Marcus Gaughan of MindsOnFire shows SMB owners how custom AI platforms can replace expensive software — and actually pay for themselves.
GROWTH PILLAR: AI & Automation
WHO THIS IS FOR: SMB owners / Solopreneurs / Corporate escapees / Leaders building systems
WHAT THEY'LL GAIN: A clear look at how custom AI tools work in real businesses — plus a practical framework for auditing your tech stack and cutting what you don't need.
Most businesses are paying for tools they don't fully use. Marcus Gaughan built platforms to fix that.
Marcus is the founder of MindsOnFire — a firm that helps ambitious companies replace scattered SaaS subscriptions with unified, custom-built AI systems. At this East Trade Winds session, he pulled back the curtain on what that actually looks like.
He walked through a trades company platform that replaced $28,000 CAD per year in SaaS costs. He showed a C-suite AI tool that gives solopreneurs access to a Chief Marketing Officer intelligence app, an AI practice maturity advisor, and an app rationalization engine — all for as little as $20–40 a month. He talked about lead scraping that identified 2,920 targeted companies in days. And he broke down the real cost of AI token usage — and why most platforms are overcharging.
The core message is simple. You don't need more software. You need smarter systems.
Key topics covered:
Connect with Marcus:
This session is brought to you by:
Join Canada Growth Network — Full GHL CRM + membership community. $1 CAD first month. $47 CAD/month after. No contract, no lock-in.
Try Web Indexer — The smart chatbot for the Knack 4 Business website. Find episodes faster, discover related content, and book a meeting.
Explore @Hive — Community, Collab and resource space for entrepreneurs
Questions? Reach us at info@kreativinsight.com
Join us free every Tuesday 8–9 AM EST — East Trade Winds networking
Subscribe wherever you listen to podcasts.
Marcus (00:06)
I thought that I would go over a couple of the systems and applications that we do. So like I mentioned, mine's on fire. We work on systems. We build everything from full custom app platforms to replace large SaaS subscriptions for companies.
Marcus (00:11)
we build and that we've been building.
Marcus (00:23)
all the way down to building small tools so the companies can take over a position, add to a department, help add some AI into their company without needing to go into a crazy implementation and flip everything on its head. So I'll show a few platforms that I thought...
We are first of all, I'll share the one that I mentioned
So these are a few platforms that are still in development. This one, I we've got a little coloring being thrown off, but we have a lot of companies ask us, know, everybody wants AI, but nobody knows where, when, how, what the best form is, how to deploy it, that kind of piece. So what we thought of doing is building a C-suite for people. This is one.
that we've been working on for about three months. We have our Chief Marketing Officer Intelligence application, give you full campaign analysis, channel optimization, funnel analytics. This is something that can both plan what your marketing structure should be, your operations, your systems, what people you bring in, as well as give you the tools. You know, what systems should you be using? How do you find leads? How do you track engagement? What's the best flow to put together for your business? We have built
another call the app rationalization pro. This is for companies that already have a tech stack. This can be anything from your small mom and pop shop or service company all the way up to an enterprise. You connect with this application. You talk to the AI about what apps you're using, how you're using them. It analyzes your system to go a little bit deeper. And then it will suggest for you.
what applications you should get rid of, what ones you should buy, which ones you should integrate, raise or lower a tier. So this is an application that in a few minutes, you can figure out what tech stack you actually need rather than waiting for an ad to pop up and keep dragging you into this subscription and this one and this one and this one. We have built one for AI practice maturity. This is to practice with AI.
Same deal. This one will help you figure out how to implement rather than your general systems, AI into your business. So, you know, it may feel like chat GPT is what you should put into everyone's hands, but maybe you should be looking at agents. Maybe you should be looking at building your own application. Maybe you can look for sources where you can build a small app that does what you actually need rather than needing to go for a large all in one program. This is an app that you can talk to to give you insight on that.
and research to help you figure out what you need to buy or should buy. In terms of the larger work that we do, this is an application that we built custom. If anybody has ever worked with trades companies, know, anybody in either a single trade or multi-trade, may have heard of the platform Jobber. There are a lot of platforms out there that help trades companies just manage themselves, manage invoicing, customers, things like that. But a lot of these are very clunky.
Jobber has a lot of good things in it, but it has no AI embedded. It doesn't allow you to have multiple divisions. There are several restrictions that at least this company was really feeling as they grew to being a multi-division trades company. So we built this application. This is a demo that we put together in about a week. We've modeled it after Jobber.
giving them the scheduling that they like, clients requests, quotes, jobs, everything that they would realistically need so that they can continue working like they did with Jobber. But then we've added in a lot of powerful features. The biggest one being this AI assistant. Now this AI assistant is not your typical chat bot. This is not one that has a few pre-recorded responses or tries to get someone's email. This AI ⁓ is trained on the entire application.
So we have some dummy data in this. Say for example, if I say, tell me what invoices need to go out this week. It won't just say, we should go to invoices and find them. It will actually start going through your business data completed jobs this week. And the formatting is a little off, but you know, job 012 that was showing is completed your parking lot, lighting retrofit that's showing finished. Here's the amount that is your final billing.
This is an AI that is trained on all of your data that you own that this company will actually own since they will own the program once they've given it to them. So they're not handing all their data to ChatGPT or Anthropic or Perplexity or training their models. And the further piece of this that makes this amazing when it's finally done, this AI will actually be able to act on the site. So if you come to it and say, what are the invoices I need? And it gives you a list of 20 that need to go out.
And then you follow up with, okay, draft those invoices for me and get them ready for send. It will do that. It will build the invoices, add or subtract line items, and then find the contact information that needs to go out to the company. If there are multiple, it'll let you choose. It's very conversational. So we build platforms, like everything from tools to help individual businesses to something that could replace, I mean, this system is going to replace.
$28,000 Canadian a year that this company is spending on SaaS payments every single year for data that they are essentially giving to job or to do with what they want. So that's a big piece of what we do. We build tech that helps companies digitize themselves, like helps them put in the systems that they need from marketing, where we use say, here's a cool platform that we've been using. We have a client that is also in the construction space.
They are trying to find their clients where they're already showing engagement. So what we did with this client, we put them onto a website and draw analytics from. We did research into their ICP, where their customers were going, what they were interested in. And what we found is the most engagement with these trades companies was with technology companies. They're all looking for the next tech and the next thing that they want.
So we set up scrapers on profiles using a program that we have built over the last three months. We turned it on and came up with 2,920 companies that were identified interacting with the pages that from our research shows is similar to the space that they're in. And we'll now over the next few days set up a go-to-market automation so they can pull all these leads out, score them,
research them, pick out the ones that actually match their IP, and then put them into an automated email list. So if they want, they can entirely automate from top of funnel to sale, going after thousands of companies in the trade space. So we love to play around with data. We love to give people the tools that they need to accelerate their businesses. And we like the where we've really been pushed into.
Cause we've been using AI and deploying it for about two years now. We got a little ahead of the bandwagon on this since we were already building applications. But what we find now is everybody struggles with what AI to put into their business. You everybody's being bombarded with the best this, the best that, the newest this. And we are trying to build apps to break through that rather than us just following what the new coolest thing that we see is.
We would rather help people research around it and find what's best for their business. know, cut out a lot of trial and error, months of not having the system you want, paying hundreds, if not thousands of dollars testing out different applications. That's what we're really trying to do. We're trying to make technology, apps, AI accessible for companies in a way that hits ROI right away. So you don't just feel like you're developing a hobby playing around with these kinds of tools. So we...
We have a lot of fun. I mean, we are building a lot of applications right now. We have another one coming out in landscaping in two months. We're building in FinTech. We're building all over the place. So I, I would love to talk to anyone here who either, I mean, it could be yourselves, but even if you're curious for your family or your friends or other businesses, I mean, we are more than happy to help people find where they need to go.
so because of the breadth of functionality of this, we ask people, what is going wrong? Like how is business what's going on when we talk to the plumbing company that we're working with now that, and I knew one of the owners, so I kind of knew the problem over a bit of time, but because of the tech that we can build and how quickly we can spin this up, it's not really about
we'll do your invoicing and get in the door. Like there are a few that we're starting to lead with, but it's more how is business, what is going on? What, like tell me about it. And they will say, I need clients. I can't get ahead of invoicing. follow ups are going to the wayside. There will always be something that they know they should be doing better or could do better at. That becomes our beachhead. Like we look at how we can fix that piece.
And we either just fix that or we start bolting on other pieces that can give them whatever else they need. I mean, it really depends on the situation. We we've played around with so many different systems and even use our own tools to research what to do in industries and what the best to do is that it's kind of the opposite. It's like, how are you? What's broken? What's wrong? OK, here's the set of stuff we need to fix it. Does that work for you? Do you like it?
there's big companies that want to do big tech implementations and AI and all that kind of stuff. And so we were having fun building these larger platforms, but they're even difficult. Like they're they're long. It takes a lot of people like they're it can be months while we're building a project. So thinking about
making it easier on us and making it easier on businesses that are not going to come in on a large program. That's what we build the C-suite for and the AI practice and the digital rationalization. there, if someone comes in and says, you know, I'm a mid-marketer and enterprise, I've got a hundred systems and whatever, it can handle that. But we're putting in very low tiers of this, like, we haven't priced them out yet, but we're looking at putting these things out for 20, 30, $40 a month.
So that if you are a solopreneur, you can come in for a small subscription and it can help you research. How do I put in one agent that can extend an employee that'll only cost me 50 to a hundred dollars a month rather than hundreds of thousands like that. That's a great question. That's basically exactly why we're building these platforms is so that we can put them out to smaller companies. You can play with them research, have it give you a very pointed, streamlined direction towards where you can build your business.
It could suggest those agents. could suggest a marketing system. It could suggest a lot of different things that you could do. And my favorite out of this is that digital rationalization pro because when that one comes in, sure. An enterprise could save 50 grand a year or whatever and be some big crazy number. But for a solopreneur, if you can cut $180 out a month.
That can be huge. Like that can get you into profit this month. That can give you that extra bit of cash that can give you money to go and do a few more networking events or things like that. So I hear you. That's that question is exactly why we built the C suite and the AI practice and the digital rationalization. It could be a suite that if you bought everything 150 bucks a month and you literally have an enterprise grade C suite of executives to talk to all day, plus an AI researcher, plus something to help you with your tech stack.
We're really trying to build those core components so that, yeah, you don't have to hire a bunch of people and drill yourself into the ground, but you can still have all the modern tools.
in terms of coding language is changing, it's not a concern now because of AI on two fronts. Like first of all, AI is coding a lot of this. This is a combination of our developers who actually understand code and structure of apps to vibe coding. Like we have accelerated our production by, I don't even know, at least 20, 30 times as fast as we were before. So because of AI, there's no way that this will go out of style. If there's a new coding language in,
Marcus (12:11)
No, right.
you know
Marcus (12:21)
I know that the coding teams have gone Florida, Texas, Vancouver, Southern Ontario. I know they'll all be able to figure it out. Cause we start with developers, like people who could do this from the ground up from scratch. Now with AI, it lets us reduce a ton of costs with us building multiple applications. We ended up being able to take bits and pieces from them and put them into other applications. So staying current is not an issue at all.
And then because of us building these ourselves, we can fix that problem of, you know, is all of your data going to the metas of the world or open AI or whatever. Now we have two options for these. The data goes to us. If someone just wants to rent this because we've built this ourselves, this is not built on top of anything. It's not on top of Claude. It's not on top of open AI. So they're, aren't eventually getting the data either way, but then the added benefit.
And this does come with a bit of an extra cost, if it's what someone's looking for, it's extremely worth it. And since we own this and built it, we also have the ability to deploy this on people's machines. And this is what we're doing for that plumbing company. There are next meeting with them after this week that we get a bit more of the demo done and feature planning and road mapping is talking with their IT department.
Because we literally want IT to be the ones that we're deploying this with. We want to ask them, what server do you want? Where is your data storage? What backup do you want? So we're hooking that up within their organization so that if we like, I mean, we will leave eventually. The plan is for us to help them turn this into a SaaS product and turn it out to the world and sell it. So this is a long engagement, but theoretically,
After this is done, we hand them a report on how to build it, the code base, we hand them everything that they would need and we walk away. Like we, there's time in future, I don't know, when we're some crazy Amazon level thing to start mining data and all that kind of stuff. But we know where the market is right now. Everybody's gotten wise to the fact that their data is being absolutely like thrown around, bought, sold, switched, scraped, absolutely everywhere. So we were just trying not to follow that path.
I mean, we have ways to deploy this into companies. Like we can truly give it away.
how you manage your token usage, because that's typically what's happening. You see a lot of applications now will give you the membership fee and then a credit fee. So that, that's a quite a big gray area. Like when we, if I had, you know, four to six months ago when credit systems rousally everywhere and coming out.
Marcus (14:38)
you
Marcus (14:52)
When I looked at it and saw that it was a dollar credit, $2 or credit, these crazy prices, I kind of took it at face value and thought, wow, like, okay, you AI usage must be insane. Like there must be so much cost to it. Once we priced it out, as we start building our own systems and betting AI, 40 users on this job or replacement platform using it all day would cost about a dollar 15 a month per person.
So about $40 a month that you would need to run an AI platform that is now a fully trained, learned intelligence platform across your entire business. So there are some of these because when you do get into token usage, it's not a token like I ask AI question, that's one token. It gives me an output, that's one token. It breaks it down. It depends on the AI model. It can be.
per letter, can be per chunk of words, it can be syllables, it can be based on the amount of electricity it takes plus the time that it takes plus the number of characters plus plus plus. So there is a lot that goes on behind it. And because of that, there's a lot of strategy that goes into chunking your data. Like when we actually build the AI, we can choose how much data will go in at one time, we can restrict outputs, we can restrict inputs. So basically you have to get into a lot of
planning with that AI usage, but it's not as much as people think. I, yeah, if you ever get a company that's like, here's your free hundred credits and it's like a dollar each or something like that, that's like highway robbery. Like they should be selling them for like five cents each and still make a profit on that. it's be weary of that usage is real, but it's not that bad. I mean, you can even rent computing power out of a data center for, think it's
$6 an hour and have enough computation power to run about a 100 person company on all the AI that you would need. Like there definitely think twice before just buying into a platform. that's all I can really say. There's a lot of complexity behind it. Even I can't go as deep into as I would want. Like I'd need a developer or one of my guys here to talk into that. But yeah, we are trying to stay away from that generally.
I mean, if it's an enterprise solution, we'll probably have a credit, but we're not planning on putting that into any of our solarpreneur small to medium business applications. It's just not enough.