Stop Chasing the Newest AI Tool | Sam Fielding

[00:00] Cold open

Sam: You can see how it thinks and what it's capable of. It's super scary to see where it's headed.

Kane: There's so many issues, like we've got add-ons on add-ons. There's integrations to pull reports because ServiceM8 isn't the best.

Sam: Businesses that are completely run by AI with no human involvement whatsoever. It's not that far away. It's on the horizon at the moment.

Kane: I do think the whole human connection thing is going to go the opposite way to AI.

Sam: Guys that get quoted 10K for a website, and I can build the same quality website in 60 minutes for a tenth of the price.

[00:35] Welcome Sam Fielding

Kane: Hey, I'm Kane Tyler Smith, the founder and CEO of Smartfix and Hypotential. This is Unlocking Business, where we have real conversations around business, leadership, mindset and personal growth. If you've been getting value from these conversations, please take a second to like, subscribe and follow our channels. It helps us reach more people, bring on more guests and share more stories. Now, let's get into this episode.

Kane: Today, I'm joined by Sam Fielding, managing director of Echelon AI Solutions. Sam and his team work with established businesses to move AI beyond simply using tools like ChatGPT and actually integrate it into how the business operates. Connecting systems, data and workflows to automate repetitive work and give teams back time. AI is probably one of the biggest opportunities for businesses right now, but I also think there's a huge amount of confusion around what businesses should actually be doing with it. Thanks for coming on the podcast.

Sam: No, thanks for having me.

[01:02] Are businesses just scratching the surface?

Kane: Are most businesses using AI properly, or are we just barely scratching the surface?

Sam: Yeah, I think 100% just scratching the surface. A lot of people get a chat subscription, or maybe even go as far as Claude or Gemini, which is becoming pretty popular now. But they'll get a subscription to a tool and think that's integrating AI into the business. They'll get it to answer some emails. They might copy and paste something from Gmail, chuck it into ChatGPT to rewrite it and then put it back in the email. And that's their thinking of using AI.

Sam: Definitely the surface. I think there's a lot of hype around it but not much practical implementation on how to use it. I think people get super stuck in chasing the newest AI tool and thinking that's going to be the change or the magic switch that's going to take their business from where it is now to some miracle AI business. So yeah, 100% scratching the surface.

[02:04] ChatGPT, Copilot or Claude?

Kane: Yeah, that's so true. There have been many conversations we've had around what's the best platform to use.

Sam: Yeah.

Kane: I originally started using Copilot because that's part of our team subscription anyway, and everyone's talking about ChatGPT and it's so good. So I moved to that, and then it was Claude. I'm like, all my projects are saved in ChatGPT, I don't want to have to change it. So I didn't change. What's your opinion, is Claude the best at the moment?

Sam: I think once you switch to Claude you won't go back. But then there's that learning curve you've got to get over. People get their project set up and then trying to switch over is like a whole new process, so people just suck it up and stick with ChatGPT. But I think if you can make the switch to Claude, you won't go back. It's so much better.

Kane: So what does your business do? How do you integrate AI properly?

[03:06] What an AI operating system is

Sam: Yeah, so we have two main engagement structures. The first one is just AI engineering, so clients come to us with a specific automation or something they want built, and then it's a simple development and implementation. The second one is the AI transformation process, which is my favourite and the most popular.

Sam: What that is, is we take them from AI-absent, or wherever they currently are, with inefficient or broken processes, whatever it might be, to AI first. And we do that using an AI operating system. Basically, an AI operating system in general is a unified workspace that they then run the business from. So rather than logging into Gmail or Google Drive or a Word doc, whatever tool you might use, your CRM, you connect it all up to this operating system. Then you open that window, talk to it in plain English and run your business from there.

Sam: It's got two parts to it. The first one is the intelligence layer, where it knows everything about your business, all your processes, your team and so on. It's like talking to a co-founder, or yourself. The second part is connecting it to your tools. So it's connected to your email, CRM, drive, Google Doc, Word doc, whatever it might be. Then it has access to real information. It can self-improve and self-learn, but it can also do things on your behalf.

Sam: A lot of tools like ChatGPT or Claude are great. They're super intelligent when you ask them to do something. But when it comes to actually implementing it and doing things on your behalf, there's a bit of a barrier to entry there. So that's the intelligence layer. You can talk to it, it knows stuff about your business and it can do stuff for you. And then obviously the automations on top, which everyone's familiar with. But I think having the combination of the two is the way to go.

Sam: You can chat to it in plain English. You've got that intelligence layer, whether that's asking what your P&L is, how many leads came in last week, marketing performance, or it spotting which channel is more efficient. And automations are pretty self-explanatory, you automate different tasks within the business. So having the combination of those two is what we've found works best.

Kane: So that operating system you're talking about, the AI operating system, is that what you build, or is there one that you use?

Sam: Yeah, that's what we built. We actually worked super closely in the beginning with a guy called Liam Odley. He's the grandfather of the AI space, and it was through him. He actually developed the whole AI operating system concept. So we worked closely with him, spent a bit of time in Bali with him and got the operating system from him. From there it's just been development ourselves internally. Working with clients, you pick up what works best and where there's a limitation, so it's a continuous development process from there.

[05:11] Custom integrations, APIs and webhooks

Kane: It sounds interesting, because my business Smartfix, we all have the CRM, we've got ServiceM8, obviously we've got Xero.

Kane: We use Podium.

Kane: All the social media platforms. So that operating system, you're saying that could potentially pull all the data from all those platforms into one?

Kane: Is that through something like Zapier?

Kane: And webhooks and everything?

Sam: So custom integrations is what we use. Zapier was great when it was first out. It was ahead of its time, and then it got replaced by Make, and then n8n came along. Now we just use APIs and custom integrations by themselves without a third party tool. So it's using an API, and then you wrap that in a Python script so it can call the API, it's secure and it pulls all the data properly. And then you get some tools that don't have an API. So then it's a custom, or our IP, that we use to connect it up and pull it out.

Kane: It sounds scary, because straight away, I'm not super tech-savvy. I'm like, what if it misses data? How do you sell your services to someone like me that doesn't fully understand, which is obviously the whole point of your business and what you're trying to do? Are there ways it can miss data or get data incorrect?

Sam: No. The APIs themselves are sometimes limited in what you can pull. ServiceM8, for example, doesn't let you pull any staff or team member data because it's classified as personal. But the rest of your data, like your job number and revenue, things like that, you can pull pretty easily. So the limitation isn't that it pulls the incorrect data, it's what data is available through that API. And if it's available, it's a pretty simple pull. It's just pulling data out of it. So you can't really get it wrong. It's more or less what you actually have access to and can pull out of it.

Kane: Is webhooks a different word for API, or is that something different?

Sam: No, that's something different. The API is literally just pulling data and also writing back to a different tool, and the webhook is like a trigger. It'll get events sent to it. A good example is we have one set up for our invoices. As soon as an invoice is paid, we use Stripe, so they'll pay through Stripe. Stripe will then send data to our webhook that says, hey, new payment received, and that triggers a flow which sends out a welcome email saying, hey, thanks for choosing us, thanks for your payment. So the webhook is basically just receiving data and then used as a trigger to set off different automations and workflows.

[08:20] AI operating system vs Go High Level

Kane: Okay. So that process you just explained, with the Stripe payment then triggering the flow. How different would the operating system be to something like Go High Level, that can have similar effects?

Sam: Yeah. Go High Level is just another software platform within itself. It's got heaps of AI features and it can do a bunch of different stuff, but it's still disconnected from the rest of the tools you use in your business. No matter how good they get, they can keep adding more integrations and more stuff, but there'll still be limitations to what it can connect to. And more importantly, what sort of data it can hold. Obviously it can hold all your lead data, and if you use it for job management or whatever it might be. But then there's going to be certain things, like maybe your P&L or your staff list, who handles what task in the business. There's certain information that it can't hold in there, or you could, but it just wouldn't be built for it. So it's a great platform in and of itself, but there's still going to be limitations as to what it can do across the whole business.

Kane: Yeah, we have issues like that all the time. We use ServiceM8 for Smartfix and there's so many issues, like there's add-ons on add-ons.

Sam: Yeah, exactly.

Kane: There's integrations, like with Winks, to pull reports, because ServiceM8 isn't the best. But yeah, we're chatting this week, there's so many things that we don't have data on.

Sam: Yeah, exactly.

Kane: And if you don't have the data, you can't make the correct decisions.

Sam: Exactly, right. And then the platforms that do offer the data through additional add-ons, obviously it starts to add up the more and more you buy.

Kane: Salesforce.

Sam: Yeah, exactly.

Kane: Build it all up.

[09:21] Building your own software and CRMs

Kane: Do you think there will come a point where you could potentially have entire operating systems built, where you're pulling data from Xero, ServiceM8, Stripe, whatever else? Do you think there'll be a point where we can make entire platforms custom built, or is that going to be still too expensive?

Sam: Yeah, 100%. And I think that's where the operating system is heading. We've already got a lot of clients. I was actually speaking to one yesterday, and when we first set up for these guys, they were probably one of the least technical guys we've set it up for. But speaking to them yesterday, they've built their own clock on and clock off software. So even if you're illiterate when it comes to technology, it's super easy to build your own software. I think we're headed on that path, and you see a lot of people building their own CRM through Claude or their own apps. So I think it's not as far away as people think. It's already easy, but it's getting easier as we progress.

Kane: Sounds scary, even just saying people building their own. I've got a friend who has been trying to build his own app for a while, and the one he's trying to build sounds so similar to something like Employment Hero, which is another thing that we use. I'm like, you're spending so much time on this. I don't know, it just scares me, because I think people will miss things. If I was to build one, I don't understand all the data that it should have, so how can I then build it?

Kane: But I suppose that's where people like you come in. We've chatted about this before, there's so many people out there at the moment that are AI experts and AI integrators. To me, again, not tech-savvy, I don't know the right questions to ask. So I just don't do it. It's easy for me just to leave it and continue doing what we're doing. But we're probably missing out on a lot of automations.

Sam: Yeah, I think so. But then it gets down to, some people decide to go and do it themselves, and then they're building all these cool software and cool apps. But it gets to a point where, are you actually saving any time, or are you just spending time building stuff for the fun of it?

Kane: That's sort of what I was getting at. I think he spent two or three weeks building this thing, and you just pay like 30 bucks a month to have Employment Hero. And it's not as good as what you've built, but yeah.

[11:25] From marketing agency to AI websites

Kane: So where did you, what did you do before you started this business?

Sam: I had a marketing agency. That was my first business, like probably a lot of young guys starting in the online space. So I did that, and then obviously ChatGPT came out. There was a lot of hype around that, and I just started experimenting with it. Somewhere along the line I got sucked into it and decided I wanted to do it as a business. I've always been obsessed with systems and the engineering side of things. So I started playing around with it, really started to like it and then turned it into a business.

Kane: So you worked at a marketing agency?

Sam: No, I had my own. I worked for myself building websites, you know, the typical Facebook ads, SEO, Google ads sort of stuff.

Kane: I had a look at your website. It's pretty cool. Did you make that with AI?

Sam: Yeah, I did.

Kane: Yeah.

Sam: It's a good story actually. When I was doing my marketing agency, when I first started, I had a team over in Nigeria. So you get the client, get the brief, show them the website, and then I go to my team over in Nigeria and say, hey, let's build this website. That would take anywhere between two to four weeks, and then an extra maybe two to four weeks with the edits going back and forth with the client.

Sam: Now, through my AI, I type forward slash new website, give it the client name or the project name, and it pulls all the relevant details from our email conversations or our call transcripts, wherever it might be. It finds competitor websites or references that I give it and then just goes away and builds the complete website for me.

[12:45] Does AI-written code need checking?

Kane: That's wild. Do you have issues? I could be wrong here, Aisha can correct me, but she made one through AI and I think it was more of the user interface. We still got someone else to build it because there were issues with Claude building all the code. I don't know how to even ask this, but do you have to check what it builds, or are you good to go with whatever AI makes?

Sam: I think with websites, the code in most cases is generally pretty simple, so you don't have to check it. And now obviously Claude's code is getting pretty good. Even with a lot of the other coding projects we do, I don't check it at all unless there's an error. For the bigger projects, we've got a proper certified developer that does have a look. But for something as simple as websites, I'd say just talk to Claude, get it to give you the website, and you're good to go without looking at it.

Sam: If there are some issues with the code, it'll be pretty evident. Maybe the website will load slow or some people get an error when they're opening it up. In that case you just say, hey Claude, I'm getting this error, this is happening, fix it for me, and then it can go through and sort it out for you. So it's pretty intuitive and good in that aspect.

Kane: So do you think the general public could do something like a website?

Sam: Yeah, 100%. When I first started doing the AI websites, there were tools like Bolt and I think v0 or Lovable even, and that is super simple. Any Joe Blow off the street could build their own website. You just open it up and say, hey, build me a website for this company, and it goes away and builds it for you. It's just like chatting to ChatGPT or Claude. You go back and forth, say, hey, I want this, hey, I want that, and it does the coding in the background.

Sam: Now obviously Claude is just a smarter model. So if you build it straight through Claude and then host it on something like Vercel, it gets you a better output. That does require a little bit more technical knowledge, but even then it's still super simple to build it.

Kane: It is. It's scary.

Sam: Yeah, 100%. I think people know AI is scary and where it's heading, but I don't think they fully grasp how scary it is and how fast we're moving. Some of the stuff it can do, and the access we have to ChatGPT and Claude is a consumer version. It's got safeguards and guardrails in place, so we don't have the full capacity of it. Then you see what the frontier models, or companies like Anthropic and OpenAI, are doing, and you can see the reports they're putting out, and what it's doing, like breaking out of these safe environments. When you piece that together with some of the stuff you do with it, you can see how it thinks and what it's capable of. It's super scary to see where it's headed.

Kane: Yeah, it is scary. That's stuff that probably the general population don't see. I suppose someone like you would look into that and understand what it means.

Sam: Yeah, exactly.

Kane: I might not look.

Sam: Yeah.

Kane: Buried your head in the sand.

[15:37] Xero invoices and reviewing AI output

Kane: What's something that businesses are doing manually that you think is crazy?

Sam: I think probably any sort of reporting at all, or manual data entry. I think manually entering bills and invoices into Xero is probably the main one there. Everyone has Xero, everyone has invoices, and then spending, whatever it is, five minutes per invoice to upload it into Xero. I think that's a super simple task that can be automated, but there's a lot of people still doing it.

Kane: Yeah. On that one, I think this week I got a notification in Xero about auto reconciling. Again, I'm like, can I trust it though? We use Hubdoc.

Sam: Yeah.

Kane: We send the invoices to Hubdoc, which is an integration in Xero, and then it sends it to Xero. But I literally noticed yesterday there were two invoices that were incorrect. I don't think it was from manual input, but it pulled the wrong job number or invoice number, so we had a double up of the two invoices. If it's not a human that checks that, I don't know if AI can pick that up, but if I didn't see it, I could have potentially double paid these two invoices.

Sam: Yeah, 100%.

Kane: I think it's just trial and error. I think a lot of it goes to trust. I just don't trust that there's going to be no mistakes.

Sam: Yeah, exactly. I think you've got to put a lot of safeguards in place and not trust the output without reviewing it first. For example, with Xero invoices, I make sure that clients are never just letting AI go and do it itself. It can upload it and put in all the details, but a human still has to go through and approve it, because a lot of times it'll get an invoice number wrong, or have a duplicate, or a due date or something simple like that, but it's still wrong. So I think you should always review the output from AI before going forward with it, whether that's an invoice in Xero, an email, a PDF, whatever it might be.

Kane: Or a LinkedIn post.

Sam: Yeah, exactly. There's a lot of AI slop out on LinkedIn at the moment.

[17:42] Spotting AI slop and AI tells

Kane: I think that's the most important thing, what you just said, you have to review it always, no matter what. We obviously use it a lot in business and there's always the extra space before the start of a word. I think if there's a tell like that and someone sees it, I just don't even read it. I know it's not been checked and it's just annoying.

Sam: Exactly. They're starting to get a lot of different AI tells, like the "it's not X, it's Y" sort of thing, and the em dash is a big one. I actually saw something yesterday. Someone made a presentation and they had a disclaimer at the bottom saying all em dashes in here were human typed, which I thought was quite funny. A lot of people that used to use em dashes or some of these AI tells are now having to change the way they talk. There was an event yesterday and there's a lot of journalists there.

Kane: And they're devastated, because we used to use em dashes but now we can't.

Sam: Yeah.

Kane: It is a bit of a shame.

Sam: Yeah, 100%.

Kane: Even for me, I never used them but now I do. Actually, it spaces the word out and makes it look different. But I don't know if I should use it, because it's going to look like a...

Sam: Yeah, exactly. And there's a lot of words too, like "game changer" or "superpowered", whatever. The words that they use. So people that use those are going to have to change as well.

Kane: You can set your prompt up, don't use these words, and you list out all the words, but then...

Sam: It's going to change.

Kane: It's so hard.

[18:44] Do you need systems before AI?

Kane: Yeah, definitely. So how important is it to have good systems in place before you try and implement AI?

Sam: Actually, I think people think that it is super important. They need to have their ducks in a row before they do it. But to be honest, it's not that important at all. The clients that have the worst systems and data in the beginning usually get better results at the end, because it's such a big contrast, the before and after. And if you're setting up something like an operating system, your processes have to change anyway. So if you systemise and set it all up first and then you set up something like an operating system, you're double handling. You've got to change it and set up your systems twice.

Sam: Whereas someone that has no systems, their business is a complete mess, then going through something like an AI operating system or setting up AI internally properly, they generally have the better results. And I think it levels the playing field a bit. You've got someone that might be five years ahead of someone just starting a business, all their systems are set up properly, everything's super dialled in. Then you get the person that is just starting a business, everything's a mess, they've got no systems whatsoever. For them to catch up to the systemised person, it's going to take them years, and they may never get there. But now with AI, that gap has closed. If the person with no systems can take advantage of that opportunity and systemise their processes with AI, then they can catch up to their competitors that have been doing it for years.

[20:15] Switching CRMs after hundreds of thousands of jobs

Kane: Such a good point. I had a conversation yesterday, because going back to ServiceM8, there's so many things that it misses out on. I've got a friend who keeps telling me to change to a better CRM. I don't know if you've heard of ServiceTitan.

Sam: Yeah.

Kane: It's from America and it's sort of blowing up over here. But obviously because we've been in business so long, we've got hundreds of thousands of jobs in this system. It's such a scary and expensive process swapping over.

Kane: I think what you just said is important, because if you put the systems in place early enough, potentially with AI, there's a chance you might not have to switch later on.

Sam: Exactly.

Kane: Because you would think that operating system that you build just gets updated, as opposed to having to replace it with a new system. Is that right? That system updates as opposed to swapping it out?

Sam: Yeah, exactly. And it self-improves as you go, and it's a neutral store. A lot of people get stuck on using ChatGPT or Claude and think that the tool is the way to go. But in actual fact, implementing AI properly just means having proper systems and processes set up that can then be used across different AI tools or different CRMs, whatever it might be. So it's having the data in the business set up in a way that you can move it across easily to different tools and different CRMs.

Sam: But people like yourself that have been in business for a long time and have all that data, it is quite a hard move. So I think starting from scratch with nothing, you can come in and build your own CRM and software up from the ground. It doesn't have to be a massive software with all the features to start, as long as it can do the basics, and then you build as you go. And I think in five years' time, people that are just starting now and building their own CRM or systems will be ahead of someone that's stuck using Salesforce or a ServiceM8 that hasn't improved with the times.

Kane: So if someone was starting a fresh trade business, is it an option for them to work with someone like you to build their own CRM from scratch, or do they still need to choose one and then have the AI operating system alongside it?

Sam: If they've got no CRM in place, then I would just make the jump and build your own straight away. If you're getting a fair bit of lead flow coming through and you've already got somewhere set up, or you may not have somewhere set up, then you'd want to make sure that if you are building your own CRM, that first MVP is good enough to hold that initial data. Whereas someone that is just starting out, maybe like a month old, with 50 leads or whatever it is to put in there, you can start off with something that's pretty much a flashy spreadsheet and that'll do the job for you. So yeah, I think the less data you have, obviously the easier it is. I think it's worth trying for most people if they're starting out.

[22:56] How Echelon audits a business

Kane: So what would the process look like if we were to get you into our business? We've already got loads of different systems in place, like CRM, Xero, like I mentioned before. What would be the first thing that you would do? Do you have to do an audit of everything that we have, and then I suppose have a proposal plan as to what you suggest?

Sam: Yeah, exactly. It's usually a three-step process. You start with that audit, and you map out literally every task within the business, who does what, how often, what it looks like, down to a granular scope, like step by step, logging into this tool, then doing this, passing it over here, and so on. So you want to make sure that audit is as detailed and documented as possible. And then you get a list of tasks. Say we found 50 tasks in your business. 10 of those aren't automatable at all. They may be phone calls, judgment calls, whatever they might be. Another 10 are partially automatable, so you can't automate the full thing, but there are parts that can be automated. And then another 30 that are completely automatable.

Sam: So we find all of that and do a roadmap saying, hey, this is what we found, this is what can and can't be automated and this is where we recommend starting. We give that to you as a menu. The operating system is the foundation, and then all of those automations are an add-on on top of that. So you can start off with just the foundation and then build it out yourself, or you can go the full transformation and get all the specific automations on top of that. We give that to you as a menu, then go away, develop it and implement it. And then teaching your team how to use it, because that's a big thing. You can spend all this money on a great system, but if no one uses it then it's pretty much garbage.

Kane: Yeah, that was my next question. So you obviously do the training on how to use it.

Sam: Yeah, I think that's a big one. A lot of people get stuck in their ways, and when something changes it's hard to get them across. But what we've found is we haven't had too much of an issue with that sort of change management. If the tool and the system is good enough and it clearly saves people time, then they're more than happy to jump across. It's a pretty simple switch.

[25:01] Saved hours vs the bottom line

Kane: Do you find in those scenarios that the managers and business owners are going to find other tasks for that person to do, because they can clearly see how much time it's saved?

Sam: Yeah, 100%. And it's quite an interesting point. Yes, you can bring in all these automations and save a lot of time. Say you save 20 hours a week, that's great. Say your admin person has saved 20 hours a week answering emails. But what you do with that 20 hours is the part where it's whether or not AI actually makes a meaningful impact in your business. You can save those 20 hours and then get that admin person to go away and do another task, which is great. You've saved the 20 hours on the emails, but you haven't actually really affected the bottom line, because those admin hours, her wage is still getting paid doing another task. So you may get more done and it may increase the revenue, but there's no actual cost saving.

Sam: On the other hand, you can save that 20 hours and that could be 20 hours less that that person has to work that week. That's when you actually start to impact the bottom line. If you're building out a couple of automations here and there, you save the time on the initial task, but the resources are still staying within the business. You've still got to pay that wage and you're just allocating it to a different task. So it's not necessarily affecting the bottom line.

Sam: Whereas if you do a full transformation from beginning to end, where the whole business gets automated and goes AI first, that's when you start to see people getting laid off and the business actually transforming in terms of team size, whether that's getting leaner, getting rid of a bunch of people, or bringing new people on to manage the new processes. So that's obviously a big scare factor for a lot of people, as we've heard.

[27:10] Job losses and AI monitoring roles

Kane: I know a lot of people are trying to defend it and saying, well, that's where the other staff will do other tasks, or you still need the human. This is a hard one. Some people say you need the human at the start, then in AI, then to test it and check it, so the jobs are going to change. What's your opinion on that? Do you think there's going to be a lot of people without jobs, or do you think the jobs are going to change to suit AI?

Sam: I think a bit of both. I think 100% there's going to be a lot of jobs that are lost simply because their tasks can be automated. Teams are going to get a lot leaner. Those tasks are automated, so they no longer need someone to do it. And the people that are still around, rather than actually doing the work themselves, are just going to be approving and monitoring the AI's output.

Sam: But then you also get to the scary point, which is sort of what I figured out how to do this week. You get to the point of having businesses that are completely run by AI with no human involvement whatsoever. I think that's the scary point, and it's not that far away. It's on the horizon at the moment. Once we get to that point, there'll have to be some safeguards in play as well, whether that's from the government saying, hey, let's stop AI development, or from the frontier companies like OpenAI and Anthropic saying we're going to halt the development. Because I think once it gets to that point, it's going to be a super drastic impact.

Sam: If you think about it, AI is the next phase of evolution. You can't get a human that is more efficient at processing data than an AI is. So AI is just going to take over, the humans are just out-evolved because it's more efficient. So on the horizon, I think there are some big, drastic changes. Maybe in the short term there'll be a few job losses and people taking on an AI monitoring role where they're approving the output. I think that's where we're going to get to in the next few months, or are getting to at the moment.

Kane: Sounds funny, AI monitoring roles. We're checking the work of the AI as opposed to doing the work and having it checked by someone else.

Sam: Yeah.

Kane: It's wild.

Sam: It's scary to think about, but I also think about it as a good thing, too. People should be paid for their passion or their brilliance rather than the actions they take. If you think about it, no one was born to type numbers into a spreadsheet. No one's passionate about that. They're more passionate, like yourself, about having these conversations or doing something you're passionate about. AI can be a tool to free up your time so you can spend more time on doing what you actually like and creating value, rather than the mundane task of typing numbers or writing an email.

[29:17] Universities and new AI roles

Kane: Is there any, that you know of, any universities or TAFEs that are training people on how to use AI now?

Sam: Yeah, I think so. You see a lot of people on LinkedIn that are AI uni lecturers. So I think unis are starting to take note of it, but again, I don't think it's enough. From my understanding, I still think they ban using AI for writing essays or doing research and things like that. I was actually having an interesting conversation with someone last week saying that at a certain university they've banned students from using it to help with their tasks. But as soon as they finish school or go to the outside world, they're going to be using AI for it anyway. So they're being tested on topics and knowledge or tasks that they can use AI to do in the future. It's sort of a waste almost. You're teaching them how to do something that's going to be obsolete by the time they finish their degree. And then you hear stories of the lecturers using AI to actually market.

Kane: Yeah, exactly.

Kane: It's interesting. And I suppose with me even asking that, even if there was a uni degree that was like an AI developer, that's probably going to cost you, I don't know how much a uni degree is.

Sam: Yeah.

Kane: What, 40k? I don't know. But you probably could learn it online anyway.

Sam: Yeah, exactly.

Kane: Probably AI systems that are made to show you how to use AI.

Sam: Yeah.

Kane: Or how to monitor AI.

Sam: Yeah. I think there's a few courses going around. Funny you should mention that. I think Simon Beard actually made a company, I think it's called CEO Partners, where that's literally what he does. He finds these chief agent officers, trains them up and then places them into a business, and their whole role is to implement AI into the business and monitor the agent use. I think that's one new role that's popped up since this AI boom.

Sam: But the roles that we use for our development team when it comes to building AI are solution architects and people that study machine learning, because the whole AI implementation process, or the system implementation process, has been around for decades. These guys have been doing it for decades using plain code, and now they're just using AI as a tool. So it's the same process, just with a new tool on top. I think the guys with those degrees, machine learning or whatever it might be called, there's going to be a big demand for them as well in that chief agent officer position that's popping up.

Kane: Who's Simon Beard?

Sam: He's the guy that started and sold Culture Kings.

Kane: Culture Kings. Yeah, I thought I'd seen that. He has a mentor kind of program as well.

Sam: Yeah, One Life Club.

Kane: I've seen that. There's so many things happening.

[31:55] The gap between AI and everyday use

Sam: Yeah. It's scary times, but obviously the development of AI is so fast and there's so much happening on the frontier. But if you look at how it's actually being used day to day in businesses and in people's lives, there's a massive gap. It's moving so fast up here, but down here, there's barely anyone really using it to that full capacity. So I think on the frontier there's not long before AI starts to have some drastic changes and advancements, but until day-to-day life and society actually catches up on that, we've got a fairly big gap.

Kane: How do you think that will happen? How do you think the general public are going to start to use it to more of its potential?

Sam: I think it'll have to be a night and day difference. I think there'll have to be a massive change, or some event that changes everything. Because at the moment, people are using it and it's starting to pick up more and more, but it's quite slow. And then you get people at different levels that might be using ChatGPT or Claude or a full-on operating system. I think in businesses, what will happen is the top competitors will be using it quite heavily, and the newer guys or smaller guys will find it hard to keep up, whether that's on cost, or speed, or just an overall lower quality output, because these big guys are using AI and they can do it so much cheaper and quicker and provide a better output. So I think that will force the little guys to then go and look at it and say, hey, we'd better start doing this.

Kane: You have to keep up.

Sam: Yeah, exactly. Otherwise you fall behind.

[33:29] $10K websites and getting undercut

Kane: When did you start Echelon?

Sam: I believe March last year. I started offering it as a service through the marketing I was doing, and then it took off a little bit and branched out into its own company.

Kane: So with your system, you mentioned before about your one prompt to build a website. Do you still build websites for people?

Sam: Yeah, I still do that on site. It's pretty easy and not much of my time involved. I talk to guys that get quoted 10K for a website, and I can build the same quality website in 60 minutes for a tenth of the price. They're more than happy to go with that. They get the same result for less. So I think that's a clear example of where AI is affecting things. If these big guys don't get up with the times, then they're going to be undercut by someone who can do it at a way cheaper cost.

Kane: Yeah, that was literally my next question, the price point. Obviously, if the general public can do it, like you just said before, there's a big gap between people like you and people like me that aren't using it to its full potential. I suppose that's where a market for you could be, saying, yeah, we can help you do those basic things and it's going to be affordable, as opposed to building a website for like 10K, which is what we just did.

Kane: And we threw it in the bin, and then we made it pretty fast.

Sam: Yeah, exactly.

Kane: Not me, the team.

Sam: Yeah.

Kane: Yeah, it's just going to make things more affordable.

Sam: Yeah, 100%. I think that will probably be the event that makes people take it seriously. They'll start getting undercut by competitors and people doing it cheaper, and then they'll have to start looking at using AI to stay in business. Otherwise, they'll go bust.

[35:34] Why the tool isn't the point

Kane: So with your business, I suppose everything you do is AI.

Sam: Yeah.

Kane: Is that right? Is it challenging for you, with the changes that are happening every day with AI, or is that the whole point?

Sam: Yeah. When I was first starting out, that was 100% a challenge, trying to keep up with all these new tools coming out, trying to always be on the front of what's happening. Back then, when AI was peaking, well, it's still peaking, but it was really moving fast. Probably halfway through last year there were new tools coming out every day, new updates. Trying to keep up with that was hard. And obviously there was the open call that came out.

Sam: But now we've got our operating system and we've changed the way we implement AI into businesses, and this is probably the right way. When I was beginning, I was probably naive and wasn't doing it properly. But working with our lead developer, who used to work with Toyota and Costco doing this stuff before AI was a thing, so he's been doing it for decades, we realised that the newest tools are not actually what makes a meaningful impact in a business. It's not how you implement AI into a business. The underlying process is just setting up the systems and processes, automating those and getting those really good.

Sam: So really, it's got nothing to do with AI at all. AI is just an enabler that helps you do that faster. The whole thing is just automating and building out your systems and then plugging AI on top of that to make it better.

Kane: Yeah, that's actually a very good way of putting it. It's all the same process that we have been doing forever.

Sam: Yeah.

Kane: Just using tools like AI as an enabler to make everything easier.

Sam: Yeah, exactly. I look at it like Facebook ads isn't marketing. Marketing is understanding who your customer is, what they want, positioning your product or service as the best solution, and then talking to them in a way that makes them choose you over your competitors. And Facebook ads is just a tool or a channel that you can help convey that message through. I think that's similar to AI. The tools are just a channel or a way you can actually use AI, but the proper way, and what AI implementation is, is mapping out all your systems, building them super robust and then automating them.

[37:39] Personal brand and AI avatars

Kane: So on that note, with AI, obviously there's a lot of human connection that's going to be lost. Is your goal to build your personal brand so that you can build the trust through marketing and then sell your services?

Sam: Yeah, 100%. I think personal brand is becoming one of the best assets you can have nowadays. Like you said before, there's a lot of AI slop all over LinkedIn and then you see all the carousel posts on Instagram. Anyone can whip up a post, not have their face on it, and it just looks super AI generated. But it's hard to actually build a personal brand and get that trust. So I think that's going to be the currency of the future. It has been for the past few years, but now it's just going to be more and more important.

Kane: What about AI video, or your, what's the word?

Sam: Your clone avatars.

Kane: Yeah.

Sam: I think if you can get it good to the point where people can't tell the difference, then 100%, it's the way to go. If you could push out content without having to spend the time filming it and it looks like you, then I think it's the way to go. I don't think we're there yet.

Kane: I don't like that. I think that's a terrible way to do it. If you do that and it's good enough to not get realised, then the one time you get recognised, I reckon you just lose all credibility.

Sam: Yeah, 100%. And then it comes back to the thing, if something goes wrong, who's legally to blame or legally at fault? You can create an AI avatar and then someone can copy that and be doing some offensive stuff or make you look bad. Who do you blame? Is it your avatar, or has someone else created it? There's a whole grey area around AI output. Does it legally stand in court? Who's to blame? Can it actually hold up? I think that's a big grey area that we're going to see a lot of important discussion around over the next few months or next year.

Kane: Yeah, especially if it's going to be able to start doing, like you mentioned a bit before, whole companies run just by AI. But obviously it's going to need that human, that person that has to be responsible for everything, if it can get proved.

Sam: Yeah, 100%.

Kane: That's scary.

Sam: Yeah, definitely.

[39:43] Five-year plan

Kane: What's your five-year plan for your business?

Sam: My vision is to make the world and businesses operate more efficiently. Following on from that, the goal is to move more enterprise level and work with some bigger companies. There's more leverage there, and more change has a greater effect. So the goal is to help businesses operate more efficiently, and I think the best way to do that is to work with bigger and bigger companies and increase our reach through there.

Kane: So with bigger companies, I suppose that would come with bigger risk but also bigger reward.

Sam: Yeah, exactly. The operating systems that we've been doing are great for one, two or three people. But when you start to go to 100, 200 or 300 people, there's a lot of issues in terms of access levels. How do you get it all syncing and updating? And obviously if something goes down, that's their whole company that's gone down. So it comes with its own set of challenges for sure.

Kane: Is that scary or exciting?

Sam: I think it's a bit of both, more exciting. There's a bit of nerves there. If you stuff it up, you can get yourself in a bit of trouble. But I think it's a cool challenge to try and solve, to figure out how I'm going to do that and what it would look like in production. It's a super cool challenge I'm excited for.

[40:55] Networking and Inner Circle

Kane: That sounds fun. So talk a little bit about networking, because we met through Inner Circle. Shout out to them. Why did you join Inner Circle in the first place?

Sam: I think entrepreneurship can be a lonely journey. So it's good to meet like-minded people that are on the same path, and you can have a genuine conversation with people you're interested in. It's also just to meet new people and see what they're doing. Perth is a pretty small place, so finding such a concentrated area of all those entrepreneurs is really cool. You meet some interesting people, see what they're up to and get different ideas. There's possibility of working with some of them too, which is quite cool. I think referrals and working with people from your own network is probably the best way to do business. You work with some really cool people, you're generally passionate about what you're doing and you want to help them as much as you can.

Kane: Yeah, I agree. I think networking is so powerful. I've only started networking with businesses bigger than what I've built over the last two years, which is pretty crazy. I always just assumed that people didn't want to talk about their businesses because it's competitive and all that kind of stuff. But the more people you meet, that's definitely not true. Obviously, some people are funny, but a lot of people are more than happy to share their stories and help other people.

Kane: I do think the whole human connection thing is kind of like the opposite to AI. So you need the human connection first and then you can offer your services, and not just get spammed by AI experts every day on LinkedIn.

Sam: Yeah, 100%. Like I said before, anyone can whip up an AI post or send out an AI generated message. I think that connection and person-to-person contact and conversation is going to be 100% the currency of the future. Something like Inner Circle is the perfect way to develop that and meet new people.

[42:51] Biggest takeaway and where to find Sam

Kane: Agreed. Is there anything we haven't covered that you think is important?

Sam: I think we've covered the key points. Probably the biggest takeaway is that the AI tool you use has nothing to do with how well you implement AI into your business. If people take one thing away from today, it should be to stop chasing the newest AI tools, just pick one, focus on it and go all in on that. Or look into implementing AI properly, whether that's mapping out your processes and then automating yourself through Claude. But I think if they can pick one tool and stick with that, it'll help them a lot in the long run.

Kane: So true.

Sam: Yeah.

Kane: And where can people find you and your business online?

Sam: If they just Google Echelon AI Solutions, we'll be on our website, and then Instagram and LinkedIn as well.

Kane: Awesome. Thanks for coming on.

Sam: Awesome. Thanks for having me.

Kane: Thanks for tuning in to this episode of Unlocking Business. If you've enjoyed this conversation, please take a second to like, subscribe and follow our channels. It really helps us to bring on more guests and have great conversations. If you'd like to connect with today's guest, the details are in the description below. I'm Kane Tyler Smith. I'll see you on the next one.