The AI Brain and the Human Brain

At the start of 2026, there was a warning that most people ignored.

The biggest problem wouldn’t be AI itself…

It would be the explosion of “cool” AI tools from disconnected suppliers.

Now it’s starting to happen.

In this episode, Mark breaks down why the rise of AI app stores and plug-in tools isn’t progress — it’s the beginning of a structural problem inside estate agencies.

No central brain.
 No shared intelligence.
 No unified view of the customer.

Just fragmented systems making disconnected decisions.

This isn’t a tooling problem.

It’s an operating model problem.

Alongside this, Mark shares a front-line insight from inside Iceberg Digital — where six senior roles have been replaced not by intention, but by outcome — and what that reveals about the real role of humans in an AI-driven business.

This episode isn’t about AI tools.

It’s about where intelligence actually lives inside your business — and why getting that wrong now creates a position you won’t easily recover from later.

Transcript

Speaker 0:01

This is the AI Operating System Diaries. I'm a Mark Burgess, Founder of Iceberg Digital, and the author of Prepare AI's Impact on Business. For the past few years, we've been building AI Systems for Estate Agencies while transforming my own company into an AI First Business. This podcast documents what actually happens during that transition. The resistance, the mistakes, and the breakthroughs that change how businesses work. Okay, welcome to this episode of the AI Operating System Diaries. If there's a little bit of an echo today, I'm sitting in my sauna. I thought it'd be a good place to do this episode because I often come here for thinking time. And today's episode was kind of split into two different parts. The first part, um, I just wanted to talk to you about something that I spoke about at the start of 2026. Um, at the time it made sense to everybody, but I'm sure that everybody's starting to forget. Um, I said something that uh didn't land particularly well at the time. I said the biggest problem coming for a state agents wouldn't be the AI itself, it would be the explosion of cool-looking AI from random suppliers. And then today I saw something that completely confirmed this, and I wanted to just talk about it on this podcast because I'm sure, like, you know, I spend my whole life geeking out about AI running an AI company, but uh, you know, you guys that are tuning in are spending your days running different kinds of businesses, so uh I you might have missed the signal somewhere. Um, so today in the industry news I saw that RePit are opening up an AI app store, um, and on the surface, you know, that sounds brilliant, doesn't it? More choice, more innovation, more tools. But this is where I was talking about how operationally this is where the nightmare begins. So, what we believed was that more AI tools would mean more progress. Many people are still thinking that at the moment. Anything like they can get their hands on AI-wise, anything that's got the two letters in it that sounds like it would be productive, great, let's just get it on board. Why wouldn't we? And that if we could just plug in the best bits from lots of different providers, we'd somehow build a better business. We that's that logic kind of worked in the software era, but it doesn't work in the AI era. Here's what actually happens. Every single one of those tools comes with its own version of intelligence, its own instructions, its own idea of your business, its own rules, its own memory, and none of them talk to each other properly. Okay, this isn't a thing whereby you go, I don't know, the AI will just figure it all out, won't it? No, it's not how AI works. Not currently and not in the near future. So instead of building one intelligent system, you build ten disconnected ones, and what ends up breaking is not the tools, what breaks is the operating model because there's no central brain. So for any of you that were there when I done my talk at the start of 2026, I spoke about the centralized brain that's being launched inside the uh iceberg AI operating system. There's no central intelligence in these systems, so it sounds great. Yeah, you can plug stuff in, it would do like some better stuff than it used to do before. And if you're a user of those softwares, that would seem like a no-brainer. Why would you stick with the old ship version? But there's still no single source of truth. Now you're just making the mess even bigger. Before you know it, it's going to be like the London tube map going on. No unified understanding of your customers, no consistency in how your business communicates. You end up in a situation where one AI is writing emails one way, another is handling leads a different way. Then you've got another one that's doing marketing based on a completely separate logic, and another one responding to vendors with no awareness of what's already happened. And none of them actually know your business, they just know their narrow task. This is the problem. We've already seen it with distributed systems, multiple tools, multiple data sources, no centralized intelligence. We already know where that leads to. It creates noise, not leverage. So AI makes that problem worse. It makes it worse, not better, because now it's just disconnected data. Uh it's it's like disconnected decision making. It's like some guy just in a field on his own that knows fuck all about your business is now suddenly qualifying your leads for you because you know it's made your life easier. And there's someone else in another field who also knows nothing about your business that's supposed to do marketing to them. And the two guys in the two different fields don't talk to each other. There's also another woman in another field who's like creating presentations for these people, it's just it's it's going to be an absolute shit show. So AI just makes all of that problem worse. The part that most people are going to miss, every one of those AI suppliers needs to make a profit from you individually. They're all individual companies, which means they're not building for your operating model. Okay, they're building for their features, their case, their revenue. So you end up paying multiple companies over the odds because they can't offset the cost against something else. This is their only business, they need to make some money out of it. So each one of those companies, you'll pay them over the odds, and each one of them will partially understand your business badly. Then something else happens. Eventually, you'll forget what you've set up because there's no centralized place where your rules live, no single system that understands. So as those tools try to become and stay relevant, they'll expand their offering. Each one of them will start crossing over with the other tools. But you forgot because you set up the other tool six months ago. You can't remember that you actually made it nurture someone for six months or do cold outreach to somebody if they didn't meet this criteria. Now you've got another tool that you've put in place that sounds brilliant that clashes with it, but no one knows it clashes with it because it's a whole nother tool in a whole nother place. You just have fragments of logic scattered everywhere, and it all starts to become irreversible because once your business is operating across 10 different AI tools, unwinding that is not going to be simple. You're not just removing software, you're trying to untangle behavior. Now, here's the bit I want people to really pay attention to. Over the next 6 to 12 months, there's going to be a lot, I mean a lot, of very cool-looking AI tools, brilliant demos, great marketing, very convincing use cases. They will all look like progress. But if you build your business that way, in 18 to 24 months, you will be in a completely different position. You won't have an AI-powered business, you'll have a fragmented one. A business full of automation but no intelligence. While you're trying to manage that complexity, something else will be happening quietly in the background. There will be agencies with one system, one brain, one set of rules, one continuous view of the customer. Where everything compounds, where every interaction makes the system smarter. The whole system, not one system, where decisions improve over time. That's the divide that's forming that most people just will not be able to see until it's staring them in the face. It's not AI versus no AI, it's system-led versus tool-led. And it's the same pattern we've seen in every industry shift. The companies that will win don't bolt new technology onto old thinking. They change the way the business operates. It doesn't matter what your CRM tries to give you to make you feel like you're embracing AI. It's just bolting new technology onto old thinking. You're gonna have to change the way your business operates. Scar tissue from what we're seeing right now is that the more AI tools you add, the less intelligent your business actually becomes. Okay, what this means next is that the question shouldn't be what AI tools should we use, it's where does our intelligence actually live? Where does it live? You've got all these different tools, 20 different AI agents from 20 different companies. Even if they're all supposedly linking back to your system like a repit, it's not gonna work. That is not going to work. That is 20 different people in 20 different fields all trying to do a job when they don't know what the other one does. By the time it's obvious, it will be too late to fix. That's the bit that most people are about to learn the hard way. So it's not about choosing the best tools, it's about choosing the way your business thinks. The other thing I wanted to talk about in this episode was what we're what I'm experiencing with AI myself in the company. Um in the last eight months, we have uh parted ways with uh six senior members of staff who if we go back eight months, all six of them I would have said were pretty critical in the operation. Like if they were to up and leave it would be a ball ache, uh to say the least. Uh we have not had to replace any of those six people. The AI is doing their jobs for them and it's doing it better, which is kind of crazy. Um now we didn't, it's not like we just said, Oh, you know what, we're gonna fire all these p people and replace them with AI. That's not what happened. Just circumstances led to them either leaving the company by mutual consent or you know, us deciding it wasn't working anymore, or them deciding that, like whatever, it doesn't matter, right? The point is is that we have not replaced any of these people, and we are in a like a considerably better position without them than we were with them in terms of the work that's being done and how it's being done. Now, the caveat that I want to put on this that I'm learning is that AI doesn't finish the work. So what people imagine in their head is that let's just say in SAWs, you imagine that like AI is gonna surface the opportunities, AI is gonna write the outreach, AI is gonna never give up, and it's gonna continually speak to these people via email, WhatsApp, phone, whatever you whatever you choose, until eventually it convinces them that like they should be going ahead, close the deal, send the contract, seal the deal, get the onboarding done, like whatever it is. If it's marketing, it's gonna come up with the marketing content, it's gonna come up with the images, it's gonna create the content on social, email, send it out, monitor it, see what's working, see what isn't, make new content, send that to the right people, and so on and so forth. Doesn't matter what the job is, that's how people imagine that they want the AI to work. It does not work like that. It doesn't matter whether you're using it just to give you some intelligence on contacts. So inside our lifecycle system, our agents now have contact intelligence. Then go to a contact, and immediately the AI will give them the lowdown on everything about this person when they came into the system, what's going on with them at the moment, what the risks are, what the opportunities are, what the next best thing that should happen with this contact is bang, just like that. Amazing. Uh at our end, you know, we've got AI that's uh looking through the database, finding old contacts that we've been engaging recently, looking at all the conversations that we're had in the past, uh creating lists of who should be contacted, creating the outreach, uh bang, just like that. So, all of these different things, but it doesn't actually finish the work. Someone, a human, has to step in and do the final bit, the human bit, and quite rightly so. We wouldn't want it any other way, would we? Do you really want it so that like there is just no role for the human? I mean, that doesn't sound great to me. What we're trying to do here with this with AI is we're trying to get ourselves back in time to that place that we used to be. If you think about the world pre-smartphones, pre-internet, pre-WhatsApp, even pre-email, you did work from nine till five o'clock, and then you went home. Right? And the work you did between nine and five o'clock was you and other humans interacting. So if I think about when I was an estate agent back in the mid-90s, the only way I had of doing any work was if a person walked in my office and I had a face-to-face conversation with them, or if I phoned them and had a conversation with them over the phone, or if I went to their house or on a viewing and had a conversation with them face to face. That that was it. That was the only way. There was no other way. I mean, I guess you could send a letter if you really wanted to, if you was really struggling, but like I can't really think many people it did that even back then. It's probably more popular now than it was back then. Fuck knows why. Um so what's happened since humans have all become their own PA. Everybody has to answer WhatsApps constantly, whether it's from your team or from the public or from the you know the customers. Have to answer emails constantly, being bombarded with fucking spam that you've got to filter out and miss, you miss the important messages, even when you're looking yourself. You've got to have Zoom calls with people. Some of those people can't do it between nine and five. Oh, it's okay, we'll do it at six o'clock in the evening. And then suddenly you're gonna sit down at seven o'clock, finally, you're gonna watch, I don't know, a bit of Netflix, do whatever, chill out. But you know what? I'll just while I'm doing that, I'll just check my social feed as well to make sure nothing's going on. And it goes on and on and on and on. That's the bit that AI can sort out for you, and you you'll probably have to overcome the addiction yourself. I I have to stop and think, what am I doing now? You know, because it would be quite easy for me to feel a bit worthless. But it's quite the opposite, really, when you examine it, you can start to do really meaningful stuff. You can sit and think. That feels like a little bit of a cop-out in today's world, but it's what we're not able to do, right? No one can no one really can say, tomorrow I'm gonna go to work and I'm just gonna think for four hours. That's not happening. That is not happening unless somehow you've got yourself into the routine of shutting yourself in a soundproof box with no internet. That ain't happening. But with this, with AI, I'm seeing it can happen. You have to rethink the roles and the jobs. You start it as the owner of the business, and then you're able to teach it to other people. Without you starting it, what happens is you start to get frustrated with your team that they're not using AI, and no one even knows what the fuck that means. But like you go, I just fucking start using AI. Fucking hell, why are you all taking so long to do this stuff? Just use AI, use AI. They don't know what to do. You're the leader, and you're not even doing it. So that's that's one of the big things that I've learned over the last couple of years. Is like wow, I have to think about what do I do every day? Because I still do random fucking shit just because it's that's what I do, and I can't, I haven't thought of anything else yet. I don't need to. The AI can sort out my inbox, I don't even need to look in there. The AI can sort out the market, and the AI can sort out like scheduling the socials and posting it and cropping them videos and putting the all of that shit. My presentations, like, you know, if I had to do a presentation somewhere previously, I've got to think up the whole presentation, right? I've then got to turn the presentation into some sort of script. I've got to take the script, break it up, put it into PowerPoint, create the images, create the slides, rehearse the script. I've got to do all of this stuff. Now I don't have to do any of that. But what I do have to do is I do have to convince myself not to have to do any of that. And that's harder than it sounds, because it's like a default. It's like what I do to make myself valuable. What you do to make yourself valuable is about to change, and only you can change it, and only you can change the whole company once you've figured out how that works for yourself, so that you can show the example of what that actually means. Okay, that's 20 minutes. The song is fucking killing me. So I'm gonna leave it there. Hope you're finding these episodes useful. I guess my last point on this uh going back to the original stuff that I was talking about, is that some agencies are gonna build a system that gets smarter every day. Uh, not just their software, their entire team. And some are gonna build collections of tools that get harder to manage every day. Again, not just their tools, but their team, they're gonna keep working the way they work. But isn't it brilliant? Because AI will allow us to work the way we've always worked. What a fucking load of shit that will be. So those two things might look the same from the outside. Both companies are using AI, but they won't perform the same. And by the time that gap is obvious, you won't be able to close it.