Every AI Mistake I Made (So You Don't Have To)

I spent months on Friday night YouTube spirals trying to figure out AI without a shortcut. 

Here's what I learned — and the five mistakes that cost me time, money, and sanity. 

Episode Summary 

 The 5-Question Checklist: Use this to filter ANY AI training before you invest --- 

✨ Connect With Brooke DM me on Instagram → @wright_mode 

Want to go deep on AI in your business? Check out the Dream Team Workshop (coming in 10 days) and the Membership (by application) for ongoing support and co-work sessions. 

Visit wrightmode.com for more

• Want to work together? Check out my offerings at Wright Mode
• Connect with me on Instagram → @wright_mode

Transcript

0:23

Welcome back to the podcast. It is Monday. I'm recording this after our touch footy grand final over the weekend. Safe to say I am a little bit sore today when we start on the win. But we played our hearts out, and that's all that matters, right? A for effort. So today I want to go through a little bit, I guess, of a vulnerability share. I'm gonna walk you through every AI mistake that I made. And what to ask before investing in AI training. So it's a little bit of a journey, I guess, of where I've come from. And I guess things that I wish that I had known before getting to where I am today. So let's take it back to the beginning. So I wanna tell you about my Friday nights in 2024 and the start of 2025. Not the fun kind, not the out for drinks Friday nights. The kind where I am sitting at my laptop, it's 8:30 PM the kids are in bed, there's 17 YouTube tabs open, and I'm trying to figure out why my custom GPT won't connect to notion. Copying code that I didn't understand and pasting error messages and screenshots in to chat. GPT, getting more confused, going around in circles, trying again, and the whole time thinking, surely someone has figured this out. Like, why can't I just find the answer? I've spent a lot of money on AI training in investing in my own learning. I've spent even more time. Trying to figure things out on my own, and I've made a lot of mistakes along the way, so I'm gonna share those mistakes today. Not so much because I love airing my chaos because each one taught me something and each one turned into a question I wish I'd asked before I invested in any training. So, whether you are thinking about investing in AI learning. Mine or anyone else's, these are the questions that will help you choose what's right for you, not just, you know, based on the best marketing or fomo. So the first proper AI training that I invested in was a VIP training day. The training itself was good. I learned how to build custom GPTs. I got introduced to tools I hadn't heard of. It was genuinely valuable, but here's what happened After I'd go to implement something and it wouldn't quite work or it wouldn't fit my business the way that they'd shown it, there was no troubleshooting, so there was no like, here's what to do when it breaks, no, here's how to adapt this to your own situation. And I just really felt like I was on my own. And this is the same problem with YouTube. Tutorials by the way. So they show you, this nice, ideal scenario. They don't show you what happens if your business is slightly different or the tool has updated or something just doesn't work. You are left trying to figure out that gap yourself. So the question I wish I'd asked was, what happens when these things don't work? Is there any troubleshooting support or am I on my own? Because learning the thing is only half the battle. Knowing what to do when it breaks, or knowing how to fit it to your specific business needs. That's where most people get stuck. So after that training, I had enough knowledge to be dangerous, I guess. Take it and run with it. I knew what was possible, I just couldn't make it happen. And that's when these kind of Friday night sessions began. So I had this project I had seen on YouTube someone using custom GPTs to like send out like very basic, like to-do lists. And I wanted my custom GPT to send content ideas directly to my Notion Dashboard. So this was before Notion had any connectors built in Notion AI didn't exist. So I was trying to set up an API to do this manually and it just would not work. So I remember watching this YouTube tutorial about these to-do list and I was like, okay, so. That framework exists. That's possible. So surely if they can do that, then I can send email marketing content that I'm, you know, working on with chat once I'm happy with it over to my content dashboard. So I'd watch like, these YouTube videos. Try something, it wouldn't work. I was copying error messages into chat GPT and ask it like what was wrong. Sometimes it would help or sometimes we'd just end up in this loop and I'd be like, I've tried that. You told me that. That's not possible, that sort of thing. So this went on for that specific project, probably took me about two weeks to actually get working. And like loads of hours in that trying to troubleshoot this. So it was a bit of a cycle of like trial fail. Troubleshooting and chat GPT, get frustrated, walk away, come back with fresh eyes, and then eventually figure it out. the whole time I felt like I was the only one doing this. Like I couldn't really talk to anyone about it because I didn't know anyone else who was in that deep or like pushing the envelope with this staff. My friends weren't exactly building custom GPT actions on a Friday night. and I couldn't really explain to them what I was trying to do. It was super exciting when things worked, but it was also a little bit lonely. Like, my husband's on the couch and I'm like, Hey, I'm trying to do this thing. And then, you know, he would see me like just getting frustrated when it wasn't working. Because I was like, why can't I make this work? Like, I know that at the core this is possible. So anyways, I got it working. But one of the questions I guess I wish I asked through that process was. Like with this training that I had taken will I be learning this alone or will there be support? Is there someone that I can ask when I'm stuck? Because the lonely DIY path is possible. I prove that, but it takes way longer and it's way harder than it needs to be. Like now that I look back on that and that template is available in. With the resource hub for the AI Dream Team workshop, and multiple clients have just set that up because I've made sure that I've stepped through all of the pieces, and it's like, if this happens, do this. If this error comes up, do this. Because I spent many, many hours testing all the paths to kind of give everyone the shortcut. Here's what actually saved me on those Friday nights, so whenever I was troubleshooting with chat GPT, I started to ask it. To explain the why it was suggesting these changes, not just fix this, but fix this and tell me why. Like, tell me what you're doing here. And that changed everything because instead of just like copying the solutions that it was giving me that I didn't understand, I started to actually learn like the logic. I started seeing patterns. You know, and learning why this broke. But that also helped me when chat was getting into a loop because I was like, hang on, you've told me that this can't work for these reasons, but now you are suggesting that. So when it was kind of getting stuck in these loops. Because I had asked it to explain its logic and I was starting to learn it. I could then pull it up when it was giving me answers that were clearly wrong. And so over time I needed to troubleshoot less because I understood what I was building. Not just, you know, do this, do that. I was understanding the why, and that's the difference between learning tools and learning how to like. Think and build systems. So AI training teaches you the tools. You know, press this button, do this, write this prompt. Here's the template, which is fine for getting started, but if you wanna go further than that, and if you want to understand why things work when something changes, you can adjust. Then you need to really dig into that. Why? Because if you don't understand why this stuff works or how this stuff works, you are just stuck every time something changes. And every time your situation is slightly different than say the template or the video that you're watching. You're stuck. The question to ask is, will I learn the why behind this or just the, how will I understand enough to adapt and troubleshoot on my own? Because the goal isn't to follow instructions forever. The goal is to actually get it, have that foundational knowledge so that it doesn't matter what tool comes out, what update comes out. You understand? On that foundational level, like this is how these things work, so I can now adjust based on these changes. All right. So mistake number four, I hopped from tool to tool like most of us do. So I started building automations in make.com. I spent ages learning it and got quite a few things working and then I discovered N eight N and I was like, oh, this looks heaps better. Now, in this case, N eight N was actually heaps better for my brain. So that visual canvas I really love and it just makes sense to me. And also when something breaks, it spits out the error in JSON and I can just cut and paste that into Claude and figure out what went wrong so that it's just so much easier to debug than make where it doesn't allow that easy copying and pasting into. Sometimes switching is the right call. Like I am so glad I made that change. But I did also waste a lot of time learning tools I didn't stick with. So in this particular training, they use tools like Cassidy and like they look really great, but that gets really expensive really quickly. And it wasn't fit for purpose for my business. Yeah, so, wasting time learning tools that you don't stick with, is not great AI strategy, there's a graveyard of subscriptions and half finished experiments out there. And look, I'm guilty of that too. So I guess the question I wished I had asked earlier was, am I learning this tool because it fits my business or because it's shiny and new and I've got fomo. And is this training gonna help me focus or just add tools to my list? Because more tools is not the answer, the right tools used properly is the answer. If you are not satisfied with chat, GPT. Like in my workshop, we have a look at how you're using it and maybe you're just not optimizing it or maybe something like Claude is better suited to your business needs, then it makes sense to switch. But for a lot of people, maybe you're just not using chat in the way that you need to get out of it what you need. And that could be an easy fix. So mistake number five. I didn't start with what I actually needed, so I did things backwards for a bit too long. I would look at something in the AI space, like a tool and be like, Ooh, I should learn that. And then try and fit that into my business somehow instead of starting with, okay, what do I actually need? What's the problem I'm trying to solve? And then figuring out which tool or approach fits that. So that notion, content dashboard, project I, after I set up that custom GPTI was like, I need a way to stop copying and pasting content from chat into Google Docs or Notion. That was the actual problem. It was that kind of like tab switching context. And I was like, this doesn't feel like a workflow to me. And so that was why I created that solution. So when you are coming in looking at where AI fits in your business, you need to think about like. Does this, like what problem am I trying to solve? Does this training solve that problem or am I just, you know, do I have fomo? Will I build something specific to my business, or am I just learning generic AI skills? Because generic AI skills will get you to a certain level. But it's building something specific to your business and your problems. That is where you will see a like ROI from your AI investments. If you are not looking at it through that lens, then you will be stuck copying and pasting in chart GPT and that really like I, yeah, you. Can do that and have a little bit of impact, but that's not gonna be a significant impact on your business. Like, and it is a good place to start when you are starting at nothing but very quickly, you need to move out of that to actually get leverage in your business. All right, so last mistake. And this one's a bit sneaky, so I'd learn something, watch a tutorial, do a training, whatever. I would just move on to learning the next thing without implementing what I actually just learned. And then you don't actually retain that knowledge. So you'll have all this knowledge in your head that you've never actually applied if you aren't implementing after learning, you are gonna forget half of it. That content context will be gone, or you'll just never get around to actually implementing it. So learning without implementation is just entertainment. Like you're just watching stuff that looks cool, all right, onto the next YouTube video. It feels productive, but nothing actually changes because. Until you start getting in there and building stuff, you don't really know what goes wrong. So like when you're watching a YouTube video, you don't get to see the Mistakes that they have made in that video, which happens, things just don't work. Even if you follow it, like step by step by step, sometimes like a setting is just switched off in your profile that like throws an error. Little things like that. And so unless you start to build, you won't see that. And you just see these perfectly edited YouTube videos of that whole process start to finish with none of those Bits that they have to troubleshoot through. And I always say to clients, automation is 80% problem solving. So you have to be good at problem solving and you have to not give up essentially when things don't go right, because it happens to the best of us. You've just gotta figure out what that problem is and solve it. The question when you are looking at AI training to ask is there time and space to actually implement this? Or am I just gonna consume this and move on and not actually build and and learn? So this is why courses with like, 38 different modules often don't work. By the time you've watched everything, you haven't implemented anything. And yeah, it looks good to have this big padded out course, but if you are. Overwhelmed at the end of it, it hasn't actually done anything for your business. So here's the thing. Every single one of those mistakes is what I have built the opposite into the AI Dream team workshops. So no troubleshooting support in the workshop. You get time to troubleshoot with me. When something's not working, we figure it out together. You are not on your own. You're not learning alone. It's live. We are in the virtual room together, so you can ask questions in real time and you're not just watching a video, hoping, someone jumps into the comments section to help you out. And also learning tools, but not the why. So I start every workshop going into, how a large language model works. And we go deep into. How this stuff actually works. So you understand enough to adapt, to troubleshoot, to build whatever you need next. If something isn't going right or you're not getting the output that you want, you have that knowledge to be like, oh, hang on, this is how that works. So I need to adjust that, and that will, sort that prompt out or that workflow. Too many tools, not enough focus. So we start with what you need. Strategy first always. What's the problem you're trying to solve? And then we build for that, not just random skills, specific solutions to your business. And then, yeah, no time to implement. So I've actually restructured this now from one four hour session where we were building. But it is now two, two hour sessions with three days of implementation in between. So you can go and learn something, go out, implement it, build, come back with questions or anything that hasn't worked or that you are struggling with, and then we get to troubleshoot that together. So the workshop is literally designed around every mistake I made and every kind of, I guess, gap that I saw in AI training. Because I don't want you suspending months of Friday nights in YouTube spirals. I did that so you don't have to. Okay, so the next one is coming up in 10 days. So if you are. Sitting here listening and thinking, I've been trying to figure this out on my own and it's not working. This is the shortcut. I will drop the link in the description and you can come and build with me. Alright, now let's get into the AI confession booth. I definitely mass applied learn by doing energy without any learn from someone who's done it energy for way too long, like stubborn levels of, nope, I'm just gonna figure this all out myself. And look, I did, and I am very grateful because I learned so many skills along the way. But I also spent months of troubleshooting alone when I probably could've just asked someone for help. If I had had access to someone who was further along than me, that would've like just sped up my learning by months. So the moment I started investing in being in rooms with people ahead of me. Not just in AI, but in business spaces, everything has sped up. Not because I couldn't do it alone, but because I wanted to accelerate that learning and that business growth. So that's my confession. I was stubborn about asking for help, and it cost me time and my Friday nights. So don't be like past Brooke. Just get the shortcuts. From me. Alright, so you are not another to-do list for this week before you invest in any training. And it doesn't need to be mine. It could be someone else's. There's heaps out there. There's plenty of pieces of this pie. I want you to ask yourself these questions. One, is there troubleshooting support that is so important? Will I learn alone or with someone? Will I understand the why? Not just the how will I build something specific to my business? And is there time to actually implement and then bring any of that troubleshooting support back? Okay, write them down. Use them as your filter and make sure whatever you choose actually ticks those boxes. Okay? So that's the checklist that I want you to measure any AI training by. All right, that's the episode. You're not behind. You're just missing the shortcut. And if you do want the shortcut, it's the AI Dream team and it's happening in 10 days. I'll link it in the show notes as I said before, and I will see you on the next episode.