Smarter Automation for Small Business: AI, Speed & Zero Bottlenecks with Carl Taylor

Episode 49 Carl Taylor
===

Understanding Automation: Beyond Just Tech
---

Carl: [00:00:00] I have a key philosophy is you should separate automation from innovation.

Carl: And so what I mean by that is if you're going, well, what should I automate first?

Carl: It's whatever it is you're currently doing manually

Carl: Map out your current process.

Carl: Automate what you already do. Once you've got that set up, you're getting yourself time back. Now you've got time back. Now you can look for where are the opportunities to enhance and to improve. ​

Introduction to the Lean Marketing Podcast
---

Allan: Welcome to the Lean Marketing Podcast. I'm your host, Allan Dib. Today I've got with me someone who's been a friend for quite a long time. And he's a super smart guy. I always get a lot of energy talking to him.

Allan: A lot of creative ideas. He's kind of a bit of a tech geek like me as well. So, we spent hours talking tech, CRMs, marketing, all of that sort of stuff. So, welcome to the podcast, Carl Taylor.

Carl: Thanks, Allan. So great to be here and yeah I remember it wasn't even that long ago, maybe a month or so [00:01:00] ago, and we're chatting, CRMs for probably a good 40 odd minutes, right? Like it was

Allan: Yep. We really went deep. We went deep. And by that I mean, I was listening and asking you questions and interrogating you and uh, and you were a wall of wisdom. It was good. I really enjoyed that.

Carl: kind of you, very kind.

Carl Taylor's Journey and Insights on Automation
---

Allan: So, Carl, for anyone who doesn't know who you are what's the two sentence intro to you, if you met someone on the street?

Carl: I help business owners become o owners of their business rather than operators. That's kind of the high level. I run a company called Automation Agency, and we ultimately help business owners delegate their marketing without having to hire, manage, and train anyone.

Allan: Yeah, I love that. You've been doing this for an incredibly long time., uh, you know, long before AI was a, whole big thing.

Allan: and you, Helping people automate workflows and just get stuff done from a marketing operational perspective, which has been a huge blocker for a lot of people because a lot of people have the strategy, they've got the [00:02:00] ideas, they've got the big picture, but they're like, I don't know how to do the landing page.

Allan: I don't know how to do the website. I don't know how to connect the shopping cart to the site, to do the upsell, to do the, technical bit that they've been. Big blockers for a lot of entrepreneurs for many years. And you've been unblocking those things. So, how did you get into this when did you get started?

Carl: Yeah, man. Well, so automation agency you just celebrated a few weeks back, our 11th birthday. So we've been running automation for 11 years. I've been in business now for. 24 years. I

Allan: Wow.

Carl: I'm about to turn 39. So

Allan: You don't even look 24 years old man.

Carl: I was 15 years old, my first official registered a BN business and it was in building websites and stuff for people. So it's kind of like I went on this big journey, had an IT company, had a whole bunch of things, but came back. So I've been in the techie space and the small business space for a long time, 24 years. The online marketing in particular came about when I wrote my [00:03:00] first book. Red Means Go.

Carl: I wrote that book and I was like, oh, I need to like promote it. I need an email thing. And so I signed up to this product called Send Pepper, which for

Allan: I remember that actually. Yeah.

Carl: the, was what Ontraport now is basically,

Allan: yes.

Carl: it was like the $20 a month version. that was my first real introduction to this idea of you can automate some form of marketing,

Carl: so that's 15 odd years ago.

Carl: And what I found is automation entity was born from talking to a whole bunch of other business owners. And they would be like, oh, I just hired this virtual assistant, or I'm trying to do this thing themselves.

Carl: Mm. They would show me their office autopilot on ontraport accounts and it would be an absolute mess,

Carl: Mm. It was like, hold on, we can do a better job. What if I could train these people to actually know what they were doing?

Carl: And then, rather than you hiring one virtual assistant, could I just give you a whole team so you're not beholden to one person and you don't have

Carl: Yeah. them and they know their stuff? And that's, what was the birthing place of what automation entity now is. And then, we implemented our first AI feature [00:04:00] in 2016.

Carl: So before ChatGPT kind of

Carl: Blew the world up. We started implementing ai because the way our model works is people send us tasks. And before we had our web portal and mobile app, it was all via email. And we were trying to scale and we're like, we need to hire full-time people across multiple time zones.

Carl: ' just to receive these emails and decide, oh, that's a design task. Send it to the design team. Oh, that's

Allan: Yeah.

Carl: send it to the web team. Oh,

Allan: Yeah.

Carl: task. Send it to the automation team. And I was like, that's crazy. That's gonna cost too much. So I hired a data scientist train our own custom classification model that could read the emails, here's all these historical emails, read them to determine, you know, what it is.

Carl: And so that's what we built in 2016. And then, funny enough we just replaced that whole system recently with a simple chat GPT prop. Now that does what cost us a lot of money to

Allan: Yeah.

Carl: scientist create, you know, years ago. But that's kind of like the world that I've been in and, I'm excited with where things are going now with the state of AI as well.[00:05:00]

Allan: That's very cool. And, you know, I was an early customer of automation agency long before, well, I mean now I've got a big team.

Allan: When it was just me and maybe an assistant and a couple of people early on I remember using automation agency, it was great because I'd submit a task, maybe it was a design task, a landing page or whatever, and boom. It would just get done.

Allan: I didn't have to worry about hiring, you know, a graphic person, then a web guy, then a whatever. So, and I think the model that you've created is really interesting because. You've selected certain tools and they're very popular tools. So things like make things like active campaign, ontraport, those kind of things.

Allan: And you're like, these are the tools we support and they, they're tools which are pretty common now in, a marketing stack. and yeah, you've got people who are just specialists as that. And then, you know, you don't have to look at five different places to, get something done, which I really like.

Carl: Yeah. The vision and where we're going even more now with some of the stuff we're building is to really be that all in one marketing team

Allan: Yeah.

Carl: [00:06:00] business, right? Like I've live and breathe small business. I'm, you know, we've got some larger clients and people who have large teams where we.

Carl: Maybe their overflow and you know, we're not necessarily big corporates marketing team, but we can be an overflow for their existing marketing team sometimes and for agencies. But really we're built for small business. And it's by using AI because now obviously things can be done faster.

Carl: So make all our team are constantly looking for new AI tools to help them achieve what clients want done faster. So the client doesn't have to figure out these AI tools,

Carl: But also, making sure there's still always a human in the loop, because anyone who's probably listening or watching who's around with some AI tools, there

Allan: Yeah.

Carl: things it can do, but it's sometimes it's really frustrating that you're just like, it doesn't quite get it, and then you feel frustrated.

Carl: Whereas imagine if you could just have the AI do that initial bit and then go, oh, let's just send it to a human to, to fix it, to polish it off, and you don't have to worry. That's kind of where we're going and what we're building, which is very exciting.

Allan: It's that last mile problem, like where you can get something to 80 or [00:07:00] 90% and you feel like it's 80 or 90% done, but it's really maybe 50% done in reality because that last 10 or 20%. There's a lot of fine detail things to that. So I've been in that situation many times where I've been. I'm like, I'm almost there.

Allan: It feels like I'm almost there with whatever, implementing something, some automation, some whatever, but I just can't get it to work or it's breaking at the last minute or whatever, and it's super, super frustrating. So, I think that's an important

Carl: sometimes, like I love chatting to the various, like, I chat to Claude, I chat to chat

Allan: Yeah.

Carl: Like I use lots of different tools for different things, but I find sometimes that it takes things off my to-do list and then adds a whole bunch more.

Allan: Yeah.

Carl: Right, because of the speed in which it can do. A simple example, you might write a piece of copy using one of those tools to write an email.

Allan: Yeah.

Carl: that's great. You've got the email back, but now it's back on your to-do list, to actually log into your platform, figure out how you're gonna send it, to actually send it.

Carl: It's amazing in that you're like, oh, I got this thing done and you get that win, but I dunno about you.

Carl: I then also then get the feeling [00:08:00] of like, oh. But now there's something back on my plate, at least with old fashioned email, which is like a to-do list. Usually when you sent the email to someone, you got it off your to-do list

Allan: Yeah.

Carl: you didn't have to worry about it until it came back. And because humans took time, it might be a day or two breathing space, but AI doesn't have that breathing space.

Carl: You give it

Allan: Yeah.

Carl: to do, it comes back within minutes. . And that's one of the challenges that I personally experience and. we've been working on trying to solve that problem for our clients, so they don't have that issue, you know.

Allan: Very cool.

Leveraging AI for Business Efficiency
---

Allan: So, I'd love to get your take. What are you seeing some of the biggest opportunities with AI right now? So, I mean, there's a ton of things you can do and you know, I mean, you watch some of these TikTok videos or reels or whatever and people are like, I. Just upload CSV.

Allan: Just go to ChatGPT, create a hundred different inspirational quotes, upload them to Canva and then upload them to Instagram. I'm like, what for, like, what's the point of that? Who wants to read robotic quotes? Autogenerated by ChatGPT and Canva or whatever, and just. It's just stupid. [00:09:00] Like, what I'm interested in is how do we intelligently create leverage and leverage meaning just get more results by doing less stuff.

Allan: Not just use automation for the sake of automation because, you know, situations like that they're not gonna help you. So. What we want is to move people to action and have people engage with our content and all those sorts of things.

Allan: So what are the big opportunities you're seeing? And perhaps I think you probably more than anyone else, are seeing opportunities that a lot of people are missing.

Carl: so. Think one of the first things there is, you're right, like we are in a huge influx of content.

Allan: Yeah.

Carl: been easier to create content.

Allan: Yeah.

Carl: I've had people pitching me to create my Instagram reels for me, where it's AI Carl,

Allan: Yes.

Carl: Like, I don't have to record.

Carl: They're telling me I don't have to record anything, and they've given me a few demos and they're. Scary. Good. I personally still have a, weird feeling about it. Like, if my content goes out there, unless I'm [00:10:00] saying, Hey, did you know this is ai

Allan: Yeah.

Carl: Like, it feels weird to pretend it's me as a talking head video.

Carl: So, so I've said no to these people pitching me, but like the technology is there and not everyone has that same stance. They're just pumping out content.

Carl: Yeah.

Carl: never been easy to create content, but. what you mentioned. At the end of the day, it's gotta strategically solve and serve your business.

Carl: Yeah.

Carl: come back to, who is your ideal client? cause for most businesses, 99% of businesses seeing a quote is not gonna inspire them to do business with

Allan: Yeah.

Carl: But there are, there's probably 1% of businesses where maybe based on their ideal client, that's actually the exact thing they need mixed into their content.

Carl: That's gonna bring them closer to what is you do. But that's where I think acan really help you. The problem with tools like ChatGPT, and it is getting better, is they don't fully understand what you do. They don't know everything you do. What you really need is you need to teach the AI who you are, what your business stands for, what are the brand values?

Carl: What does your ideal customer actually look like? it needs to have that extra context

Carl: Because if you use an AI tool like. Google search, you're gonna get very [00:11:00] limited results,

Carl: But when you treat it more like this is a human being, this is a new junior hire and I need to give them everything they need to be able to win at this task, I'm about to give it, and you give it that in your initial prompt, the results you get are astronomically different. And so you really, you need to look for AI tools that can. Learn over time more about you and our brand. Like one of the things we're building is this integrated memory. So when you are delegating a task related to your brand, if it asks the question, you answer it, it remembers that over time. So you don't

Allan: Yeah.

Carl: you know, explain yourself over and over again so that the content gets better and better.

Carl: It's always gonna be refined. But coming back to the strategy pieces, you don't wanna just produce content for content's sake. You could say Give me a year's worth of content, it's gonna be meh.

Allan: Yeah.

Carl: You gotta come back to what is it you are trying to sell? What is that emotional journey you're trying to take them on?

Carl: Where going through your standard tools of like the one page marketing plan, like being able to understand who are my ideal client, what does I sell? What are these key components? But load that up into your AI tools, help it understand [00:12:00] that, and then ask it questions to help craft what that journey looks like before you actually start generating the content. that's from a content piece. But I think that content's an obvious play. A lot of people are already pushing that. some of the more interesting AI opportunities for businesses right now are, it's going to get clamped down on eventually, but right now, AI phone. Is really interesting.

Carl: Like there are some, depending on who you sell to an AI receptionist is super easy. Like if you, right now you've got a lot of calls in your business going to voicemail,

Carl: Yeah.

Carl: should replace that straight away with an AI receptionist. You can do that with tools like go high level. We can help you set that type of thing up, but like a, a simple receptionist where instead of it going to voicemail, you forward it to a number. Where AI can answer that and you can load it up with standard questions people would ask. You can give it the ability to book appointments or to send SMSs to that client with links to do different things. So you can give it the ability to actually be useful rather than just leave a message. You can still leave a message, but you can actually achieve some of the things that, that someone [00:13:00] might be calling about, and so you won't miss those leads and those opportunities get them booked in the calendar.

Carl: Another opportunity that I see for business owners is. When they're building out their automated flows. So for decades we've been able to build these automated flows. You know, someone fills in my form, they get this email, then two days later they get this email. Then three days later they get this email. some really interesting situations of can you personalize?

Carl: I recently signed up to a tool and I was blown away, and we are now to put this into our workflow of onboarding too. But like the moment I signed up to this tool. They had an AI agent that went and crawled my LinkedIn, so use my email address, crawled our LinkedIn and it personalized the very first email that went to me to say, Hey, welcome aboard. I can see that you know, you're the CEO of automation entity and you've been doing it for this many years and like it personalized this whole thing. And based on everything I know about you, I think this would be the very first function in our tool that you should use,

Allan: That's cool.

Carl: right? Like that level of personalization [00:14:00] impressive.

Carl: And so those kind of things, they're not doable straight in tools like your active campaign and stuff anymore at the moment. I'm sure eventually they will come, they are doable. Go high level, you can do it to a degree but you can definitely do it with tools like Zapier and make to make that possible and build these workflows. But I wanna go back to something earlier you said, 'cause I think this is really important.

Carl: I have a key philosophy is you should separate automation from innovation.

Carl: And so what I mean by that is if you're going, well, what should I automate first?

Carl: It's whatever it is you're currently doing manually

Carl: Map out your current process. This might be your customer onboarding process. This might be your content creation process. This might be your customer service flow. Like it could be anything, right? But look at what you currently do. But the big trap we all fall into is as I'm trying to automate it, I'm trying to get creative and put all these fancy bells and new features in

Allan: Yeah.

Carl: But if right now your [00:15:00] first email that goes to people is fairly templatized, but you're still sending that from your Outlook or your Gmail or whatever, an easy win. Don't

Carl: Yeah.

Carl: the personalized email that I just mentioned. Just literally take that template that into your standard autoresponder. Boom, it's done. Automate what you already do. Once you've got that set up, you're getting yourself time back. Now you've got time back. Now you can look for where are the opportunities to enhance and to improve. Where it gets messy is we go, oh, this is how we currently do it, but we wanna build a whole automated flow that's gonna add 20 new steps of things we don't actually currently do,

Carl: Makes it a bigger project.

Carl: It makes it harder to do. It means things might go wrong. And might make people go, oh, the automated process isn't as good as the way we used to do it because you actually changed too many variables to know what's happened. if you walked away with nothing else and you're looking to automate processes in your business, they marketing processes or non-marketing processes, and when I talk automate here, I want you to understand my definition of automation not just tech. my

Carl: definition Of automation is [00:16:00] me not doing it.

Allan: Yeah.

Carl: And that can be tech or it can be team that's are we automation entity? It's

Carl: because we give you the tech and the team to help get this

Carl: stuff off your plate

Carl: It's just you not doing it. Not everyone thinks about it. So to me, automation is about tech and team.

Carl: It's not about just a tech tool. And so this principle of automate. Separate automation from innovation is true for when you're building team pipelines. And I also want to talk about like even in that content creation flow don't always go, I need to automate the entire flow. It might be one simple step. Maybe right now you spend hours trying to come up with the perfect script for your Instagram video. Well, maybe if you just spend a little bit of time and go, okay, I'm not gonna change anything else. I'm still gonna change, still do the way I come up with ideas, but the scripting process, I'm now gonna start using ChatGPT or whatever tool.

Carl: Here's my prompt that I've crafted. I've spent a bit of time to build something that, that allows me to just say, Hey write this. Like the way I get my scripts done. I open up ChatGPT and voice mode only [00:17:00] because automation, entity's copywriting doesn't have a voice mode yet. But when it does, I'll switch to that and I basically just say, Hey, I need a script for my Instagram. is who I serve. This is the Instagram account. This is the audience. This is the hook that I'm going for. This is the comment or the caption that I plan to use. And here's the general idea of what I'm thinking, and I'll just talk through what I'm thinking. And then eventually what'll come out. Is it'll come back to me with something and then I'll just iterate back and forth like a person, I'll be like, oh, that's not quite right. I wouldn't say that. Or, actually, what I'm really saying is this, until I get to a point that I'm pretty happy with it, I exit out of voice mode. I've got a written script there, and then I go away and record it. That was a process that took a long time that now I can bang out a bunch of scripts in a few minutes and it's, it's made a huge difference. And over, time, I've taught it the framework of what's the ideal script. It's not just general chitchat.

Carl: But that's the example of like, I haven't automated my entire flow.

Allan: Yep.

Carl: I've just gone, here's a step that was taking a lot of time. Can I find a way either with tech or team to [00:18:00] make that easier?

Allan: Hey, Alan here. Want to dive deeper into today's episode? Head to lean marketing.com/podcast for links to all the resources we mentioned, plus some exclusive ones. Just for podcast listeners, you can also subscribe to get notified when new episodes drop and receive my latest marketing and business tips right in your inbox.

Allan: That's lean marketing.com/podcast. Now back to the show.

Allan: Yeah, it's very common trap where people are like, Hey, I wanna automate sales qualification, or I wanna automate some marketing process. And often it's like, dude, you don't even know how to do this manually, so how are you going, you know, how are you going to you have to tell the machine, do this, then do that, or.

Allan: Sometimes it's machine, and like you said, sometimes it's man, right? So, some steps are done by teams. Some steps are done by tech, but if you don't know how to do the thing you're gonna have a pretty hard time in computer science. They call it GIGO. Garbage in, garbage out, right? So if your dunno how to, if your inputs are not good, your [00:19:00] outputs are gonna be not good either, right?

Allan: So, very common thing. And, I did an AI workshop a few months ago in San Diego. So, one of the things that we were working on is how do we create content? That sounds pretty good because the stock standard ChatGPT and large language model output is pretty cliched, pretty crappy most of the time.

Allan: And so we got to a stage where we could pump out stuff that was. Pretty good. Still not perfect. It is still definitely needed. Editing, still needed improvement, but we did that through essentially a four step process. First, we would tell it what the goal was, then we would tell it. So for example, we might say, Hey, we're running a campaign where we want people to click on a link and opt in.

Allan: So that's the, we're telling it the goal. It'd be more detailed than that, but and we'd tell it about the audience, all of that sort of thing.

Crafting Effective Prompts for AI
---

Allan: Then we would tell it the format. I want it in this format. I want a 600 word. Whatever LinkedIn post or whatever it is tell it the format, then we would tell it what we didn't want.

Allan: So we would say, I don't want hypey kind of [00:20:00] language. I don't want cliched words, all of this sort of thing. But then here's the really important part. We gave it a ton of context. I gave it like, you know, 50,000 words that I'd written before in my style. And I said this is some of my past writing.

Allan: Use a similar style, use, similar sentences and structure and all of that. And it came out with. Like, like I said, it wasn't perfect, but much, much better than if I had just said, write me a 600 word LinkedIn post for this way better. So you, so the key is really with large language models, like ChatGPT, like Claude Deep Seek, whatever you're using is framework.

Allan: Plus context, Like what sort of process do you want it to follow? What format do you want it in? What do you not want? So what are the negative things that, you want it to look out for?

Allan: And then context and the more context you can give it. And now these large language models, they're allowing for more and more context, I think. Google's one that's probably the one that allows for the biggest context. Gemini, I think you can have 2 million units of. Which is [00:21:00] essentially like, like maybe a full book of context.

Allan: Yeah. So, really powerful. So I'm using that all the time and I get much, much better results.

AI as a Writing and Research Assistant
---

Allan: And I use it as a tool to really augment my abilities. I don't think it's gonna replace me with writing. I'm write, I'm currently writing my next book. I'm working on you know, my, you know, content all the time, and.

Allan: I still can't get it to come up with completely novel ideas. So that's on me. I still can't get it to just completely do this stuff, but it's like a really powerful research assistant on tap. It's like a really powerful copy editor on tap. It's like, you know, a maybe like an idea thought partner where you'd say, Hey, does this make sense?

Allan: Is there a better way to say this? Or whatever. So, in that way, super, super powerful. I think of it like the Iron Man suit, right? You know, Tony Stark, geeky guy. Normal guy, but he puts on the Iron Man suit. Now he's got all of these powers, you know, he can fly, he can shoot, you know, all of this sort of stuff.

Allan: So I think of it that way.

Carl: Yeah. I think of it in the same [00:22:00] way. You know, it's it's funny you use the Iron Man suit. That's one of our internal things we talk about, like, that, you know, our human heroes, the guys that still do the work. Like they're the humans still in the driver's seat, but it's like they put on an Iron Man suit when they're leveraging the AI tools.

Carl: Some of them are internal built, some are, you know, external tools that we use, but like, these allow them to have more superpowers to deliver a better result. So I agree. Like I think we are coming into a wave economically that we are gonna see a huge shift in how work happens. I think What that looks like. Everyone's still clamoring to guess of how that's gonna look, but I think there is gonna be a big shift. But right now that I don't see them eliminating us, they are still very much augmentation tools. And if you use them well you really can move so much faster. Like, as I said, the downside is it spits it back and it's now back in your plate again.

Carl: But like you as the human are the bottleneck far more in terms of the output that these ais can help you produce. There are things that I've brainstormed with AI that I know. I would've probably taken months percolating to think [00:23:00] about the best way to do something, and I've been able to just back and forward over a few sessions with these tools to come up with a plan. Is the plan as good as what I might've spent months percolating to come up with? I don't know, but the difference is it took me maybe a day or two, and I'm now at a point where I'm executing on that versus I spent months thinking about it. And that to me is hugely valuable. That idea of being able to get more done with less time, less thinking. We don't wanna outsource the thinking completely though. I think that's a

Allan: Yeah.

Carl: Stage where humans might delegate all their thinking. I read something recently about, apparently some young singles are taking the, their chats that they're having with dates, like SMS chats and they're pasting it into their tools like chat GPT and asking for how to be a flirtatious response to this and stuff like that.

Carl: Which, you know, that's good 'cause it could be a way to learn. But it does scare me that. If you use something like chat GPT to decide who's a good partner for you or not, like that might be a dangerous space to, to move into.

Allan: Yeah. The other thing I found with some of these [00:24:00] tools. Especially ChatGPT I think for not so much Claude, but certainly ChatGPT, it'll often try and confirm some of your views. So like, it's almost like a hype man which is okay sometimes, but a lot of times I want to know when I'm wrong.

Allan: I want to know. When I'm off track or whatever, and sometimes it'll do that, but I find Claude is more honest. So Claude will be like, no, that's incorrect, or that's not exactly right or whatever. Whereas I feel like ChatGPT tries to just confirm whatever you put in generally.

Allan: So

Carl: hundred percent. one of my favorite things to do whenever I go brainstorming, as soon as I feel like it's too much, like in the positive, I will then go, okay, great. What am I not seeing?

Carl: Tell me why all of it's wrong. Like give me the other perspective. because you're right, it won't think about it all or expose it to you unless you ask for it. You know, I remember like when Google first kind of hit, the scene and it was like, people don't need to learn anymore, really. They just need to know how to find the information.

Carl: Now we need to know how to properly discern information. We needed to do that before AI hit, but I think even now it's like. [00:25:00] You need to avoid the echo chamber of the AI telling you everything you want to hear, to be able to discern and ask it back and forward, which is in one way is better than maybe your social media feed that might be fully reinforcing your views.

Carl: And you can't ask someone in your bubble to give you the opposite. Whereas ChatGPT can at least access the opposite view to give you a more balanced approach. But yeah, it's a very interesting time we live in.

Allan: I found it really powerful using it as a thought partner where it's like, does this make sense? Or, does this analogy work for this or that, to explain that. And it'll say yes or maybe not quite, or whatever you are, mostly right, but you've missed this that, or the other.

Allan: So I found that a really powerful way of using it. But I think one of the mistakes we always make with any new media is we use old media paradigms on new media. So like when radio came out, it was just someone reading the newspaper on a radio, you know, when TV came out, it was just someone basically being a radio announcer, but You are watching them tell [00:26:00] you about the weather, then the sport and all of that sort of thing. When social media just came out, people treated it like a broadcast medium, just like tv, you know? Hey. And so it's important that when a new media, and I think we can think of aand large language models as a new media, we, they're essentially replacing search engines to a lot of extent.

Allan: Like, I can't even remember the last time I used Google for anything other than maybe looking a phone number or something like that. It's definitely my. Second or third priority when looking for information.

Allan: So, if you are using just large language models like ChatGPT, just as a search engine or whatever, okay, that's fine, that's level one. But there are so many better ways to use them and so many more powerful ways to use them, especially if you're thinking about augmenting your abilities. And then also taking into account their limitations. So the limitations, especially now, is that they're essentially giving you the median response, the average response. So, which in a lot of situations is great. You don't want to [00:27:00] come up with wild stuff that you know isn't really done in the tech industry generally.

Allan: Same if you're drafting a legal agreement, you want it to be a easily readable legal agreement that any lawyer might have come up with or whatever. But where they suffer is where you're coming up with a novel idea, something brand new, something really novel. They're not really gonna come up with that.

Allan: So that's your job. And then you can augment your abilities and help you basically compound that idea and implement it in a really powerful way by using it as a thought partner, by using it as a writing partner, by using it to help you generate ideas around it.

AI in Business Communication
---

Carl: Yeah, some of the other unused cases, I think that businesses don't think enough about is if you have any kind of form of, let's say for example, you do sales appointments, someone books in for a sales call, and maybe you send an SMS reminder and someone replies to that. And the other option is it goes in an inbox and hopefully you've got a human team member who can deal with [00:28:00] it. Well, you can have an AI that can deal with those things, you know. Another business I had that I recently kind of wound down what we did is when someone downloaded a lead magnet, we had an SMS going out. And the SMS was a hard coded SMS of like, here's the link, here's the thing.

Carl: But it started with a question. It was a conversation starter,

Carl: Actually replied to that. In the beginning, I had a human who was doing those SMSs. then time, once we'd done a few of those conversations really dialed in our process, goes back to the earlier point of like, I didn't innovate and try to automate something I hadn't figured out yet.

Carl: We dialed in the process with a human, and then once we had that dialed in. We then trained an AI that had those same, basically did that same conversation. And so when the human replied the, instead of it waiting potentially 10 minutes, 15 minutes, 20 minutes, up to three hours sometimes for the human to be able to jump in and reply, the AI was able to respond basically immediately.

Carl: And in a medium like SMS, you know, you're gonna have a higher response. Like all the stats show that if someone fi fills in [00:29:00] a form online, if you call 'em or SMS, them, like within seconds of them doing it, your conversion rates go up because it's just. You've got them while they're hot. Their intent is high while they filled in that form.

Carl: So if you're able to actually connect, the opportunity of a sale goes up dramatically versus the longer time goes. And so the SMS responds right away, asks another question. The client responds. And so the, this lead and aare having this conversation. And the objective for the AI was just to qualify them to book a call.

Allan: Yeah.

Carl: set done over SMS. Boom. Great. Hey, here's what I'm thinking. and it's not just scripted, like this is what's the paradigm shift? Pre-AI and large language models mixed in. It was just a standard, like if they say this, send

Allan: Yeah.

Carl: they say that, send that. Now you still got the general, if this, if that. But now it's like if they don't say that exact word, but it can at least know that it was a positive response. how to respond. And when it says, say this, it doesn't say the exact same thing. It'll be like, oh, you've got three kids. That must be a handful. Like that's some example that [00:30:00] the ahave said that's not programmed anywhere. It's just decided to say, oh, you've got three kids. That must be a lot. And so it feels like a human. And you get these back and forward SMS until eventually they go. Here's what I'm thinking. I think we should book your call with Carl.

Carl: He's amazing. He can chat through and we can put together a bit of a plan. Would you be up for that? They reply back with, yes. Great. Here's the link. And it just sends 'em the link to like the Calendly type booking. It's the same thing that someone, a human used to do. And then we just replaced it with AI

Carl: like those are the examples of if you've got any kind of appointment setting like that over a text-based medium, ais a huge opportunity for you if you haven't explored it, especially if you've proven the model. You've got a proven process humans are doing, you can augment, which allows you can, those people you've got doing that can now potentially hit the phones if you're not willing, like there are AI outbound calls. I think regulation will dial that in eventually.

Allan: Yeah.

Carl: and scammers will

Allan: Yeah.

Carl: I'm not so sure where I sit on that right now, but I think inbound calls with AI answering, I'm all for that. I think that makes sense. Just don't hide that. It's AI you say, Hey I'm [00:31:00] AI and then I think SMS or email interactions back and forth, that's a standard flow, not an open inquiry, but a standard flow. but outbound calls, I still kind of put in the human category for now.

Allan: I totally agree. I was experimenting with a voice agent from 11 labs and I loaded. All the transcripts from successful sales calls that our sales team did. And man, it is really good. Like, you would go through, you'd ask at a bunch of questions or whatever. It responds very low latency really powerful.

Allan: I'm very tempted to put it in now sales process and just test it out for real. I think you're right. You should probably disclose that you're talking to an ai because again a lot of times you can maybe. Not even know. But I think it'll be one of those things where, you know, like right now if I need help or support with something, I'll go to a company's contact page.

Allan: And I'm so relieved when they've got just an online chat where I can just chat or, you know, I'm like, ah, great. I don't have to call someone and wait in a queue and all of this sort of [00:32:00] stuff. And even if you have to wait in a queue for a chat, it's better than waiting in a queue on a phone. I don't know why that is.

Allan: It just is. I think it'll be that way when AI gets really good, you'll be like, ah, thank God I don't have to chat to a human. I can just get AI to solve my problem really quickly and instantly. So,

Carl: I think, but I think you hit the nail on the head. 'cause I've been thinking a lot deeply about this. Obviously with what we do at automation and see and where we're going, what AI unlocks is speed,

Carl: Right? Like it, it's previous systems. We were it gave you the ability to automate certain things and create processes, which was very beneficial to the business owner. Didn't necessarily always add more value to the end user could in some ways, but it was really more beneficial for the business owner. But when you mix AI into the mix here, can now benefit both the business owner and the end user because the, like it or not, social media, the world where it is wired us to enjoy speed. And as soon as we get used to a certain level of speed with AI that expectation gets pushed onto humans. [00:33:00] I've

Allan: Yeah.

Carl: in myself literally in the last few weeks. I've been working deeply with our development team, and of course I work with Claude and some of these other tools to brainstorm and do some developing too. And when I give something to my development team to build, and it might take them a day to do something, and I'm used to Claude or a tool like Lovable or one of these other, you know, vibe coding tools, is able to spin something up literally in minutes. I'm frustrated that they took a whole day to do something

Allan: Yeah.

Carl: You know, it's crazy because I've been rewired, but the reason this is important is your consumers is this is happening to them too. And if it hasn't happened yet, it's only going to happen. And so it's gonna become a requirement for us to have these AI components because we can't scale the speed like AI can humans.

Allan: Yeah.

Carl: So humans need to be that personalization. Like a long time ago, an early mentor, Brad Sugars said, you know, systemize the routine and humanize the exceptions. I think AI just magnifies that so much more and looking at humanizing [00:34:00] the relationship,

Allan: Yeah.

Carl: . So people are building relationships with AI already, so. AI can actually do. It's no longer humanize the exceptions. It's kinda like humanize and augment AI exceptions, because the routine is easy. Yeah wild times, and I just I think it's exciting, but you've gotta consider AI has to be in your marketing mix and your overall business mix, because consumers will demand the speed that only AI can currently deliver.

Allan: Yeah, you're freaking me out, man. And people are building relationships with AI I mean, there's now some of these AI flirting chat bot things or whatever, and I mean, I'm on some of these AI reddit forums where you actually run AI on your own local hardware.

Allan: And a lot of the reason they do that is so that they can have explicit chats with the AI and all of that. I don't quite, I don't quite get it, but I mean the Reddit forum is just full of all of that sort of stuff, and there's now AI girlfriend services or boyfriend services or whatever, and I'm like.

Carl: I think you know, there's a lot [00:35:00] of only fan people I've heard that don't actually exist. They're like, they're probably men, entrepreneurial men who have gone, Hey, I can make a bucket load of money. Let's launch four or five or six different OnlyFans models that are all completely agenerated and AI chats. And they're probably raking in a huge amount of money. And

Carl: I think, I believe some users know that they're AI but they don't care. Like they're happy to pay for. Like, it's bizarre you know, it was an article I was reading just a few weeks ago.

Allan: Carl, be honest. Are you the one behind all of these AI OnlyFans?

Carl: I I wish I was that clever and didn't have my own morals and ethics that made me feel like that wasn't a place I wanted to go. But

Allan: Fair enough.

Future of AI and Automation
---

Allan: Where are we gonna be in a year from now in five years from now?

Allan: In terms of AI in terms of automation, are we gonna have much smaller teams? I hear Sam Altman keeps talking about the one person billion dollar company kind of thing. So who's gonna be the first person to build a billion dollar company with just one, person by leveraging AI or do you think we're just gonna [00:36:00] increase productivity per unit? Where are we going?

Carl: So increasing productivity per unit? A hundred percent. That's already happening. We can see it. We see it in my own company. I mean, I've already said to, got a very small engineering team who builds our platforms, and I've already said to them like, I'm expecting that, you know three full-timers should be able to output it like 30 full-timers by the end of the year, just by, smartly investing in leveraging tools.

Carl: 'cause one of our core values is work smarter, not harder. Like I don't care how many hours you work, I care that you worked really smart, that you,

Carl: Yeah.

Carl: do something in 15 minutes. You know, instead of four hours, like do it in 15 minutes. and that, I don't think that's a stretch to go, hey, like one person should be out to put times 10 within, around 12 months from now with the rate that the AI tools are building.

Carl: And also when you've got the attitude of, I'm looking for ways to leverage this, to get,

Carl: Hmm.

Carl: So I think that's happening and going to happen whether you like it or not. The more interesting part about the one person business is so interesting to me because. For the last probably 20 odd years, maybe just under 20 years now, [00:37:00] I've really had this idea of a business is not a business unless it can work without you. And the only real way you could achieve that, you could build a business that allowed you to truly be the business owner and not the operator of the business, was by building a team.

Carl: Yeah.

Carl: you could leverage tech to do certain components, but you needed humans. You needed great people. I don't think it's gonna be as fast as a year from now or two years from now that gets us there.

Carl: But I do think the ability for creative founder, the visionary founders

Carl: Who can now talk to AI and bring, put businesses into existence.

Allan: Yeah.

Carl: On a micro level, like for example, if you have an idea for an app right now, you can go to a tool like lovable.dev, for example,

Allan: Yep.

Carl: and you can just tell it what you want and it can whip it up. It might not be perfect, but like, if it's a specific, like if you're trying to build a complex app like Facebook or whatever, or even like automation is built it's a nightmare. But if you're like, I'm gonna build a headline generator app

Allan: Yeah.

Carl: like that, like it could whip that up in a [00:38:00] day. And you could then start selling that. I think that there really is an opportunity now for teams to get smaller. Are we at the point where a one man business, don't think we're there yet. think people will try, but I think you're still gonna need some level of assistance. Or if it's not assistance, you're gonna have other agencies.

Carl: For example, like you might not have a marketing team, but it's you plus automation entity, which is leveraging AI and people. Or you've got a finance, hopefully your accounting company now. not just relying on humans. They hopefully have some AI bookkeepers and other things in the mix. So you're, you might still be a one man business delegating to these other more specialized companies that are leveraging if they're smart. I think the future for the short while at least, is AI plus humans, which is where we are going.

Carl: eventually though there's a real reality that we might end up in a world where. You got robots and AI doing most of it and it brings up the whole economic model collapse because for there to be a supply and demand, there must be demand for there to be demand.

Carl: There must be people who are earning an income to be [00:39:00] able to pay for those things. You know, they can't just want something, they have to have money to pay for the thing in, in the economic

Carl: Yeah.

Carl: that we live in. And if that doesn't exist. And so, you've got people who swap their time for money like labor.

Carl: Yep.

Carl: You got property like ownership of IP or physical property

Carl: Mm

Carl: And things like that businesses and then you've got governments provided subsidies like UBI or Centrelink and those kind of things, like they're the three main income sources that. Drive the economy and

Carl: Yeah.

Carl: labor market falls out. There are people way smarter than me that have been thinking about this for years trying to solve this problem, and

Carl: I'm hoping that's years from now. But definitely a path I think we're heading towards.

Allan: Well, it's always fascinating talking to you, Carl. I, like I said, I always walk away with more energy, more ideas, more inspiration. So, um, I really appreciate your time today. Where do people find you? And of course we'll link in the description. Is that just automation agency.com?

Allan: Is that the best place?

Carl: Yeah, if you're a small business owner in particular or a small marketing team in a small [00:40:00] business, and you're just struggling to find the time to get stuff done, then come and check us out. Automation agency.com. help you clear your backlog. We'll help you launch faster, we'll help you just get it done,

Carl: take it off your plate.

Allan: Nice.

Carl: Love to do that. And if you just wanna connect with me on like socials and stuff, I'm just, Carl Taylor, Carl with a c.

Allan: Carl, thank you so much. you've come with profound wisdom, great ideas. Thanks again. Always enjoyed talking to you, my friend.

Carl: Absolute privilege and pleasure to be here. Thanks.

Allan: Thanks for tuning in to the Lean Marketing Podcast. This podcast is sponsored by the Lean Marketing Accelerator. Wanna take control of your marketing and see real results with the accelerator. You get proven strategies, tools, and personalized support to scale your business. Visit lean marketing.com/accelerator to learn how we can help you get bigger results with less marketing.

Allan: And if you enjoyed this episode, please leave a review or share it with someone who would find it helpful. See you next [00:41:00] time.

‍