#302 - 3 Hacks to Stop Your AI Content Sounding Like AI

Okay, funny story. So a little while ago, I got publicly accused of being a robot. Let me set the scene. I'm sitting in an airport lounge about to jump on a flight, and I realize there's a gap in our posting schedule, a pretty crucial one. And there's this one post I really want to get out into the world, and I want it out today. So I do a very rare thing. I actually sit down and I write it myself on my phone, thumbs going, boarding call happening in the background, fire it off. Done. I get off the plane a few hours later and there are a few comments waiting for me. Now, we hardly ever get comments, so a couple of DMs and comments was a big day for me. But every single one of them said some version of, haha, I can tell this is AI, or AI slop, or Kate, did a robot write this? And I just stood there in the arrivals hall laughing because here's the thing, that was the one post I actually wrote myself. Every other post that week, that month, the ones that nobody blinked at, they were the ones my AI wrote. So either I've read so much AI content, I've started writing like it, or my AI is now so well trained on my voice that it sounds more like me than I do. Either way, I'm taking it as a compliment. And today in this episode, I'm going to share with you 3 of the kind of quirky hacks I'm using with my own AI. So my content sounds more human than when a human wrote it. Build it, automate it, scale it. This almost feels too easy. Time to build a grown-ass business. The Doing It Online Podcast. Here's your host, Kate McKibbin. Hey, hey there. I'm Kate McKibbin from hellofunnels.co and welcome to the Doing It Online Podcast. Now, here's why today's episode matters, and it's the bit hiding underneath that whole airport story. Doesn't matter how good your content actually is. The second someone gets that little AI ick reading it, their brain just checks out automatically. They stop trusting it. They start discrediting the whole thing, even if what you're saying is genuinely brilliant, even if it is based on your actual teachings, your actual expertise.. And if you can feel that when you read other people's stuff, your audience feels it when they are reading yours. So a lot of people have just given up trying to make their AI sound like them. And I reckon you're probably in one of two camps. Camp 1 is, eh, it sounds a bit like AI, but who really cares? Camp 2 is, I cannot get this thing to stop sounding like AI, so forget it. I'm just going to go back to writing it myself. And I get both, but I want you to know there's a third option because neither of those is the one that I picked. Because if you're feeding your AI a one-line prompt and hoping for magic, yeah, it's going to sound like everyone else. But that's not the AI's problem. That's the initial setup's problem. The people whose AI content actually sounds like them, they didn't just find a better prompt. They didn't use a better AI. They built a proper setup behind it. So let me walk you through the 3 pieces of mine. AI content hack number 1 is a skill I built and I called it Eel, That's AI. SideQuest, if that word skill is new to you, basically what that means is it is a saved set of instructions that your AI runs on command instead of you re-explaining yourself every single time. So I built this one and I trained it first on all of the obvious stuff about AI content that I just can't stand. The em dashes, the endless bullet points, the way it structures every single thing pretty much the same. And the words. These are these words that just scream AI once you start noticing them. Things like silently, quietly, delve, meticulous. There's a whole list, and honestly, more of them keep turning up. Now, here's the bit that actually makes this work. It's not something I run when I remember to. It is actually built into our process as one of the very last things that fires during a content creation session. So I get everything written, I'm happy with it, and then the eel— that's AI— skill sweeps through right at the end, pulls out and corrects every single one of those tells. And the other thing that makes this powerful, and this is how you uplevel pretty much everything when it comes to AI, is it is a learning skill. It's got a loop built in. Once a month it goes off and it educates itself on the newest batch of Clear AI tells. It adds them to its naughty list, and from then on it's always running on the latest version. You can run that more often if you want, although, although honestly, once a month is plenty, but that's up to you. And yes, before you ask, this is one of the skills we hand you inside Claude like a CEO. If you've seen me talking about that, just copy and paste. You don't have to build it from scratch. Hack number 2, and this is a big one. This is how you get it sounding exactly like you. Because when I trained mine on my voice, I didn't just train it on how I write. I trained it on how I speak. I uploaded a stack of our client calls. And as part of building the skill, it went away and analyzed how I actually talk. And it came back and set its own rules off the back of that. And that one's a learning skill too. But the way it learns is the important part. So stay with me. At the end of any big content session, the final job is it goes back and it looks at the feedback I gave it, but not the little one-off, like change this word here, move this comma there. It looks for feedback with intent behind it, the why behind a change, not the specific fix. Because if you train it on every little one-off correction, you're going to end up with a giant pile of disjointed rules that is actually going to make it really hard to sound good, let alone human. And here's my favorite part of this whole thing: when it was learning my voice, I got it to actually hone in on the stuff I do that is actually grammatically wrong. The imperfect stuff, the bits that are a little bit off but make it very human. The overcapitalization of things, overuse of ellipses, because I want to keep those, right? That's part of what makes me, me. Look, I used to be a copy editor, full disclosure. I know the rules. I just don't care to use them all. And those little broken rules are exactly the nuance of how I write and how I talk. So I told it to protect them, not to polish them. And the whole thing, this whole process, it runs, it's just a simple setup wizard that I created. And it walks you through analyzing, learning, getting it all bedded in properly at the start. And then that feedback loop just keeps tightening it and tightening it. And that setup wizard is literally what we walk you through inside Called Like a CEO Day 2. It does the analyzing, the learning with you. So you're not staring at a blank screen trying to describe your own voice or just cutting and pasting in 5 of your favorite emails and wondering why everything out the back of that sounds exactly the same. Hack number 3 is the one that saves me the most time because here's the thing that nobody's really talking about with AI. Even a beautifully trained AI prompt, skill, agent, whatever, it will drift the longer that you're in a single session. So if you're writing out a whole month of emails in one go or a big batch of socials, by the end it's soaked up so much from all those little tangents and side conversations that it slowly wanders off from how you actually want it to sound. So I trained mine on a set of trigger words to pull it back in line, little code words, which means instead of typing out that same feedback for the 50th time, or even, you know, copying and pasting a saved snippet, which is what I was doing. A best example of this is the word Jess. So Jess is what we've named our ideal client avatar. We've got a full avatar breakdown for her and my AI knows her inside out because that was part of one of our original trainings too. So when I paste the word Jess, it stops whatever it's doing and it takes on the persona of Jess. Then it goes back and reads whatever was just written as her and it comes back and it tells me her gut reaction, which bits she stumbled on, where it got a bit waffly, maybe she started to drift off or click away, which bits actually landed like straight between the eyes. And then it gives me a score out of 10. How likely is Jess going to be to do the next thing, to click through, to register, listen to an episode, buy whatever it is? And then it writes it all out with that in mind. It is such a fast way to tighten things up. I've got a few other trigger words. Beige is another one that I love, but just putting in that one word It triggers all of it. It's all in there, ready to go. Now here's the really honest bit. If you try to do these 3 things one at a time, haphazardly as you go, it'd probably make your content take longer, not less. And that's not how we want this to work, right? So the best thing to do is you wanna be able to sit down once, you wanna take the 20 minutes to set it up properly, and then you just have it running underneath everything that you make forever. That's the difference between having a few clever prompts and actually having a content department that is learning and getting smarter every time you use it. So that's what you want to get to. If you're ready to go from copying and pasting prompts or a few disjointed projects to feeling like you've genuinely installed a proper content system, come and grab Claude Like a CEO. It's the whole setup, including that setup wizard, so you're not building any of this from scratch., and right now is a massive $150 off, but only for a very limited time. The link is below in the show notes. Oh, and for anyone listening who's already got our cash money content or even our previous program, Course GP Team, you might be wondering how this fits. So basically, as far as Course GP Team goes, this is an upgrade. Course GP Team was over 2 years ago. So many things have changed. If you have that and you've been running it, this is how you now uplevel everything to what is the latest and greatest and the most efficient in AI now. Now, Cash Money Content program, this is the piece that sits on top of it. Cash Money Content is a system for coming up with ideas, the strategy, the mapping, and creating all of your organic and social content in one sitting, right? This is a layer that makes everything that comes out of those sessions sound even more like you and land even harder with your people. So it just takes Cash Money Content and it makes it even better. The first lesson unlocks tomorrow if you're listening to this the day it drops. Don't worry, it's not live. You can jump in at any time, but tomorrow is where everything starts. So that's the 3 hacks, guys. There's the eel, that's AI, the skill that sweeps your content right at the end, the voice training that keeps your human imperfections in, and the trigger words that pull it back in line when it drifts. None of it is about finding one single better prompt.. It's about setting up the right systems underneath. And if you ever get accused of being a robot on that one post you actually wrote yourself, honestly, do what I do. Take it as a compliment. It means your AI's working. All right, that's it for today, guys. I'll see you all next week. That's it for this episode. Now go build it, automate it, and scale it. I love the show, guys. You're awesome. For more strategies and resources, Visit hellofunnels.co.