Episode 31: The Revenue Architect: Vincent Oliver on Why Your AI Agent is Overpriced

00:08

Welcome back to the Marketing Boost. Today I have Vincent Oliver from Oliver Revenue on today.  You want to introduce yourself? Yeah, so I’m Vincent. A little bit about what we do is we go into businesses who are looking to  automate, improve processes, and we implement custom software, uh AI solutions, agentic AI, uh and even some marketing solutions to help people grow their businesses. OK.  And  so  we had a mutual friend introduce us.

00:36

I talked to a lot of the founders and everything out there. There’s a subset of businesses that are popping up right now that I  both find interesting, but also kind of a, uh, the, it’s just like the marketing business. It’s like anybody and everybody low barrier to entry is AI business. Right. I really liked your take on like where AI is going and,  all of that. Like  I know that you’re building out some tools. want to kind of, yeah, sure. The, the,  the 40,000.

01:05

view picture of it first and then we can dive into it. So that’s exactly OK. So a lot of people are getting really obsessed with this idea of having  AI agents that  are in every part of your business that you give unlimited access to that can just do uh any task  in the background, almost as if it’s like hiring an employee.  But I don’t actually see the best use cases for most companies being that now people like.

01:34

myself and maybe even Leon like  like, okay, yeah, there’s there’s something to be said about using agentic AI. But when you’re talking about token usage,  unless you have a reason to spend that many tokens, I don’t really think it’s I don’t think it’s worth it personally. ah I think that what people ought to be doing and I think that me and Leon are both doing this  use AI to code and developed like

02:01

true software, custom software that we can place into companies. Right. So that companies can have software that costs them a couple hundred bucks a month and not some AI agent trying to do everything.  And it’s costing companies, you know, two thousand bucks a month because they have to spend a hundred bucks a day or whatever on it. So I really see that where people need to go is let’s not stray too far from like the basics of like computer science. Right. Like let’s not stray too far.

02:31

from that, but rather let’s use agents as a tool to write better code because code is what’s most efficient and most effective. It costs very little to keep that running in the cloud than to have an agent running in the cloud connected to everything that you do. Now, I want you to hear me. I’m not saying don’t use agents. I’m just saying for most people, for most practical solutions, you don’t need an agent to do it, right? think there’s a lot of hype with like Claude bot.

03:01

ah You know, where people basically got this like taste of  what  an agent can do. And then they treated it like it was,  you know, AGI, you know, general intelligence where it was. It could do anything for you and you just let it live in your business. And  I just don’t think that’s where we are yet. We might get there in the future. ah However, the expense that is going to cost to run and manage that is going to be too high for.

03:30

small business owners and even maybe even medium sized businesses would do better with just like real software hands on software that is code. It has a trigger.  You know, it has a dashboard where people can interact with it and you can use it for simple use cases. I like  the dashboard part because I’m seeing a lot of people all of a sudden get like consolidated reports  on

03:58

what is happening in their organization? there’s a coffee company that we work with and they were like, Hey, I need to know what’s our inventory. What’s, know, what’s out there on Amazon, what’s out there on Shopify, how much are we moving in retail? And then what is it by store? And normally that would have cost them, you know, a hundred to $200,000 minimum to get that done. But their operational efficiency just improved because now they’re able to say, man, we’re moving a lot in this, this portion of the South.

04:28

Versus somewhere in the north. They’re just like hey, we’re not really hitting on those flavors We should probably consolidate those flavors down there and move other stuff there  in the north Right.  I like the the spin that you’re talking about agents because you’re right Claude bot  one It’s really hard to set up like exceptionally right? It’s a security nightmare. It’s a mess. I’m seeing um The a 16 C and even white combinator those right?

04:56

you know, the, the great startup like spheres out there,  um, talk about these agentic enabled companies. they’re talking like marketing agencies, dev shops,  uh, tax people, accountants, lawyers, all become like software enabled. And I think when people hear that, they think, you know, an agent going and doing stuff versus probably an automation that collects the information, passes the chat, GPT real quick or Claude.

05:24

parses it and then gives somebody information versus a lawyer just saying, no, this whole brief was prepped by AI and who knows if it’s correct or not. Right.  A hundred percent. And so what we want, because you started this off by saying like, you know,  any  general marketing company now is an AI company. Right. And so what we want is we want to make sure that when we’re plugging in agentic AI, we’re plugging it, we’re plugging it into a container that makes sense. oh

05:53

right into a small environment where it makes sense for it to live in, where it has  where we can actually put detailed uh instructions where it can do exactly what we need it to do with the level of consistency. Because when you start looking, OK, let’s take uh let’s take the idea of some type of general agent that might that might be OK for small business in terms  of  security issues. They’re not worried about getting

06:22

a massive lawsuit or whatnot. However, they can’t even afford it. But then you go to big companies and the security issues, the risk of, you know, certain executions not being done properly. If let’s say 90 % of the time, the agent does whatever you need, just right business, that 10 % can land you into a million dollar lawsuit. That 10 % seen it. The airlines, the Canada Air, I believe.

06:48

They gave free tickets away and they  up had to in court uh do that because that agent was acting on behalf of the business. uh I’ve seen rental car companies. There’s been tons of cases where large companies adopted and then go, oh my, uh these things were actually communicating like contractable elements of our business that we should have had a human kind of view. And I think too many companies have veered

07:18

100 % AI even even there like I see you know, we’re we’re a geek them today and yeah, everybody’s an AI consultant But when you press them, I don’t I don’t hear the conversations about what’s an agent? What’s a skill? What’s you know,  what’s a piece of software? I hear a lot of like oh you should use this tool or you should use this this like this routine or any and I’m like, yeah But what am I gonna do with an eight and like like is it are you just using it as a beer or are using it as a?

07:47

an actual like orchestration platform. So it is a very interesting space that I don’t think people are talking about nuanced enough.  I think it’s so new that  people… Oh, it’s new and it’s a hype cycle.  Right. So like you were talking about the  large corporations. Basically, what happened was some really good sales professional.

08:13

came in, was able to whisper in someone’s ear, this is the next big thing, convinced them that if they don’t hop on it right  now,  they’re going to miss an opportunity to beat all their competitors. And then they let the thing run wild. It happened to Facebook. Look, Facebook went from being meta because they were going to go the metaverse and VR to, I think, closing down majority of their meta labs and saying that they’re going to do AI fully.  And meta has not been traditionally good at making

08:41

that that creative next bets right now. Right. But uh AI is is a very big hype cycle. I mean,  the fact that I have to say Facebook, I mean, meta shows the issue because they’re not even,  you know, we haven’t  shipped it to the metaverse yet. You know, and I mean, they they had a real opportunity to to say, OK, how are we going to architect like this digital landscape because we’re all living in some way a digital landscape. But  I guess they just can’t quite figure out how to do it.

09:10

Um, but, uh, back to what we were saying, these large companies that have the money to  pay and  maintain these agents, it’s too risky for them, you know, you know, it’s just, it’s just too risky for them to just let something go loose. And so we need kind of a hybrid model where, uh, because AI is naturally probabilistic. So I know everyone who’s watching this is  use chat GPT.

09:39

If you put the same prompt as a chat GPT  in two different chat windows, you’re going to get five different answers. Well, a business that relies on making the correct decision every single time or the correct execution every single time, it can’t survive that level of probability, right? Things need to be the same. The good thing about code is it’s very deterministic.

10:05

Right. If I say this does this, it’s going to do it 100 % of the time whenever it’s triggered. And so the idea is not, OK, how do we let AI go rampant and then figure out what to code? Let us figure out, OK, what needs to be coded? Have the AI build the code and then ship that off to the company. So really what I see has happened is the cost for a company to develop software.

10:32

has, you know, is a tenth of the cost now. You know,  really, it could be maybe a couple hundred bucks to make really good software. Now, it ain’t everyone’s going to charge a multiple. So it’s going to be ten thousand, please. Right. uh But now  companies can have custom software that’s just focused on what they do. Or here’s some of the things I’m building. I’m building what you would call middleware. OK, so  let’s say a client has uh

11:01

So  I’ll share the uh invoicing solution that I have. So a client uses QuickBooks. They can go into  my software, log into their QuickBooks, and this software lives on top of QuickBooks  and does executions  determined by the person  automatically, right? So that way they don’t have to go into QuickBooks and type everything out, as still days. They can have basically what we call middleware where

11:30

the person interacts with this  with my software and then that software will interact with all the different components of QuickBooks based on their custom needs, right?  And so what I did was I built in a voice feature where someone could basically manage uh a specific client or their book of business using their voice, right? Hey, send uh so-and-so uh a  payment request, right? Send so-and-so a link so that they can see

11:59

how much they’ve paid and what they owe. I think a solutions like that, that is, we’re really going to see is going to take over because now we can customize software. So that way it’s not just some VC funded a hundred million dollar company saying, I think you’ll like this. It could be someone like you or me having a consultation. So I was saying, Hey, I really need this and QuickBooks doesn’t have this or

12:23

I really need this and HubSpot doesn’t have this or I really need and then we can build a layer on top of it. And those have been un-economical before like  no, like that middleware.  You know, you weren’t going to spend $25,000 or maybe go hire somebody in India or the Philippines to build it and cost maybe 10,000, but you had to manage them. had to deal with them  and you had all these problems.  Um, I, I still think that there’s some compliance and security issues around there, but like when you explained it just now, it’s like,

12:53

That seems like a really cool tool and feature  over like a, you’re building on all the other quick books. It’s like, no, it plugs into the system  and you can do that. love MCP servers. Right. The fact is that I can go ask, you know, Google or Claude, what did we spend on Google ads? You know, what,  is the trends? What are the projections? What are, what are all these, these parts that normally would get dumped into Excel sheets? I would have to look at them kind of like

13:22

uh you uh

13:50

45 minute process if you’re not paying attention and you get distracted, which I will get distracted. Right. But imagine if you could just say,  make an invoice of so on so mark that is for three hours of work.  Let them know that it’s due next Friday. that have to be middleware? Could that just be an MCP server? Like, could it, could it be something like that and built in Claude? Like, do you think you’ll ever be? Well, yes, I  absolutely. But, but then we’re talking about a different type of client. Okay. Right.  So then what would be that client?

14:19

the MCPs versus the middleware one. Where oh is that delineate? Okay. So, so if I have some form of dashboard  or basically a screen interface, right?  That’s going to be for people who  consider themselves to be non-technical or less business owners. So a lot of business owners. If I build an MCP server, what that’s going to be is I’m going to basically put that on GitHub and I’m going to have 50,000

14:49

Right. Software engineers star my thing and download it and use it. Right.  We’re talking about totally different, totally different uh markets  and  even what you’re talking about. Right. So you’re talking about basically really accurate reporting uh research.  And then even if you build in some functionality where, you know, it can even investigate from you like, hey, what do you think you spent this much on Google ads? But this  ad

15:18

perform better, do you want us to  shift funds to this and that? Well,  if you offer that solution to a client who does not use Claude, they’re going to need some type of graphical interface. They’re going to want to see some type of image because they can’t explain it. The way that I can explain it is in a simple  natural language form versus.

15:44

the business owner, let’s say a nail salon is not gonna be able to go like,  is that ad where I should put the money? You  have a different sort of thought process. Okay, I see that.  And so, mean,  the thing that would be  really great  is  for people like us  and whoever’s watching who’s more of a consultant, marketer style of person uh is, okay, how do we train people so that they will just go right into

16:14

cloud code or load up cloud code in their Tylenol. And then it makes it super easy because we don’t have to build an interface. Like you said, we  literally build skill MD files,  markdown files, which costs like nothing. And these people could get  10x productivity improvements in  multiple small domains. Now, if you’re going to do agents, that’s how you do it right. What you don’t do is give it access to absolutely everything. And you say, let it run wild, unprompted.

16:44

you know, unscripted  and without even having any triggers built in.  So I do think that the goal, like  when I just gave that presentation, what I was  telling people was like, in three years, you’re all going to be doing what I’m doing right now.  Using cloud code, using it from your title, using, you know, an  IDE integrated  development  interface, right? You’re going to, you’re all going to be using this, excuse me, environment. You’re all going to be using this.

17:12

to build things. It’s just I’m three years ahead of you and you don’t realize yet that you don’t need a graphical interface. I’m excited for the time when we don’t use screens. I’m excited for screens to go away and I can just communicate with my agent. And it’s just like living in my glasses. maybe maybe one day. I’m kind of I’m I I think I’m actually more scared of that future because I’m seeing more and more people like and I don’t know if it’s right or wrong, but.

17:41

I like I have kids. You do too. The 10 year old doesn’t know what a file system is. Like I think that’s an important skill no matter what. It’s an organizational skill there that you always have to think through. The typing is, you know, something that you don’t have typesetters anymore. You know, you don’t you just type on a keyboard. Like I, I do wonder where it’s going. Like if it’s all voice, I’m like, Ooh, I don’t, don’t know if I like that future. I just use whisper flow.

18:08

Okay, what’s not it was very good. I don’t even I mean, I don’t even type because it slows me down. I’ll read because I can read I can process. I can process, you know, written word better than I can process like spoken word. Yeah, right. And so very similar. Yes. And so but I’ll talk because I can speak way faster than I can type. Yes. Voice notes are my friends. Exactly. People hate me. But once they figure out how to how to

18:37

get people to be able to uh or how to get the thing to be able to translate whatever it needs to in a way where I can hear it and it be  and I can understand it at the same  speed or comprehension  as me reading it.  I think things are going away. Absolutely. And then here’s so you should look into uh Google just put like a 200 million dollar investment in Warby Parker glasses. I where I’m really. Yeah.

19:04

They’re trying to build. I think they’re trying to put their agents in glasses. I wouldn’t be surprised. They’ve been trying to do that since what? The 2010s when Google glass glass, Robert Scoville. Um, you know, he, he tried to do that. Um, when I talked to him, maybe like a year ago, I, uh, he was launching an AI newsletter and I, I said to him, I was like, you know, the glasses, was like, what, where, like, did you ever think we would get to this AI moment? And he did, he said that, that essentially

19:32

he knew that it would happen, but he didn’t think it would be this quick. And I think that is the big part that’s happening right now.  And that’s why I say like, dude, I think we’re going to move to  no screens faster than you think.  here’s the thing, man.  think you might like it more than you realize. I probably will. It’s one of those skills, just like I hate that nobody can write anymore.  think there’s a component of you do have to work your brain.

20:01

Your brain has to work in various particular patterns and I think like writing you you retain more when you write it down even versus typing and I think screens Good or bad. It’s just I’m gonna need something to like have it in front of me so that I can just read it because if somebody tells me a bunch of words and a bunch of numbers I’m gonna look at them and go I Need to able to read this and reread it five times because I need to process it. So it is is interesting world that we’re heading on

20:32

the actual AI usage in the company’s app. oh You talked about uh building an invoicing system. You talked about some other stuff.  How are you reaching these people? How are you educating them? I’m sure that a normal business owner is not thinking the way that we’re thinking or the way the internet’s thinking or even  X or  what’s happening on LinkedIn. How are you getting to them and how are you pitching them?  So I think a lot of it, and I think this is

21:01

I think this is why even if you’re a builder, you need to have a consulting  lens. You need to kind of understand how to communicate to people because people aren’t looking for AI.  know, people are looking to do less work.  know, people are looking to make more money.  And if  I can speak to you and say, hey, you’re looking,  you’re super busy. So lawyers are really easy because they’re some of the busiest people.

21:27

uh in the country. They work like 80, 90 hours a week. And they bill all 80 or 90 hours.  Right. oh But, you know, if you could convince them like, hey, let’s cut down on the task that you can’t bill on  using AI so that way you can replace those hours with billable hours. Now they’re thinking, what do you got? Right. So I think really it’s about  messaging. that’s why a marketer

21:56

is always going to do well no matter what season of technology, what’s going on is because I’ve heard that a lot more that the marketers are actually the best position. The engineers could become the way of the dodo bird,  but I’m hearing marketers, especially particular types of product marketers that understand product and marketing and like customer communication cycles that can  really go in there  and do it. I guess you would kind of consider yourself.

22:25

rooted in marketing or? Yeah. So, I mean, I started in marketing consulting. Okay. It’s just, it just so happened that now I can create software to do the implementation of some of the things that I’ll tell you. never done. Exactly. And so, I mean, it, it’s so luring.

22:45

Even I think software is like, because I’m like super, I’m like, oh, I can make software tools for super specific problems. And so I want to float in that direction. And it’s like, no, man, you need to get back to consulting so that way you can talk to people and hear what the problems are, you know, and then maybe when you get a great grasp,  know, you have so much data from your clients,  then spin it together, make a software and then, you know, try to  blow it up. But so you’re talking a little bit of like

23:14

You being a consultant, but also having  a sassified product on the front end to some degree. like it, and I’m hearing this a lot more.  There’s somebody down the hall for me from the studio that they have uh a tool that tracks competitors from marketers. And then he was like, well, maybe on the backend, do like  marketing consulting for how to compete against those competitors, battle cards and all that. And I was like, that’s a sassified agency.

23:44

Like I’ve seen a lot more of that.  And I think, and this is what I’m noticing, it’s on the front end and the back end. Okay, so think about it like uh I can have a software tool that pays me that will also kind of be like a lead magnet, right? It’ll attract customers to me. They’ll do a free trial. And then, you know, if they actually adopt and use the software tool, well, eventually they’re going to be like, oh, who made this? And they know my name up there.

24:11

And then I reach out and say, oh, OK, I can actually do XYZ for you. Now they’re now consulting, though. But then even still, I’m going to have other software products that I know they’re going to have need of. Right. And those the software products are probably pretty niche, smaller problem solving. They’re not, you know, a 50 million dollar business, but they might be a hundred thousand dollar business. Right. Like exactly. And so, I mean, so if I’m if I’m getting, you know, let’s say I’m getting.

24:40

let’s say 6,000 a year from software on the front end to bring someone in there. They’re paying 500 bucks a month, but they get into my ecosystem and I have,  you know, I can consult them on many different things. I can do  training for the executives. can do,  you know, I can offer licensing seats for training for some of the employees, whoever they want to be trained. I can  help them make strategy and plans for like long-term implementation.

25:10

AI led growth, you know, maybe even how to satisfy some of their business processes, right? I can sell all of that. Well, I six thousand to get the customer. We would make it. The hope is a couple hundred grand on fulfilling the customer for, you know, a small business, so let’s say between a million or three. And then on the back end. Well, if I have.

25:38

three other software products that I know that they have need of.  Who’s doing your lead follow-up? Oh, I just got a sales guy. Okay. You know,  what’s his numbers? How many leads fall through the crack? Give me an exact percentage. I have no idea. Oh, oh, you need clarity and reporting  and wouldn’t it be nice if he had like a sales buddy that like could automate a lot of the follow-up so that you can, so that he could do more work and be focusing on closing deals and this thing just got him booked to

26:08

book deployments, right? Well, we can build out three, four or five. We can have a suite of products. I always think about it like it’s like the Microsoft like business model where it’s like, okay, you came in for word. Yeah. You pay for word, you get teams and all of a sudden you’re in their ecosystem. Now you’re doing you, you will,  if anybody likes Microsoft products, I feel bad for them, but like just like on the art of a space mission, Oh, I’ve got two outlooks like,  know,  and then like,

26:35

Dude, mean, but the companies that use Microsoft, they’re like the stickiest companies. They never move. mean, the fact that I still get emails from people who from like Outlook email accounts is like shocking to me. But it’s just because of like my age. I’m not 50. I’m not 60. Right. So I’m not one these old dinosaurs. But once they’re in, they stick. the reason is

27:00

Who’s gonna migrate all of this data? Who’s gonna migrate all of this information? Who’s gonna migrate all this documentation that you built up, right? Well, in the same way, like who’s gonna migrate all of this software that because I’m consulting you, I’m gonna make sure you implement and I’m gonna make sure you adopt to it I’m gonna make sure everyone in your company does. I’m gonna train them on my software. Okay, yeah, our consulting package is complete. I got you on the software. So that’s kind of where I’m going with it.

27:30

But the trap for me was I was like, oh, I built something really cool that a lot of people can use. I’ll go all in on that.  No. Well, for me at least, I don’t think that’s it because here’s what we know. In a year from now, cloud is going to drop something that’s going to make my software obsolete. Right.  And I don’t want to be one of the types of people where I’m trying to sell someone on something that they don’t really want.

27:57

where they don’t really need there. it’s like, Hey, but you got to use the software. It’s like,  I’m using cloud and it’s like, no, no, don’t. There’s a of FOMO out there and I am  sensing it from people. And this is why, like, just, like marketers that  they’re compared to use car salesmen, because most marketers don’t know how to actually implement. They’re more of, um, consultants that can’t do right. And it becomes a really problematic thing down the road as, as essentially, you know,

28:24

If you’re doing Google ads and you don’t know how to actually implement, you have people outsource. It’s like, Whoa, things don’t go well.  Same thing. And I’m sensing it out there with AI, like telling them, Hey, here’s an, uh, here’s MD skill file. Like you use Claude code and it just does it. And then here’s five tools. Oh, by the way, those five tools subscriptions cost you $5,000 a month plus Claude usage and all of that.  And I liked the way that you’re thinking because.

28:54

It’s not playing on the phone row so much of,  of it, but it is playing on the, like you need to make more money. need to save time. You need to increase revenue like, or even just decrease expenses and keep revenue the same. So it’s a different take. You you need to lay some people off so that you can make more money. mean that that’s a totally real, you know, that’s a real take. And you know, a lot of people, they have  five team members in  like, let’s say one workflow. Well,

29:24

If I could replace what two of those people are doing with some type of agent bridge or software bridge, right? uh Or just a task driven agent, not something that even needs to really think too much, you know, or, or figure, you know,  that’s another thing that kind of case at me when people are like, all right, agent, figure out my idea, figure out how to do. No, no, no.  You need to have clarity, you know, because

29:52

especially for like an AI consultant or an AI implementation specialist, like you need to have clarity. You need to know how it needs to be built and why it needs to be built and when it needs to be implemented. You know, uh because if you don’t have an understanding of that, you’re basically saying, Claude, I want you to teach yourself how to build this thing that I think would be cool.  It’s happening so much out there. That’s that’s yeah, that’s been my problem with it. It’s an interesting cycle. I love

30:21

the direction you’re going um for all of those folks that are starting to build an AI enabled company or even building consultations around AI. Like what advice would you would you give them right now? It seems like you’re  I think you’re in the right headspace and cutting edge of it. So I think what I would say as you’re building, we’re not going to kid ourselves. We know you’re using AI to build.  You’re not software engineers. We know you’re using AI to build.

30:48

What I say is I built myself a skill that I think it’s called like education K through 12, something like that.  And what it does is anytime I make something, right? So first, just start by making things for yourself, for your own business. Don’t go and say, I’m going to make things for other people. And then what I’ll do is I’ll run this skill where it will take me, it will teach me, it will actually create  a uh react. uh

31:18

a visual dashboard. uh And what it will do is it will teach me from  a kindergarten level all the way up to like a PhD level of like, here’s what we build. Here’s the functions. Here’s the actual coding commands. I use  this bash to do this. Hey, by the way, this isn’t JSON. You should know that. Right.  All of these nitty gritty things that,  I wasn’t  I’m not a trained like software engineer. Like I said, like we said, we started in marketing. uh

31:47

However, we’re gonna beat all the software engineers because we have we understand things a little bit better than that We’re gonna the term I keep hearing is like a marketing engineer, right?  Oh, I like it.  It’s it’s just like go to market, you know  Marketers like there’s so many little  things and it’s always rooted in marketing because it’s a social skill It sells and it’s ability and knowledge to know how to use the tools,  right? You know  because I mean for a solution there always has to be a problem

32:16

Right. Marketers are really good at saying what’s the problem? What’s the pain? The psychology.  Software engineers are really great when they work for like large businesses because large businesses have so much data that they know the problems and then they build the blocks. Go, go build this. OK. But anyway, build a skill, right? In Claude while you’re building, build some type of skill  where it’s going to walk you step by step of what it did, why it did it. And then bonus points.

32:44

Make the skill make you be able to uh build  some piece of it, know, uh like  type,  getting back to typing, type out some code, right? Have some type of quiz. Hey, do you know what a crawl-in function is? If you’re building things and you don’t know what a crawl-in function is, you’re listening, like you should know that it’s a scheduled trigger. Like  that’s something that you should know and it’s important for you because the more you understand the language,

33:13

the better the prompts you can give and the more you can have a watchful eye over what Claude is doing. And you could say, why would we do it like this when we could have a very simple code function? And that’s why I’m so big on like, hey, let’s get back to code. And I’m not a software engineer, but I understand like I’m not trying to spend a thousand tokens, man, like, you know, or a thousand dollars on tokens. I should say I was probably I spend.

33:39

You know, I’m going to the thousands of a day. Oh, well, I’m trying to spend a thousand dollars on tokens like once you make the process, figure out how to make it into like, you know, some type of code where it’s easy and all I have to do is just hosted on a on a server and it can just exist in the cloud. And that server cost me like 10 bucks a month. Like that’s where you should go. But you have to you have to you have to have some ability to learn. So have Claude teach you after you’ve built it. Right. So you almost

34:08

You kind of have to have some understanding of two languages, English,  right?  And like pick a coding language, right?  You know, and then have some understanding of both so that way you can know how to guide it on the front end and understand  somewhat what it’s doing on the back end. would that would be my recommendation. But  like we said, it can be a markdown file. It can be a skill like and you could just run the skill after you’ve made something. ah

34:38

Like you say, you were talking about file systems and organizations. I have a skill called file organizer or file cleaner. Yeah. So if I build too much, I’ll run the file cleaner skill and it’ll be like, OK, here’s everything. Here’s the things you can delete. Here’s the things that actually got made in in the wrong place. We’ll put it in the right place. Right. How do I want to put this? You can have all of these really fundamental. You can have these fundamentals built in.

35:06

to where it can run the fundamentals for you.  know, and so I think that’s important to know.  So it sounds really about learning, continuing education, and just like understanding what you’re putting out there. And the key part of it, just like the cron jobs, it is an important thing. You want to fire something off of what a  Docker containers are,  just figuring out those things. And that’s what people are paying you for.  Yes. Anyone who’s hiring you for AI.

35:33

automation, agenting, software, whatever.  They’re trusting you and they give you money that you understand the newest tool systems capabilities that they can leverage. So they’re getting the most bang for their buck.  You spoke about innate in earlier, right?  You can keep building innate in automations, but over time  that’s going to be  less efficient.

36:03

from building your own automations  using  Claude. I mean, Claude is dropped like, you know. It’s on every single day. But anyway, so for instance, I have a uh product, software product that I’m developing just based off of work for clients uh where it will follow up with leads. If you miss a phone call, it’ll follow up. If you send a proposal, it’ll follow up. Right? Well.

36:32

So Claude, I have an N8n MCP. And so Claude can build into my N8n and can build whatever I need. And so Claude was like, oh, let’s build it on N8n. I was like, dude, I’m not paying you $200 bucks a month to build on N8n.  dude, just write the JSON code. And let’s put it in the Cloud. then whenever we get some data, use HTTPS nodes. I mean, I hope this isn’t going over too many people’s heads. Right?

37:01

Maybe, right? But that would be a problem, right? Because these are very simple things. If you don’t know what that is, then you should use N8n because it’s almost like training wheels actually. But whenever data comes in, it triggers a workflow. Well, I don’t need that workflow right on N8n. I don’t need the visual component, right? I want this thing running in the cloud so that I’m not tied to N8n and the terms of service and, we don’t like you. A third party vendor, the price increases.

37:31

It’s a hard world out there. Well, Vincent, where can somebody find you on the Internet like website socials, all that? OK, so you can look me up. uh Vincent Oliver  on LinkedIn. You’ll see Oliver revenue solutions. can look at the notes below to you can look up Oliver revenue dot com. That’s where you’ll find me.  That’s my website. um And, you know, Leon Leon said he might help me  start off a little bit of like maybe a podcast or some type of uh social media presence. So

38:00

Whenever that comes around, I’ll let y’all know  and you know, probably be on YouTube because I like YouTube the most. I like it. Thank you so much, man.