Bryan McAnulty [00:00:00]:

There's a big difference between creators who casually use AI and creators who set up real AI systems to grow their business.

Alex Bacon [00:00:06]:

It's not prompting skills, but actually knowing your topic and your subject — knowing what good looks like is still going to produce the better outcomes.

Bryan McAnulty [00:00:14]:

Alex Bacon has spent more than 20 years helping companies scale. He helped grow a tech services business from £4 million to £38 million in revenue, supported the launch of a legal tech company that generated £22 million in revenue, and led teams through rapid growth and acquisition. Today, he's the founder of BrightKeel.AI, where he helps founders and growing teams use AI and automation to save time, improve their processes, and achieve more with the people they already have.

Alex Bacon [00:00:38]:

If you don't understand how you do something and you don't have clarity on that process and you're not confident it's the best process, then you can't really work out where to put AI into it. Whilst I and a photographer can both use AI to go and say, "Right, create me an image," I don't necessarily know what a good image actually looks like. A photographer is going to be able to prompt that and say, "Okay, use this lighting, position the subject here," and they're going to end up with a much better photo out the end of it, even though we're effectively using the same tool.

Bryan McAnulty [00:01:10]:

AI is at the top of everyone's mind right now. But if you imagine a year or two from now where everybody's got access to the same AI agents where we can just press a button and they do all the work for us, then there's an important question that we need to ask. What makes your business different or better than your competitors? I believe that it comes down to how you do things. I believe that in business, how we do things will be everything in this age of AI. So I started asking founders this question, and something surprising happened. These conversations were so valuable that I knew I had to share them, and this is why I'm starting this new series inside The Creator's Adventure called Founder Stories.

Bryan McAnulty [00:01:42]:

Hey everyone, I'm Bryan McAnulty, the founder of Heights Platform. Let's get into it.

Bryan McAnulty [00:01:53]:

Hey Alex, welcome to the show.

Alex Bacon [00:01:55]:

Thank you, Bryan. Thanks for having me on.

Bryan McAnulty [00:01:57]:

Yeah, excited to talk with you today. My first question is, in this kind of series of interviews, I've been talking with founders such as yourself and thinking about the premise: basically, if everybody has access to this magic AI agent a year or so from now where we all just press a button and the work is done for us, then what's going to make your company different or better than your competitors? And my belief is that essentially comes down to how you do things. The Mac versus PC commercials of Apple talking about the way that we approach building the hardware and building the software — that's what makes us different. So I'm curious, for your business, what's your take on that, and how do you see that for yourself?

Alex Bacon [00:02:36]:

Yeah, it's a good question. I think it's something that we're all going to face. And I think certainly at the moment it's about having the experience and the domain knowledge to know what good looks like. So I produced something the other week with a partner that I'm working with, and he was really impressed with what it came out with and said, "Oh, is that a particular prompt that you've put in?" And I mean, yes, it is. It's some prompting skills, but actually it's my knowledge of what to ask AI in the first place. How you ask it is one part of it, but actually knowing your topic and your subject, knowing what good looks like, is still going to produce the better outcomes. And I think that's the case across the board. Whilst — and I always use the analogy of a photographer — whilst I and a photographer can both use AI to go and say, "Right, create me an image," the reality is I don't necessarily know what a good image actually looks like. I might see one and go, "Yeah, that's good or bad." Whereas a photographer is going to be able to prompt that and say, "Okay, use this lighting, position the subject here, make it look like it was taken with such and such a lens or such and such a camera." And they're going to end up with a much better photo out the end of it, even though we're effectively using the same tool. So I think it's still going to come down to that side of things. And so certainly in my business — well, I effectively have two businesses at the moment. One is AI consultancy, and I'm helping organizations and leaders understand how to use AI and what AI looks like within their organization. So kind of AI strategy and adoption. And then marketing. So I'm a fractional CMO, marketing director. And obviously the two crossover quite heavily. And I see those two coming closer and closer together, because I think actually when you look at AI strategy and AI adoption, there's a layer that goes across the top — and I use this term of universal parameters. So I think there's something that is true across a business of how you use AI and what you want AI to look like and be like within your organization. But then you're also going to have those functional domain experts. So someone in finance, for example, who understands AI is going to get a lot more out of it in there than I could, because I don't understand the finance function. So I can go into a marketing function and really help them understand how they can use AI within their marketing, but with the background of expertise in marketing strategy, expertise in knowing what resources they actually need, what agents need to be built out — and not just doing something because you can do it, but doing it because it's the right thing to do for the organization. So I think it all comes back to actually using AI to amplify and give you superpowers around the skills and knowledge that you already have.

Bryan McAnulty [00:05:45]:

Yeah, I completely agree. I think for those who are somehow kind of worried that their skills are not going to be relevant, this is absolutely the reason why they are relevant, because you're going to be able to take that specialized knowledge and get the better result out of something where somebody else doesn't have the knowledge to be able to ask the right questions and put in the right information there. But it's not only that. Sometimes people may be on the opposite side. They may realize that, okay, yeah, I have this knowledge. Yeah, I'm not worried about AI. And then they go to use AI and they're not imparting that knowledge into the way that they communicate to AI. And so they say, "Oh, write me a blog post" or "Go write this thing for me." And they haven't actually told it anything about their specialized knowledge. And now you end up with AI slop. And so you have to make sure that you're putting that knowledge into it and explaining your expertise where you need to in order to direct it the right way as a tool.

Alex Bacon [00:06:51]:

Yeah, exactly that. And I know it's a well-trodden cliché, but I think thinking of your AI as an intern, as an employee, is important. Because if you had someone new joining your organization and you were asking them to write a blog post, for example, one of the first things you would do is bring them up to speed on your organization. Get them in touch with the subject matter expert. Make sure they understand your tone of voice, your brand, what key messages you want. So all of those things are completely true for AI as well — you need to give it that context, you need to give it that information. And I think the other thing to do as well is to understand how you break down a task into a series of individual tasks. So taking the analogy of a blog post — or I quite often use the example of doing a press release — if you simply say, "Write me a blog post" or "Write me a press release about X," you're going to get something back that's pretty mediocre.

Bryan McAnulty [00:07:58]:

Right, you just say, "Go do it," right? "I'm just going to do it."

Alex Bacon [00:08:02]:

Exactly. And so actually, when you think about, well, how would I actually go about doing that? You might say, well, first of all, I'll interview the subject matter expert. Okay, well, let's go off and do that. Come up with some questions, interview them, get that. Then you might plan out a structure of what that blog post or press release looks like. Okay, let's have AI sort of use that and get the structure, and make sure we're comfortable with the structure first. Then you might go out and look at the competitive landscape and say, "Okay, is this a unique idea? Where is it going to fit into the market?" So again, you can build an AI agent to go out and do that part. But you're instructing it to go through these series of steps and building it up. And then you might say, "Okay, now I want to dictate into it to get my tone of voice into what's being produced." I might then almost write it in reverse and kind of write it, but then have the opening paragraph I write at the end, once you know the rest of what you're writing. Same with the headline. You come up with sort of a killer headline, or a stat, or quotes that you want to put in. And then you might go and research where you want to post that. So whether it's just a blog on your own site, or whether it's something that you want to go and sort of guest post or put in external links. So all of those things culminate in a much better output. Knowing how to do those things is typically from your own knowledge of how you approach things and how you've done it in the past. And so take that same example and apply it into anything — whether it's legal, whether it's a consultant, whether it's the photographer that I gave the example earlier — for them to be able to say, "Well, actually, this is how I break down the task, and this is how my skills and knowledge are applied in different parts of that task," that's going to really give you a much better outcome.

Bryan McAnulty [00:09:58]:

Real quick, Bryan here. You know me as the host for this podcast, but what you might not know is I'm also the founder of Heights Platform. It's an all-in-one platform that over 10,000 creators have used to build their online courses, communities, and digital product businesses. We recently added some awesome updates to Heights AI to help you turn your idea into a viable business. Heights AI can build entire product offers, review your content, and even coach you on how to grow. You can try for free for 30 days. Links in the description. Now, back to the podcast.

Bryan McAnulty [00:10:26]:

Yeah, I completely agree. And that aligns with what I was talking about — that the process is almost everything, and the how you do things. I remember a few years back when we started using these AI tools and ChatGPT and everything, people would say, "Okay, act as an expert" or "You are an expert at this thing." But especially now, if you think about how that is, there are reasons why it made sense to maybe do that a tiny bit before. But especially now, just saying, "Oh, you're an expert marketer" is very ambiguous, and anybody can do that. And then anybody has — we're in the same problem, right? Then everybody's doing the same thing as you. How are you going to stand out? And so you said exactly what I would agree with, that it's the processes. So telling the AI, "Well, these are the things that we have to do, and this is the way that we have to do them," now you end up with a result that is much more similar to what you could have maybe been able to create yourself, or more similar to the thing that you want. And that's how you're going to come out ahead of the business who is still saying, "Hey, you're an expert marketer, just write me a blog post." I'm curious — before you started this consultancy, what frustrated you about the industry, and how did you work towards kind of solving that?

Alex Bacon [00:11:42]:

So my background has been largely in IT services, but in commercial roles. And so most recently, before doing my own thing, I was marketing director for a fairly large Microsoft partner. And, as you know, I run my own podcast, Scaling with AI. And the premise really for that, and then subsequently the business, was that there was this mismatch between what the vendors — so the big players — were kind of promising, and then the reality of what people were seeing. And this is going back probably 24 months now, and we were one of the first ones to have Copilot licenses. And you're being sort of promised it's this magical tool that can do everything. And actually, we got given the licenses and kind of thought, "Well, whilst it's cool, actually it's not really that great." And you suddenly realize that there's a big gap, both in terms of the capabilities, but also — you know, we're being given this really powerful tool that no one's been taught how to use. No one's really thought about how to use it. And so it's identifying really that there's this gap, and being able to, first of all, through the podcast, understand well, how are people actually approaching this? How are they using it in their day-to-day, sort of beyond what we're being told by the big tech giants of what's possible — actually, how is it being applied? And then through the business, through my business BrightKeel.AI, it's then helping organizations make sense of that and say, "Okay, you're hearing the headlines. What does this actually look like for your organization? Where's the gaps in terms of your ability to adopt? Where's the opportunities that you can use it? How do you take your business on this journey and start to adopt it?" And that's really kind of where it's gone, and it's going very well, luckily.

Bryan McAnulty [00:13:37]:

Yeah, it's great. It's very interesting, and it's difficult, and I think there's a need for people like yourself, because you talk about the Copilot licenses back then and how it kind of overpromised what it could do. But now we move forward a couple years, and I think that not necessarily every tool, but a lot of these AI agent related tools now, all of a sudden the capability is so far beyond that, and is, I think, beyond what a lot of people even assume is possible. But it's very difficult to even understand that without really experimenting and digging in, because we still have the same text input, right? So the text input a couple years ago could answer a question. Maybe hopefully it could access some of our docs or something like that. But now it's like the text input can potentially go and work for a couple days on these very large tasks. So it's definitely — there's a lot to have to be figured out by not only an individual entrepreneur, but especially an entire business and their employees, and learning how to kind of work in this new way.

Alex Bacon [00:14:36]:

Yeah. And I think when I reflect, I guess almost on my own journey of starting to use AI and figuring it out — sort of back then it's looking at prompting and then building up to things like custom GPTs, and a lot of those foundational things of, what does a good prompt look like? How do you break up a task into those individual tasks? I think those are quite important foundations to be able to then move into sort of the agentic world and start to think about how you build out those agents and what those agents are applied for. And I think I'm almost lucky in that respect that I've kind of gone on that journey because I've been there as it's matured. And I think actually, if you're coming into it now and you're just jumping from zero knowledge into agents, it's kind of easy to do in some respects because you can spin up Claude or ChatGPT and put on agent mode and create an agent. That's, you know, that's great. That's quite easy. But I think that you're almost missing that context part if you haven't sort of gone through and understood, well, actually, where does this fit in? What do I want it to do? What do I not want it to do? What's the individual task that I'm asking it to do, and the sources and the context that I'm giving it to do it? And all of that thinking helps get a better outcome, even when you're sort of spinning up an agent.

Bryan McAnulty [00:16:05]:

Yeah, you said it so well. I was going to jump in and say the same thing, that I feel lucky that I was able to be involved with this stuff earlier. And I think anyone watching or listening to this should probably feel the same, because you have this advantage of knowing the failure modes, and how these things worked, and what they could and couldn't do back when they could barely do anything. And so you kind of have this built-in understanding, whether you realize it or not, of like, yes, these tools can do some of the same work as people, but the way that they work is different. Having that understanding can be very helpful, because now when you're working with an agent that's doing all these things and working for a longer time and dealing with all this context, it is very helpful to understand well, what is the way in which I can communicate, or that I can make sure it has the right information, to actually get the positive result. And it's definitely a lot more kind of opaque or hidden for somebody who's just starting right now. So, you created the AI Maturity Pyramid. Can you kind of explain what the different levels are, and how an entrepreneur could figure out where they sit today in that?

Alex Bacon [00:17:12]:

Yeah. I haven't got it in front of me, so it's going to be stretching my memory a little bit. So it's something that I walk through with businesses to help them really understand where they are. And sort of at the top, you've got fully autonomous agents running, integrated with your data. And it's kind of the utopia of what we're promised ultimately. And so the maturity pyramid sort of works up through from there. The foundational level is actually around culture and curiosity from the leadership team. And that's not necessarily where everyone starts, but it's the foundation on which to build, because a lot of what I talk about is saying, well, actually, it's down to the leadership team to create the right culture for people to feel that they can safely experiment, and to make sure that they know why the organization wants to use AI in the first place. Is it using AI for the sake of it? Is it using AI because it wants to reduce the headcount, or is it using AI because it wants to enhance the team? And a lot of that comes back to the business strategy, not so much the AI strategy — but the business strategy of what's the purpose of the business, what's it trying to achieve, what are its values — and putting those things in as the foundations on which to build on. And then really, it's then where on that journey, sort of between there and that utopia, the pyramid walks through. So it starts off with experimentation and prompting, and then starts to look at how they're using that in a more structured and strategic way — putting in guidelines and policies so the organization understands where they are and what they need to do. And then the layer that often blocks a lot of people is then data. Because you can start down that agentic route, but actually, if you don't have clarity on your data, you don't have data that you can trust, you don't have data in a structured, secure way that you can feed into AI, then actually you're quite limited on where you can go beyond that. And then it goes on to more automation, integration, and more in terms of the agentic side of things. And that's what moves you up to that utopia that we're all looking at. And so really, it's a model for organizations to identify where are they, where are the gaps, maybe what steps have they missed along the way, and how do they move forward.

Bryan McAnulty [00:20:01]:

Yeah, I think for probably the average person watching or listening to this, you can understand like, okay, I get it — messing with ChatGPT and getting a response versus that utopia of these agents doing all this work for me. The data part, I agree, is an especially big challenge for large organizations where this data is everywhere and shared between so many different people, and there's a lot that has to be considered there. But it's still a problem for the smaller teams and the solo entrepreneurs as well. And I think that it's been one of my concerns in building our software products, where customers would say, "Well, it would be great if the AI would just be able to maybe do support for my customers, or maybe do this or do that." And the problem that I've always seen has been that the average entrepreneur does not have the data. And so either does not have the data, or the data is in their head or something like this. And so you need to get that out and get that accessible to the AI, so it's not just like, oh, somebody said this — okay, I'm going to make something up. So yeah, organizing that and being aware of that is very important. And that kind of brings me to the next question, which is: why do you believe that companies should improve their processes before adding new AI tools? Because I agree with this as well, but I want to hear from you.

Alex Bacon [00:21:14]:

Yeah, I think — I mean, I'll just close off sort of the data part first, because that structuring of the data, and it's linked to the process piece as well, is really important. So even in my own business, I'm probably quite risk averse in what I allow AI to have access to. So I've got certain folders and I've got certain drives that I'm happy to connect into.

Bryan McAnulty [00:21:47]:

And I think that's a good point, actually. Sorry to interrupt, but I want to say that I think some people might be hearing like, okay, these two guys are talking about AI, so they're the kind of people that just turn on something like OpenClaw and have it go run wild on their computer. And I'm absolutely not going to do that.

Alex Bacon [00:22:07]:

No. And I'm exactly the same. I wouldn't touch OpenClaw. And even what I use — as I say, I've kind of got certain folders, certain drives I'm happy to connect it to. But there's a lot of context in my other files and my other folders that would benefit AI to know about me, but it's not structured in the right way that I'm comfortable to let it go in, because a lot of it's just in my historical personal folders. So I've got work folders in the same drive as I've got bank details, photos of my kids, you know, all of that sort of stuff. And so I don't give it access to everything. And I'm in the process of sort of starting to try and move things over to make sure that I've got certain folders that I can put it into. But yeah, as you say, that's on a very simple solopreneur sort of journey. So you can imagine then, as you get into the context of a business, it becomes even more complex and even more messy. And then, sorry, to move to your question. On processes — I think one of the main things is if you don't understand how you do something, and you don't have clarity on that process, and you're not confident it's the best process, then you can't really work out where to put AI into it. And again, taking a really simple analogy of thinking of it as a new starter — if you're not able to articulate how you approach something and the tasks that are involved in a process, then you can't expect a human or an AI to start to do that. Let alone then, when you start to think about integration and the flow of data and the flow of information, that if you're not confident you're doing it in the right way, then all you're doing is kind of amplifying the noise and amplifying the mess. And probably the biggest part of all of that is understanding where you still want a human involved. So again, that idea — a really simple analogy that we used earlier of building a blog post. As we said earlier, if you just say, "Write me a blog post," then you kind of get to the end of it and go, "Oh no, it's completely missed the mark." Whereas actually, if you're putting in that human check and saying, "Okay, well first of all, I want to see the structure before you go away and write it." "Okay, yeah, the structure is good. Okay, here's — a bit more, I want you to focus more on this angle." So again, if you take that same thought process to a more complex process, you really start to understand, okay, well, here's where I can put in a human to make sure that I steer it in the right direction, before it completes everything and certainly before it produces something that I go and push out into my customer base, for example.

Bryan McAnulty [00:24:56]:

Yeah, the AI is really just going to amplify whatever kind of comes into it, in a way. And if you're not understanding what you're asking it to do, you're not going to be really happy with the result there. And yeah, definitely realizing where are the parts that it's going to help me with versus where are the parts that I need to be involved. And I think one of the distinctions I like to think about with this is that it can help do the work, but the ideas, the taste, and the why behind the things — that has to come from the human. And so, back to the blog post example, if you had a blog post where there had to be all this research done to find these numbers and stats and things like that, that's the perfect thing for the AI to go and do. So provided that you're explaining what your actual intention is, then it can do that. And then the ideas behind it, the creative part, that should come from you. Because yes, some models are better at being creative than some others and things like this, but I think humans are still the best, and we understand what we actually care about. But yeah, I agree. Your ability — how quickly and clearly you can communicate some idea or intent — is like the ultimate skill nowadays.

Alex Bacon [00:26:07]:

Yeah. And I think just going back sort of almost full circle — part of that, we talked around the importance of understanding the process, but also the importance of understanding your business. And this is where my marketing background goes alongside the AI side of things. To that point that we said about amplification — actually, if you've got clarity on your ICP, your ideal customer profile, and you've got clarity on your messaging, so you know what you want to say and who you want to say it to. You understand your customers. You understand why they buy from you. You understand your competitive differentiation. All of those things that would make up your marketing strategy, those are fundamental for putting in in terms of AI, because again, back to that amplification idea — without that, you're really in danger of amplifying and just having a real scattergun approach of loads of different messages targeted at loads of different audiences. And it might look good and sound plausible, and you might think it's great, but actually, if you're firing that out, then all you're doing is confusing the market, and you're not getting that consistency, you're not getting that clarity, which goes back really to the first principles of marketing.

Bryan McAnulty [00:27:27]:

Yeah. And I think being able to — the problem of the way that AI works, that I think is often misunderstood, again by people who maybe did not have the opportunity to use these tools early on, is that it just wants to do the work for you. Whatever you say, it's going to do the thing. And so if you told it to do it, it's going to do it. It's not going to stop you and push back and say, "You know what, I kind of want to know why we're doing this. I kind of want to know who we're actually targeting here and how we're planning to help them." It's not going to just naturally say that to you. And so if you're not sure that you've communicated all these things, what you can do is instead ask it and say, "Is it clear why we're doing this? Is it clear where we're differentiating here?" or this kind of thing. And then get that really clear. Once you have that really clear, now you can go and tell the AI to go and do that thing. This brings me to the next thing, which is that I know even myself, using these things all day long, I still will come across some kind of AI agent tool or something like that and think, "Wow, this seems really cool," but you still have just this text box, and so it's like, "Well, what can I actually do with this?" I find myself even struggling of thinking there's a lot of potential here, but what do I actually do with it? So how can an entrepreneur or a small team identify that first thing that they want to delegate to AI?

Alex Bacon [00:28:49]:

It's a good question, because there's kind of several ways to approach it. So when I'm working with organizations, I tend to look at what's the opportunity there. So what actually is it going to do to enhance what you're trying to achieve? So a lot of that comes back to some of those earlier points I said of having clarity on what the business objectives are, why do you want to adopt AI, and what are you trying to achieve? And then recognizing, how is it going to help you achieve that? Because actually, if it looks cool and shiny and it does something, but actually it doesn't get you closer to where you're actually aiming for, then it's a distraction. So one of the first things is understanding how it's going to help you, what's going to be the upside. And then looking at what does that mean in terms of our readiness? So do you have — when I'm talking with a business, more it would be data availability, clarity on the process, maturity of the people and the skill set that's needed there, the tooling that's needed, and what time and resources are needed — both, I say, time and money. And typically then, you kind of end up with a scoring that says, okay, this is the potential upside, and this is our readiness or ease of being able to adopt it. And you can then say, okay, well, here's a bit of low-hanging fruit. Here's something that actually is going to be a bigger project. And typically, they're kind of categorized between what can you just do with AI now — you know, what can you just put in as a prompt and use? What do you need to build, but you can build yourself? So kind of what agents or workflows might you want to build? And then what maybe is a bigger project, so needs maybe a data project or an integration or an automation project or some custom build. And you've got those three tiers, but you're able then to map it to the upside from it and your readiness for it. So that's kind of for a business. I think for individuals, it's kind of similar but on a more straightforward level of just saying, well, what do I want to achieve? What's a problem that I'm actually trying to fix? And if I had a magic wand, what would I fix first? And I think thinking about that, rather than saying, "Here's a cool tool, what can I use it for?" I think if you start with, "Well, what would I just want to solve? What am I doing that I'm not enjoying, or what am I trying to achieve? What would I like to do that I can't do?" — and then being able to find the solution for it that way.

Bryan McAnulty [00:31:26]:

Yeah, it's a great point. I think also people are eager to just have it do everything and reach that utopia of like, wow, I have this AI doing all the work for me. But there are sometimes these low-hanging fruits where it's actually a very simple thing that you could have AI do that would take a very large amount of your time if you had to do it otherwise. And not everything is like that, but sometimes you can discover one or two things like that, and just getting started with that can be a massive difference for you. I know one example I'm thinking of for us is that we used to have a human go in and look at every new creator that came to Heights Platform and try to understand what is it that they're actually trying to build, and what is the niche that they're in, and all of that. And then we would see how we can help them from there. That's a lot of work, to have a person have to go and check every single thing. Now AI can do that, and it can figure out very quickly and very cheaply what is this person doing here. And so now we just have that information, and now our time is able to be spent on the higher leverage things, and the actual cost of running the AI through all of that is like pennies, even though it would take many, many, many hours of human labor.

Alex Bacon [00:32:38]:

And I think it's a good example, because you're not only benefiting from the immediate task, but what you're doing there is building up an information and a knowledge source. And that's probably more the bit that gets lost if you've got a human doing it. They might do a very good job at looking at an individual case and saying, "Okay, right, here's Alex Bacon. Here's some information about him." Great. But actually, what AI is doing is building up that picture. And then you're looking across your membership and saying, "Okay, here's the trends. Here's the insight." Whether you use it immediately or whether you use it down the line, you're really capturing that. So I think thinking around those additional benefits is important as well. I mean, touching on what you said, one of the analogies I sometimes use is just saying, well, just because you can do something doesn't mean you should. And actually, you know, if you go and pay £200 for your Claude license and you build something that goes and researches people and blasts out a thousand emails every day or every week — is that right for your business? Because actually, that's not new. You could have probably paid someone in another country £200 a month to do that for you. In reality, you could do that now without AI. But you probably wouldn't, because you'd probably be fearful of the quality and the reputation damage, and it probably doesn't necessarily fit your business model. It might be a numbers game, but actually, for a lot of people, it's about managing their reputation. And so just because you can do something with AI doesn't mean that you should immediately jump at it. I think you still need to understand the context to say, well, does that represent my business? Is it the right model for my business and where is it going in the future? So again, that example you gave, saying, well, actually, yes, it's much more efficient, but also you're probably getting a better output from it as well.

Bryan McAnulty [00:34:38]:

All right, I want to get into like a quick lightning round here. So, looking for a 10-second answer or a quick take on these next few. What is the biggest mistake that entrepreneurs make when starting to use AI? Sorry, not a very quick one.

Alex Bacon [00:34:46]:

This is probably the toughest of all of them, I'll let you know that.

Bryan McAnulty [00:34:52]:

What's the biggest mistake that entrepreneurs make when trying to use AI?

Alex Bacon [00:35:02]:

I think initially overestimating it — sorry, sorry — initially overestimating it and then underestimating it.

Bryan McAnulty [00:35:12]:

Yeah, I like that. People start early to try to have it do everything. It doesn't work well. Then they get a little bit of success and then they have it do this one thing, but these tools are evolving so fast that now it can probably accomplish a lot more than you imagine.

Alex Bacon [00:35:27]:

Yeah, exactly.

Bryan McAnulty [00:35:29]:

Is prompt engineering becoming overrated?

Alex Bacon [00:35:35]:

Yeah, I think so. I think the tools are getting better and better at understanding, and I think a lot of the things that we spoke about is not necessarily the quality of the prompt per se. It's more about understanding the context, giving it the context and understanding the process, more than the actual prompting itself.

Bryan McAnulty [00:35:57]:

Yeah. I find that whereas people might have been concerned about doing this a few years back, now you can kind of just ramble on with a voice-to-text thing to the AI, and your thoughts may not be super clear. But if you get out all that information and all the right information, it's a lot better than either not sharing that information or spending hours to make sure it's all perfectly formatted and explained in the absolute clearest way.

Alex Bacon [00:36:24]:

Yeah. And the best thing you can do is then ask AI to structure that into a prompt. So if you're building out a prompt for an agent or something like that, then as you say, ramble into it and then ask it to give you the prompt.

Bryan McAnulty [00:36:39]:

Should every creator build an AI agent? And I'll preface this by letting everyone know we're in end of July 2026. I think the answer may be different now for you than it might have been, I don't know, in like January, when OpenClaw first got hyped up.

Alex Bacon [00:36:56]:

Should everyone do it? Probably yes. But does everyone need to do it? Not necessarily. I don't think everyone is there yet. And I don't think that's a problem at all. A bit like we were saying earlier, depending on where you are on your journey, I wouldn't rush to build out an agent for the sake of it. I would get comfortable with it. But there's going to be benefits to everyone. Everyone will have a use case for an agent. You just need to be able to figure out what it is.

Bryan McAnulty [00:37:29]:

And then on the show, I'd like to have every guest ask a question to the audience. So if you could ask our audience anything, whether something you're curious about or kind of just want to get people thinking about, what would that be?

Alex Bacon [00:37:41]:

Probably, touching on everything we've spoken about: do you really understand your customers? And the reason I ask that is because I think there's so much insight we get from our customers. And the danger is using AI assumes knowledge or gives you the generic answer. Actually understanding your customers, why they buy from you, the language that they use, is really, really valuable for marketing and AI.

Bryan McAnulty [00:38:13]:

Yeah, I like that. Alex, it was great talking with you. Before we get going, where else can people find you online?

Alex Bacon [00:38:19]:

Yeah, thanks, Bryan. So I'm quite active on LinkedIn — Alex Bacon, or Alex Bacon from BrightKeel.AI. You can look at BrightKeel.AI, or my podcast, which is Scaling with AI.

Bryan McAnulty [00:38:34]:

Awesome. Thanks so much, Alex.

Alex Bacon [00:38:36]:

Thank you. Thanks, Bryan.

Bryan McAnulty [00:38:41]:

I'd like to take a moment to invite you to join our free community of over 5,000 creators at creatorclimb.com. If you enjoyed this episode and want to hear more, check out the Heights Platform YouTube channel every Tuesday at 9:00 a.m. US Central. To get notified when new episodes release, join our newsletter at thecreatorsadventure.com. Until then, keep learning and I'll see you in the next episode.