and then your profit margins start to drop. When, in a matter of fact, if you just focused on that one vertical and really doubled down and scaled that one thing, this is where people have...
more, drive more revenue, drive more profit, and actually have more freedom in the long run, instead of trying to add eight, nine, 10 verticals, because AI is just constant transformation and communication from one agent to another, just like your businesses and humans. So the more steps that it has to take, the more likely it's going to mess up.
And this is where I just think businesses are trying to automate and AI their whole broken business. But a matter of fact, if they could just identify the one thing that they're really good at and double down and scale it, that's where they're going to have more freedom and be able to really make money in this AI space.
Yeah, it makes a lot of sense, right? People overcomplicating it because they think AI, they think complicated. Why is it that AI, you think people are overcomplicating it?
What about AI and some of the different ways to implement it in businesses? Are people not fully understanding? Like what's kind of an unlock for them that can help them picture how it should be implemented?
One of the biggest issues with AI is all the gurus on the market. This is, I mean, you give a digital marketer the ability to make money from something, they're definitely going to do it. I just believe the markets are being taught wrong.
That's just generally what it is. It's unfortunate. But AI is just not as good as what people say it is.
It takes work. You have to put your boots on in the morning. You have to continue to work with it.
I mean, you've launched your own things as well. It's like these things don't get set up overnight. And the markets are just teaching people, it's like you can click a button, it's going to run your whole business.
Like, yes, that's exaggerated, but that's essentially how people think when you're having conversations. What I also think has been a real issue is it's making people lazier. Like the common sense in people is really starting to be even less.
And then people are just assuming that it's going to do everything and it's making us lazier and lazier, which in a matter of fact, like the people who are winning are the people who are putting in more work. That's a really good point. And so.
When it comes to AI, you mentioned that it takes work, it takes diving into it. We're going to talk about some of your favorite tools, some of your favorite processes. Also, I know you built your own, you know, tool or actually company that can actually help people build tools too.
So we're going to dive into all of that coming up on this episode. But when we look at how people are implementing AI, what do you think are the best things that businesses can use? AI to automate and to make better?
And then what are areas where maybe it's just not quite there yet? Maybe it will be in the future, but people are kind of getting stuck trying to trying to build. One of the big ones I try and get most people to stay away from, unless you really know what you're doing, is like financial, financial world.
It's just like because if one number is wrong, they're all wrong. So it doesn't have as much forgiveness in it. I know people are trying to do all the spreadsheets and analysis and all that stuff like I get it.
But it's just, to me, it's like if you look at these big companies, let's say like QuickBooks, most people know what QuickBooks is. I know they have some AI features within their system, but if they're not full -blown AI yet, I mean, why do you think that you can do it in your business? And like this kind of like, again, back to the common sense piece, where we're seeing people have their biggest success is all of us have some sort of framework, strategy, expertise, something that we've done continuously over and over, something that's built us to what our businesses are today.
If you look at those frameworks in some sort of capacity and you build an agent around them, it's just like you did with your tool. It's like you can build an agent that understands these because that's what gives you the scalability. It scales what you're already good at.
So instead of trying to figure out what AI can do for my business, it's looking at it in a way where it's like, what am I amazing at without AI? And how can I scale it with AI? And that's the difference between people who are succeeding and people who aren't.
Yeah. I think that also is a really great point, too, is using it as an amplification tool versus trying to build something completely from scratch or areas where it maybe doesn't work quite as well. We're going to get into some of your favorite tools or things to utilize.
But I think one thing that's also valuable to understand your perspective on is when businesses are implementing AI. One common thing, and I know I've experienced this and a lot of people are experiencing this, I'm sure you've seen the same thing, is one person on the team, whether it's the owner or a marketing director, you know, or somebody on the team gets really excited about AI, starts implementing it.
But maybe the adoption is. longer or a little bit more challenging to get other team members to adopt AI. What are some of the techniques?
I know you had a background, too, in consulting businesses on AI. I know you've evolved. Now you've got your own, everything that you're building with Cloud 37, which we'll get into in just a second, right, that helps build AI agents.
But I know at one point you were doing more AI consulting. Why do you think it is that people are resistant to AI? And what are some of the ways that...
When you were consulting businesses on this, you helped get full team buy -in to start implementing AI within a business. Good question. I don't think people are buying in unless it's like a shiny object and they're seeing that, which are typically people that we don't like the tire kickers.
We stay away from those people. It's the fear of unknown. That's really what it is.
Like there's still a massive education curve that needs to come in. Like, I mean, we're... We're in it all the time.
But you think that the amount of people that aren't in it, people are just using ChatGPT still to run their business. Like all of these things, like the numbers are still quite large on just the adoption of the education piece. That's where it is.
Like people, there's just a fear. And like people don't buy when they don't have that real understanding of what they're going to purchase. And that's the unknown piece.
So I mean, how we were able to get people to buy in the beginning is like we... we had a tear effect. Like we went through and I used the analogy, we would build a blueprint for somebody.
So I mean, we would charge people for this blueprint and it's just like building a home. Like we're gonna map this whole thing out for you so you can see what you're going to do. And at the end of that is like, now you can see what we're gonna build, here's why.
And that this was in the early consulting days where in those times, we really started to see what was working, what wasn't working. It's like people wanted to implement AI into their systems. And as we were doing it, it's like, this is broken.
It's not really giving them the uplift that they want. They're spending time and resources on things like basically in six months that AI might replace. And it kept going back to literally, what are you guys the best at and what people are passionate about?
And that's always in somebody's business. Why you started a business, why you continue to drive in a business and it continued to lead back to those. And unfortunately it's like sell is what is it?
So sell them what they need, give them what they want. So it's like you sell them into what they need and then with your consulting and with your expertise and with your guidance, you essentially push them down the road that's going to give them the long -term sustainability. Yeah, I think that makes sense, right?
It's like they have certain things they're excited about, but you want to make sure that they actually get the things that are actually going to create the biggest impact. What parts of a business, I kind of asked a version of this question earlier, and you had a really good answer in terms of looking at kind of amplification.
But now let's actually break it down in terms of parts of a business. Like if you look at departments from marketing to sales to you already kind of touched on like more of a finance side, but maybe operations, client fulfillment. What are the areas that you see benefit the most from AI?
And then do you have some examples of how different businesses are implementing this? I know there's a wide variety of different businesses, but a lot of people that watch. this show they have online businesses maybe other coaching consulting course creators or their agencies so i'm curious what are some of the the areas within maybe a client -based business a lot of people watch this have are involved in client -based businesses that you've seen have been beneficial to automate with ai so it's it it's it's tricky and this is i mean this is what people pay us for so yeah In our businesses, I'm an open book.
I mean, it doesn't really matter to me, but it's like, let's say sales for an example. So instead of implementing a sales caller for a business. Where it is, is if you're somebody who's a sales expert, you understand objection handling, how to like pacing of calls, all of this fun, all of the things that a sales expert would know.
So if you build agents to be able to teach people how to become a sales expert as well, you're transforming people into a better life. AI is transformation of data. That's all it is.
It's taking knowledge and it's transforming it into the next thing. So sales, that's one aspect. Let's say we have a client on our platform that...
In two agents, he's generating healthy six figures a month, teaching people how to write faceless TikTok scripts. But the difference with it is, and this is where it becomes powerful and where there's a gap in the market, is I've said for over a year, the most important agent that you can build in your service -based business is an onboarding agent.
Because from the time a sales happen to the time you do fulfillment, there's a gap in the middle, always. And people really, your salesperson neglects it and your fulfillment people don't really support it, but there's never really that good onboarding experience. But if you can have an onboarding agent, let's call it, that trained all the agents to do the work for them, so you dynamically train them, then the output becomes custom to the business.
And this is where it becomes very, very scalable and very, very custom. So you're not a generic tool on the market. You're hyper -focused expert in your field, whether it's sales, email marketing, jar, writing newsletter, doing YouTube scripts, doing YouTube thumbnails.
It doesn't really matter. But when you onboard people, you have to collect their business information so you can train the agent doing the work to be able to give it that hyper -focused, hyper -specific output that they want. And that's where we're separating ourselves in the market is we're dynamically training your customers, your clients, information before the agent that we've built for you does the work.
Makes sense. So the best agent you can build to answer your question, onboarding agent number one, to train your expertise, whether it's sales, email marketing, newsletters, YouTube scripts, HR, like creating content, it doesn't matter. But you have to think in a way it's like, how do I transform somebody to level them up?
It's a transformation of education. That's where I believe AI is the best at and where it's only good at right now. That's a really good way of putting it, too, and also leveraging the data that the existing client that you're working with has.
Could you share why it's important to almost build, and this should honestly be a softball for you because I know this is a big part of what you do, but why is it so important for people to build? kind of a custom AI setup for their own business versus using out of the box solutions. Just let's say using standard ChatGPT or Gemini or an out of the box like.
AI sales assistant or let's say there's an onboarding one versus building something more customized that's attuned to their to their business and their particular needs. I mean, just like really anything, it's like you can go and I can search Google for pre -AI and I could get an answer. But if I would still rather hire or go to an expert who would give me the answer that I need, like it's like it's the two parts of it.
It's like, is it a trusted source? Yes, no, maybe. It's like and.
is the output that I'm, how quick can I get the output convenience? Because we're lazy people. Like that's just, that's just really what it is.
And then the last one is, is it, is it tailored to me, my business, my personal life, whatever it may be. And most of these tools on the market, and this is why you see it's very hard for large softwares and large platforms or like LMS, like Kajabis and stuff like this. are struggling to implement AI into their system, like true AI.
Because you have somebody who's teaching like a dog walker, and then you have somebody who's teaching how to create financial freedom. The spectrum's way too long. It's too generalized to be able to do it.
And this is why a lot of these big platforms are struggling to adopt AI properly into their systems. So I mean, like the general tools, there's cases for it. But also the problem with general tools is...
And this happens from my experience losing my first business to Facebook, our good friend Facebook. When the new tool comes, you chase that next shiny object. So you keep training your business on these next tools that come out.
I mean, how many AI sales callers have you seen over the last three years? A lot. So it's like if you train it on one, then the next one comes out, then how do you train the next one?
So it's like where we're trying to really teach people is like, don't look for the tool. train your knowledge base first, train something that your business is on, and then find the tool if you don't know it, or the expert, let's say you in the YouTube space, and you work with the expert to be able to train your knowledge base to give you that output.
This is where you start building the verticals within your business. And I just like stop looking at the tools on the market. Like they're, most of them are garbage.
They really are. Like, you know what, you know how to run your business. Like, I mean, well, most people, most people know how to run their business, not everybody, but you know how to run your business.
Like if you could remove AI from the world and you had to step back and start your business again, what would you do? And not enough people are taking one step back to move forward because they just, they're chasing the shiny objects and they just believe it can move that quick. Yeah.
Yeah, I like the way of looking at that. And I think people try new tools all the time. We see it all the time, you know, bouncing around.
But how deep have you actually gone to build out exactly what you're looking for and actually have it in a way that it's going to make an impact for your clients? And so in terms of some of what you're doing, I know with Cloud37, I think it's actually a good kind of transition in a second here. You know, could you explain the difference?
Before we get into more of specifically what you do, the difference between using an AI tool, an AI chatbot, and an AI agent, right? So the pre -built kind of tool, using a chatbot like ChatGPT or Gemini, and then also what an agent does and how that's a little bit different than a tool people might be familiar with, right?
Where it's kind of a pre -built setup. chatbot that people are familiar with. You kind of ask ChatGPT versus an agent that does things on your behalf.
Could you maybe share? And I know the answer to some of this, but this is really for the viewer or listener who's watching. And I'm curious what your definition is going to be between those, because I think some people get confused.
They hear AI, they think it's all the same. Yeah, I was going to say we all have slightly different definitions to the whole thing. I mean, like AI tools selling you convenience.
At the end of the day, underneath the AI tool is a prompt. And if you knew how to build the prompt, you wouldn't buy the tool. You're buying convenience, but the prompts really aren't that complicated to create.
The other one is an AI chatbot. So chat GPT. I mean, it's generalized knowledge is basically the new Google.
Like it really is. You can ask it anything. It'll tell you anything.
The only problem with it is it'll tell you that you're right. Oh, yeah. Alec, I love your what you're talking about with Jeff.
I mean, they probably don't at the end of the day. That's I mean, that's just it's the new Google to me. And like you use it for those sorts of things.
But I just don't believe you can run a business on it. And that's my stance. I know there's going to be a lot of people like, well, I do.
But I mean, at the end of the day, it is what it is. And then your AI agents, depending on the definition, to me, AI agents are. transforming and predicting the next thing.
So it's like, again, it's transformation. So I'm taking it from this agent and I'm taking the data. So agent number one, I'm just gonna use the onboarding agent because I think everybody needs to hear about these more.
So I take the onboarding information from my client. What do I do with this information? This agent already knows how to take this information and do the next step with.
So then when it's in this agent, then I take this agent. So it's doing... The next step is understanding is being able to take the data and transform it into what the next step is, which is kind of your prediction side of, and this is where like the proper training, you can get into hallucinations, context windows.
I mean, there's a whole bunch of things that come into play there. But to me, it's like an AI agent is essentially a well -trained human to predict what the next step is with the data that it's given. I think that makes sense.
It's basically taking the role of what a person would do if they were following an SOP, but you actually have the AI doing it. What I'm curious about and the reason I want to know right now a little bit more about Cloud 37, because I've seen your evolution, right? You know, being one of the forefront leaders in the world of AI and consulting different businesses on how to implement it to now having Cloud 37 that builds AI agents for businesses.
I think it's really relevant to understand exactly, you know, how that works and what that is. And then that's going to allow me to ask some questions around not just what you do, but also. how these agents plug into different businesses, what you're seeing right now.
So could you tell everybody a little bit more about Cloud37, what it is that you do? What does it mean to build out AI agents for a business? Could you share a little bit more on that side of things?
Yeah, so I'll give you a kind of a quick backstory to the whole thing. So I mean, like we started as an AI agency three years ago at the very beginning. And as we were working with clients, we ran to the same situations.
Like we had throwaway code with building agents and then we still had to collect all the business information into a centralized place so the agents could essentially do the work. This is when... Two and a half years ago, I started selling knowledge bases before people even cared what a knowledge base was.
And those were extremely tough sells. Where the agents side came is, is because we kept having throwaway code, we started with like, well, if we build a platform, we can just copy and paste the exact same agents because they're all the same at the end of the day. And this is kind of where Cloud37 evolved.
Now, with that being said, there's lots of tools that can build agents and knowledge base and all that fun stuff. But it's the scalability behind them. And my brain always goes into productizing what people are really good at.
That's just always, always what I've done. So we have a knowledge base. You can build agents, but...
agents and all this stuff is becoming commoditized. But what still is not commoditized is like our expertise in the community that's built around it. That's essentially your only moat.
So what Cloud37 allows people to do is to essentially build tools or what we call expert studios around your expertise. So let's say like, we'll go back to this guy who helps people do write faceless content TikTok scripts. And people will compare to ChatGPT in like custom projects, which I can explain the difference, where he can resell them.
So we help manage like the processing, the tokens, the security, the DevOps. So it's even more so you don't even need vibe coding. Like we've literally skipped the vibe coding stage.
You come in here, you talk to an agent because we've had it. We built an agent that helps build agents, that helps launch agents. So you can get these things done.
It hosts your knowledge base, which is your most important thing to your... business and then the difference between like us in a custom gpt is these expert studios which are essentially white labeled apps of their own it has an onboarding agent so when you're onboarding people your clients but after they pay they go through and they basically answer a series of questions whatever it may be and it dynamically trains the agents in their studio to give them that custom output So you remove the onboarding stage, you give them custom outputs, and it creates complete scalability for your business.
So we work B2B, and then they go B2C. Whether you're an extension of another client, whether you're bringing it to consumers, again, you're transforming people into what your expertise is. Because as a service -based business, I can only work with so many people.
I can. But if I can bottle up my expertise, and you could get 80 % of it, and then you have to do the other 20%, because all you're trying to do is teach people what good output is. That's all you're trying to do.
So, I mean, you've been in this space for a while. It's like people don't go through info products. People don't watch courses.
They never have, they never will. So, but you combine agent, the output, what people want with video to teach them what good output is. The videos are now two minutes long with the actual output itself.
So you're teaching people what good output is for that very specific role. So you can go and you can find five, 10 different experts. Use their agents, which are custom to your business, which are tailored to their expertise.
And this is where we're scaling it. We're helping people turn their service into a subscription, which then falls into that Ascension model. Because now you've taught people that you are a trusted source and they can do it yourself.
Now you can go in and you can start picking these people out. Or they're like, yeah, this is a low ticket offer. I trust you now.
Now I'm going to buy your high ticket. And this is what Cloud37 is powering. We're powering experts.
studios at the end of the day. And so do you recommend, and I'm sure you may see a little bit of both, do you recommend people roll this out as kind of an individual, almost like an agent software option that is a lower ticket subscription that then ascends people? Or are you seeing people implement this into their main kind of services business or both?
Like, what are you seeing on that side? Because you kind of talked a little bit about... how this would work onboarding.
But then you also kind of alluded to almost like a software that can ascend. Are you seeing people use it for both of those use cases? It's both.
It's both. And at the end of the day, even if so, there's let's say you're a complete do it yourself. Like that's, that's literally what it is.
So people can come in $97 a month. You got people in there, they're getting their agents, they're getting their output. They got some training.
There's a community aspect. I mean, it's all good. And there's a lot of people think of it like school, but with AI agents, that's how, that's actually how we're getting compared to right now.
So it's like, school's great, but it's like, you don't have the agents. So it's like, if you could have agents attached to school, like this is where it becomes very valuable. Then you could have your next ascension where it's more done with you.
So you have like weekly calls, you're doing these things. It's the same agents, but you're a little bit closer to me. Because people want closer to the convenience.
At the end of the day, talking more talks about the other time. Sell convenience. This is what he talks about.
Sorry, he talks about compression of time. That's what it is. When you work with me on a one -on -one basis, if you do it yourself, you can get the output.
But if you do it with me on a done -with -you basis, I'm going to get you there a bit faster because you're going to be able to ask questions. You're going to be able to do all these things. And then the last one is the done -for -you.
It's like, you don't want to do any of those. No problem. I can still do that work for you because there's always going to be people who want to have the done for you service, the white glove service.
But the problem with service -based businesses, they've never had the ascension. And it's too hard to go from a webinar to a $10 ,000 a month package in the service -based business right now. It just is.
People are struggling to buy high ticket services. What do you think is causing that? And this is a little bit of a different question, but what do you think is causing that shift?
Do you think AI has to do with that shift? Do you think it's more economic? Like this is a little bit of a tangent, but I'm curious you saying that, of course, you're talking about AI as the solution to that, but do you think that this is brought about because of AI?
And I definitely have seen people saying that. Do you think it's more economic? Do you think it's just kind of the general fluctuations?
People are saying the trust recession, like. I'm actually kind of curious what your take is on that. And then we'll loop it back to AI and how the agents and this setup actually resolves it.
I think it's just the economics, just where the world is right now. Like people are struggling, world is struggling. There's the AI is definitely putting a play because it's teaching people that you can do it for cheaper.
When a matter of fact, you can't, you still need to know how to do the actual job itself. And then where that stems to is just the arrogance in people, the egos in people. that think they're smarter than they actually are.
It's just absolutely mind -boggling to me. Like the great thing about AI, which I love AI the most for, is exposing people for who they're not. It'll expose lazy people and it will expose people who are not actually experts, which is my favorite piece of all of AI.
It really is because it's going to give people who work hard, who do, who actually put in the time and really know their stuff is going to give them the opportunity to be a part of this world. Yeah. And how much data do you think people need to be able to build this out?
Because I know, you know, we've done some things to build kind of like a fine tuned YouTube ad script. You know, I don't want to say agent because I think agent like you're describing is a little bit different. But in terms of like a highly optimized, fine tuned GPT or LLM.
Right. But it has a lot of data and it's going to be way better than if you just put it into standard chat GPT or cloud. How much data do you think people need to be able to actually leverage this effectively, you know, in terms of existing data of how something's been done in the past?
I mean, if you're an expert in something, it shouldn't take that much data because you know what good output is. So it's like if you don't, then you need to fine tune it on the Internet. I mean, it's true.
It's true. Like, if you know what you're doing, you can get yourself there. You absolutely can.
But like when we're building agents, you build them backwards. It's like when a client comes to us, it's like, I want to build this.
doesn't really matter what it is. Well, the first thing I ask is like, can you show me the final product? Like what the final output is.
If they can't show me what that final output is, I won't build it for them. Because that means they're trying to make up a system to be able to create a hypothetical output that they have no idea what good looks like. But if they can show me it, I can build an agent, which is essentially like metaprofting, which you can build a system to be able to match what that output is.
So it's like if they've got the output, that means they have somebody in their business that knows what the good is. Building the prompt is very easy. Then all you just keep doing is just feeding it more information.
Like you just keep feeding the beast and it just keeps spitting things out. So in short, you don't really need that much data if you know what you're doing. If you don't know what you're doing, then you need to really.
put that data in there and start figuring it out and using the data points to decide. And that's the thing. It's where you're going to be.
And it's very interesting because I see people with AI say, oh, it's just going to make everybody automatically an expert. But if you don't know how to actually set up the AI agents the way that you're describing it or actually translate your expertise, you're not going to know what's a good output. You're not going to know actually how to reverse engineer what needs to be built to have a good output.
So I think that makes a lot of sense. And it also... A lot of the people who are watching this episode are expert -based businesses of some kind or service -based businesses, client businesses.
And so that should give a little bit of a sigh of relief. It doesn't mean that people are off the hook. They need to start leveraging AI, right?
Because you'll be left behind if you don't. But it also doesn't mean at the same time that people who are doom and gloom saying, well, the expertise isn't worth anything, the AI can do it all for you. I think you answered really well how that's not the case, right?
The expertise with AI. is gonna be what gets the best results. That, and then the last thing that we preach is building community.
If you don't have community, you're gonna get smoked in the long run. And you don't need to have 100 ,000 people, but you need like. a hundred, a thousand like diehard people that would do anything with you and stay with you through and through, because that's going to be, that's going to be your only moat.
It's like at the end of the day, if, if AI comes to a point where it can do everything, it's going to go, the pendulum is going to switch backwards. People are going to want to have that true human connection and you've got to build that community now so you can, you can get yourself there. Yeah, absolutely.
Absolutely. The community is so, so key. And when it comes to.
Some of the clients that you've worked with, with Cloud 37, I think some people, when they think about this or they hear this, it gets theoretical. Could you share some specific examples, case studies of this actually being implemented into a business? Where were they at before?
And then how did this, especially maybe the onboarding or, you know, walking them through the client fulfillment process, whatever it is, like how did this transform things and what were the results? I mean, we're just building out one of the case studies today. So we were just interviewing the guy before us.
It goes back. It doesn't really matter. So I want to step back a bit.
Like all agents are built exactly the same. There's your output. There's your expert who trains it.
There's your prompt. And then there's your input. So it's like wherever whoever's watching it, those are the only four verticals.
I don't know any others. I really don't. Sure.
In the prompt, there's tool calls and there's APIs and all those things. But those are the four pillars you need to think about. So if you can think of it as what am I the best at?
Here are the four pillars. Now I'm going to listen to how Jeff and his clients are building the things. So the one client, faceless TikTok scripts.
That was his business. But he was doing it on ChatGPT. He was doing custom projects.
When he had somebody come in, there was a couple of problems here. It's like one, he had to build the profiles because every single custom GPT had to be trained on an individual audience. So people paid for an audience is what he calls them, which is essentially what their scripts are going to write about.
That's just really what it is. So he had to build each one of those individually. He charges about $500 an hour and he was spending six hours a day building these for customers.
Okay. So it was a lot. That's not including...
Then he had to set up the custom projects for people or custom GPTs. And then he had to teach people how to use custom GPTs. Because at the end of...
After a period of time, when the thread becomes really long, you have to create a new thread, create a new thread, create a new thread, create a new thread. And it's like the demographic he was working with didn't understand that. They just thought you could just keep going, you keep going, you keep going.
Oh, yeah. Which is not the case. Again, it's something that we take for granted.
We really do. But you think about the markets just don't understand that. So when he came to us, he had a good business, but he was capped.
He had nothing else. So back to the onboarding agent, it's like you have the same questions, you ask the same questions, you build the same audiences, all the frameworks there. So we built an onboarding agent so when his clients now pay, his customers now pay, they fill in the same questions that they had to fill it anyways.
Because we trained an agent to build the audiences. it automatically builds an audience for them. So he doesn't have to do that work anymore.
So right there, he's getting $3 ,000 back a day based on his time. From there, it automatically goes into their studio or his app. Where then the script writer that is also based on his frameworks on how he writes scripts, he helps people create faceless content.
So the scripts, and then he's got the images and all that fun stuff. Whereas then now it reads the audience automatically based on the script, based on what they paid, and it outputs the scripts for them. But because we've just simplified the whole user experience process, these people understand it's like, you've got to switch these things, the context windows, like all of this stuff just becomes easier and easier and easier and easier because it's more of an app than just like an ecosystem.
It's very tailored to him. This is like one of the best use cases that we've seen, but it can work for HR. Like let's say you're hiring somebody.
Okay, so you've got to hire somebody. You've got to ask questions. You've got to get them to submit a resume.
And then when it comes in, you have an agent that essentially qualifies if they're good or not. Then when they go in, it's let's say you hire the person. Great.
Now, how do you onboard them into the system? You have all your SOPs, you have your documentation, you have all these things, but it's personalized to them because everybody learns differently. They really do.
So again, it's a matter of transforming the person that you're selling to, to make them better into your expertise. That's all it is. And doing so in a way where you can scale it because it's used by AI.
So you have that personalized touch. but across as many clients as you can enroll. Correct.
Because now that, let's say the TikTok guy, now what he's doing is he, because he can just keep adding agents to it now. So now he's added in a prompting agent. So based on the script, because again, this is where like the agent stacking comes in.
So now he just adds another agent where based on the script, it'll write a prompt, which then will generate the faceless video. So now they don't have to worry about that. So now he's getting into other platforms because like.
on TikTok and Facebook and all these platforms are slightly different. So it will help create different content for different platforms as well. So he can help them scale horizontally while he scales vertically.
Yeah, that makes a lot of sense. And I think that one of the things that is really valuable is what you're doing is helping to set this up because I think a lot of people don't understand how this works or how they can actually put this in action. After you share a little bit more about how people can learn more about Cloud37, I do want to ask.
some bigger picture questions about where you think AI is going, how you stay up to date with AI, with the changes and all of that. So we'll take a little bit of a shift. So make sure everybody stick around because we're going to get into that in a second.
But in terms of Cloud37, could you share maybe a little bit more about... how your process works and you're working with clients. I know you shared kind of like a big picture of your building out the AI agents, but how can people learn more about you and what you do?
How can they build their first agent with your process? Maybe you could share a little bit about that. I'll put any links you mentioned down in the description and show notes as well.
Yeah, I mean like cloud37 .ai, I mean, that's our website. It's just, it's in the process of being revamped right now. Really excited about that.
Our vision is to help people build these expert studios out and really scale. the profit of their service -based businesses. It's really tough to watch so many amazing people get left behind because they're trying to solve problems in their business that are just irrelevant if you learn how to look at what the goldmine is.
On social media, I mean, Jeff Mack AI across any of the socials. I mean, I'm putting stuff out there all the time, trying to keep people up to date on what I believe. I come more from like how businesses should implement AI, just not another AI tool to implement.
That's kind of where my branding is coming in. And it's just worked extremely, extremely well for me. That's great.
So we'll put those links down below. Definitely go and check out cloud37 .ai. And I also know you put out some great content around AI as well, which is great.
And so in terms of the future of AI, this is where I kind of want to shift the conversation for these last 10 minutes or so that we have on this podcast. It's a big topic. I'm very curious because you are so...
inside of AI and building this out, working with clients, creating this with these AI agents. Where do you think things are going with AI? Optimistically, I think it's going to give people time back with the things that matter most.
Family, friends, stuff like this, being a family man myself, I think in time it's going to be able to give people that back if used properly. There's always bad apples out there. It's just regardless, but AI is just a tool.
The people who run the tools are the people that are going to ruin it. I do think there's going to be a huge gap between upper class and lower class. I think the middle class is definitely going to start to spread in time.
And our goal at Cloud37 is to bring as many people up into that upper class with us as possible and really be in that top 10%, whatever the number may be. But at the end of the day, there's a lot of really good... When people try and challenge me on AI...
If you I mean, AI has been around for a long time, more than just chat GPT. It's in your phones like Alexa and like Siri on your phone. That stuff has been that stuff has been around in the health care world.
Like it's finding diseases that have never been found before. It's helping doctors with surgeries that it was never been able to do before. It's in like building products that people don't even know about.
So it's like people are resistant to AI when it hurts them. accepted to it when it benefits them. And those are the unfortunate things.
But I mean, at the end of the day, you can only stand in your lane for so long before the world will just force you into what is going to happen. And that's just the world of AI. Yeah.
So you might as well get on board and start building with it and embracing that. I think fighting against it, you know, at this point with where we're at and where we're looking at where it's going is that's going to be a challenging proposition. So might as well embrace it and then be a creator on it, like you're saying, and be in that upper echelon of people.
And how do you keep up to date with AI, right? There's so much stuff going on. We even started the episode talking about that.
How do you stay up to date with all the latest in AI? I'm actually genuinely curious, not just overall, but like, are there tools or are there, you know, specific places that you go, like AI processes that you have to keep you up to date with everything? Like what's your process for staying up to date?
Well, I mean, I'm fortunate that I got a business partner with 300 ,000 followers with AI tools. So that helps. He reviews them all the time.
What we've noticed with it though, is like 2025 was definitely an AI tool gold rush, but people are starting to become really understanding that not all the tools on the market are helpful. They really aren't.
So that helps. I mean, at the end of the day, I'm so convinced on where this AI market is going and what we're building that it's helped me remove the noise. I'm just like all signal.
I really am. I read a lot of business articles, more so as like how businesses are adopting it and how they're not adopting it, most importantly, because I'm trying to find trends on like what's working because there's a ton of articles that say it's like business impact and the actual use of AI in the businesses. Like it's not really, there's no ROI to it.
Like there's all of these articles that are just being put out everywhere. So that's kind of where I stay up to date with most of the things in terms of AI tools. I stay away from it.
I think selfishly cloud 37 as a knowledge base, you're scaling your expertise and building agents, cloud code. It can do basically everything as well. At your, at your next level, we have the ability to like.
People build agents on our platform and use them as skills on Cloud Code. So that's kind of cool because we do want to have that overlap there for people who are a little bit more experienced. And then there's the aspect where it's like, sure, you can get into these open clause and stuff like that.
But I still think those are way too early. I mean, if you don't know what you're doing, that's going to destroy you really fast. But between Cloud Code and or like Cloud in general and...
Cloud 37, I just don't believe the tools on the market where they are right now, unless you really understand what that output is, are going to give you the value that you need. You're going to spend more time setting them up. than scaling your business.
That's really good advice, especially coming from yourself and, you know, and then also Matt as well with the AI, you know, tool overviews and stuff like that. It's good to get like a sense of what all that looks like. And I agree.
I think like this is the year of focusing and building it out and also building your own agents and platforms. And so that's literally what you do, which is phenomenal. Yeah, well, I mean, like even think of like yourself, like you built your own tool.
But that was based around years of trial and error and your expertise. And you had the willingness to go and build it yourself. And you're transforming people into better YouTube.
Like that's what you're doing. That's literally all AI is. And that's why you're having success with it.
You said it's not like your number one thing you focus on, which is fine. But at the end of the day. That's what AI is.
And it's building your community. It's building more opportunities. And that's all we want people to do is what are you the best at?
Bottle that up and scale it as much as you can while building a community and just buy your time to the next thing. That's a great way of looking at it. And it's a way to leverage AI and have it leverage your time and your expertise and your skillset.
It's a 10x multiplier, but you have to have something there to multiply. Like you were talking about earlier, right? There's a big difference.
If you actually have that expertise and you're a true expert. then AI, when done the right way, can help you multiply that. And it gives you massive leverage, more time, more ability to execute with clients really well.
So onboarding, different apps, things like that. Or even attract new people, like what we're doing with Thumbnail Creator. That's a great way of looking at it.
Anything else you think is valuable for everybody who's watching or listening? This has already been such a fantastic episode, dive in everything. And I'll definitely encourage everybody to go and take a look at cloud37 .ai.
But anything else you want to share as it relates to AI? Anything we didn't really cover? Any other golden nuggets before we close out the episode?
One, everybody needs Cloud37. Selfish plug. Now, at the end of the day, the big thing that I try and really...
push on people. It's you need to slow down in such a fast paced world. You have to stop for a second.
Like you just, what is my business? Where is it going wrong? Mapping these things out, looking at your SOPs, looking at like everything in your business.
If you've got them, take a step back before you move forward, before you start buying tool, before you start buying into all these AI systems, even our own, like it's like at the end of the day. Yes, I want as many people using Cloud 37, but at the end of the day, it's not going to help everybody. That's just not realistic either.
But unless you really know what you're trying to accomplish, AI can't get you there. It's not going to bridge that for you. You need to know how to bridge that.
And that's just remove AI first and then implement AI afterwards.
That's great. Exactly. Get down to simplicity and then build on top of that with your expertise.
I think it's the classic rule to apply. AI shouldn't overcomplicate. It should actually simplify things when done correctly.
So I think that's a great way of looking at it and a great final golden nugget. Well, Jeff, thank you so much for coming on the podcast. This was phenomenal.
No, I really appreciate it. And always good to see you.
The Hook
The bait, then the rug-pull.
The episode opens with a statistic handed back to the guest like a loaded question: 95% of businesses are implementing AI wrong. What follows is not a tool roundup. It is a forty-three minute argument that AI has no opinion about whether your business is any good, and that the thing most people are automating is the part they should have fixed first.
Add Modern Creator as a preferred source and Google shows you more of our breakdowns in Search, Top Stories, and AI Overviews. It only changes what you see, and you can undo it in your Google settings anytime.
Add to Preferred SourcesOpens your Google source preferences with us pre-loaded. Tick the box and you're done.
Marketing Minds host Aleric Heck interviews Maria Wendt on the content system, ad economics, and offer-naming rules behind an eight-figure $27 product business.
A 27-minute conference keynote where a marketing agency founder shows the live agent architecture replacing his headcount — six named agents, one shared brain, $500K in attributed value from $2,500 in tokens.
Eric Siu opens his own AI agent stack, the bots that track demand, draft content, and prep paid media, and argues marketing judgment is the one thing none of them can do for you.
A YouTuber live-builds a five-agent Claude Code pipeline, inside his own production website, that takes a feature request from spec to merged pull request with one human click.
01:10
01:53
02:17
02:47
03:29
04:20
05:05
06:21
06:33
07:29
07:34
08:25
09:23
09:58
11:09
11:29
12:32
13:27
13:32
14:08
14:46
15:42
16:10
16:55
17:37
18:31
19:07
20:18
20:44
21:08
21:58
22:42
23:39
23:44
24:36
24:47
25:56
26:27
27:15
27:45
28:37
30:04
30:38
30:50
31:35
32:04
32:25
33:01
33:35
34:45
35:44
36:29
37:06
37:49
38:47
38:52
39:37
40:00
40:54
41:06
42:18
42:43
1:02:56
27:39
14:14
28:30
18:02
28:02