Transcription
I consider this the most important video I ever made. I launched this channel in 2017 with the first videos about productivity and my experience working in corporate and climbing the ladder there. I shared over time all the insights and struggles that we have thriving in a business world.
In 2022, I met my co-founder, Paco Cano. He built four businesses in parallel in the past 20 years, and he came to the same conclusions when it came to productivity in a business. So, we talk here about decades of productivity advice that helped us to be highly efficient inside business, and we have seen this over and over again.
Today, it's AI. In the past, it was automations, and before that, it was workflow optimizations. And now you see people saying AI is just hype. It's not working. We have even studies showing that 95% of the companies who tried to implement AI in their businesses failed. And to me, this is much more worrying because it has nothing to do with AI but with the fundamental setups in these businesses when it comes to productivity. And to me, failing of implementation of AI is just a reflection of the underlying business processes that are just unclear.
And hear me out. In this video, I will get step by step to why this is, and we have a lot of ground to cover in here. If you're one of the people saying AI is just hype, this might change your mind afterwards. And if you're already using AI to make your personal work more efficient, then maybe you want to send this video to anybody who just doesn't get the point about the potential of AI as a professional.
I will break this video down into three parts. First, I will dive deeper into what I just claimed and explain why AI is failing to be implemented in the workforce and businesses. Then, I will share with you in part two things that I'm leveraging AI on a daily basis and the return of investment related to money and time, which to me is obviously the proof that it's working. And in third, I will show you the basic setup of any AI agent to really make it work for you, which is directly connected to setting up any other team communication system too.
And right from the get-go, I want to point out this is not Claude specific. This applies to any AI that you might want to use, no matter if it is Gemini or any local models. It doesn't matter because we have to think in a tool-agnostic way about these problems. And once you get the understanding about this, you can use any tools you like. And by the way, that's something we taught since 2018. It doesn't matter what tool you use. If you don't understand why you're using the tool in the first place, it will never help you or your business to improve anything. And you will be always on the hunt for the next shiny object that another person recommended to you. And that's what we are here for. And the ICORE methodology to make you aware of the things that surround you at work and make you aware of this. So it makes it much easier to find the right tools to fix it. AI really helped me to remove the friction of the things that I just don't want to do and really focus on the things that I love to do that I'm an expert in and just offload the work about work.
Let's dive into this. Let's start with part one. Why AI is failing for individuals and businesses. I want to give you a very practical example. And in order to do this, I have here a Miro whiteboard open that I created in August 2020. This is where I was still working big corporate, and this is where I came up with the first ideas about ICORE and how I consider what continuous improvement is because here is a very practical thing that pushed me towards this conclusion. And you see here, it was hand-drawn on my iPad. I put it on this on this board here, and that's something that still applies to date.
I worked in multiple projects in parallel, global projects, local projects, a lot of combination, millions of budgets. So it's not something where you have a small solopreneur startup and really complex work with over 60 people involved. And what I recognized back then that whenever we completed a project, we had a lessons learned session, and we sat together talking one, two, three hours, concluding things about the project. What went good? What didn't go good? And then we kicked off the next project exactly the same as before. There was no change. They were doing the same mistakes, and then we ended up again in the lessons learned session. People started complaining again, and then I recognized there are people constantly complaining around me, but nothing is changing. And this is where I drew this where I considered each loop a project. So we have the project start where we gather the information, the input stage, we organize the information around the project that we had. We produced the output during the project. In this case, by the way, it was building packaging lines for medicine. Very long projects, two to three years. And then the refine part was always missing. Well, it is the lessons learned, but then it was cut off. And I thought about how can you bring this over into the next project. So we have a continuous loop going on where each project becomes better than the previous one. And that's what I called continuous improvement.
And anyone working in corporate knows that continuous improvement is a buzzword, and you have to optimize your work and all these things. But if you're constantly just a firefighter, you have no time to improve things, and you just stick to buzzwords, and actually nothing changes. And this was to me the first conclusion that continuous improvement needs to be a positive feedback loop where everything improves the next time we start again, and this will remove frustration.
And then I thought, why is this happening? And I was thrown into a project that I took over from another person who was leaving after 15 years. And they told me, "Yeah, you have to take over his work." And they showed me a slide deck where they showed his name there. And they said, "You will do this, and you will do this instead of him. So you replace your name with this." And I said, "I think this is the wrong approach." By writing down a name on the slide who is responsible for this, you build up a bottleneck based on a human, on an individual. And if this guy leaves the company, as it was the case in this moment, everything fails due to one single person. And I was shocked in big corporate that something like this can happen. And obviously, over the past decade, I learned work coaching many businesses worldwide that this is industry-independent, company size-independent, these issues happen over and over again.
And what I did back then, I said, "I'm not doing any of what you're telling me." Instead, I sat down and I created this map. As you can see, this map was also created in 2020. So, this is where you see there was directly the frustration. This was the birth of ICORE as we know it nowadays. And this is a real practical thing. I was taking over a team back then with five people. One guy coming back from burnout after one year, another guy was away with a heart attack. And this was the kind of environment I was working on, and my responsibility for people in there. So the pressure was on. And then on top, as a team leader, I should take over project work from a guy who just left after 15 years. So this map didn't exist in this company yet. And I sat down, I said, "I'm not doing all these things because I have no clue what this guy was doing. I have to wrap my head around what the scope of this project is." And I mapped this out, the first things that I understood so far. I laid out on a map. Then I went to the different departments, engineering, quality assurance, the project manager, and so on. And we sat together, and we stick together the different pieces. So you see the big boxes, and these are different departments working. And then I laid this out, and we met again, and then I proposed, "Look, there are parts in this project that can perfectly run in parallel, which was previously just sequential. There's a lot of things where if the departments directly would do the work, the whole work would be much more efficient. There's much faster handover than actually having one single bottleneck of one person going away." And it's not only a bottleneck, the thing is just also not scalable. And this guy was taking on more and more projects, and he was always constantly bottleneck. And we realized this single person was doing all this and all this for himself. He learned this in the past 15 years because he had to wait for people. So he was ambitious and he learned so many different things that helped him to move fast, which in the end ended up that he is the most important person there. And that's what changed the moment I created this map because we ended up with, first of all, I didn't do anything in this project. In the end, my team took over this single arrow here. So I first broke it down to this box here where I said, "Okay, I see my team doing this." And then even inside our department, we just realized that there are different groups inside the department who are more efficient doing individual parts of this. So we ended up just doing this individual arrow there. So you can see how one person doing all this, and how slow this is, switching to functions and defining what function should do what in this project. And therefore, it was then able to hire new people based this function. We had identified the bottlenecks, and we ended up increasing project performance by 40% with the same headcount, in fact, even less people involved than we had before the handover. We became so efficient doing this one arrow. And if we look into this arrow, it is a whole bunch of other processes behind this again. So this is zoomed out, and as I said, that's a very complex project we are talking here about. And if it is working for this, it works for any smaller projects or smaller things that you are doing the same way. And that's where I was born. That's the foundation of getting an understanding of business processes that's going on here because here we are just talking about business process optimization, and not the optimization of the expert doing the work. So this became highly efficient.
And you say, "It's not worth for a one-off project to create a whole map." But the point is, in any project in your business, you might realize that there are recurring tasks, workflows that you do for any project over and over again. Even that you have variations from project to project or the outcome is different. But you will be able to identify parts of this project that you do over and over again. The team does over and over again. And by identifying this, you will optimize the whole teamwork and the whole business work. And this is the crazy part here because back then there was no AI. I wasn't allowed in the pharma industry to use any automations or anything like this. This was just human work optimization. And this allowed me then to provide the right numbers to the leadership, to the management. People got fired left and right. And I saw it is based on wrong conclusions, based on the wrong KPIs. And I back then, I kept saying KPIs like watermelons: on the outside green, and the inside red. We don't do us any favor if we force the things to make the numbers look green because then the wrong decisions are made in the next level, and then we complain again about the management. That's why I always told my team, "Never blame the management. It's on us to provide the right numbers to them so they can make the right decisions."
So on the side of this, this is now project work that I was involved in. But at the same time, my team was doing operations in this business. They've been constantly behind that the management was complaining how slow they are. At the same time, they started firing people in operations because they thought the budget is too high for operations and all these things. But nobody actually ever tracked on how many projects my team was working and involved in. And again, this was historically grown. There was another department reaching out to my people, which I wasn't leading back then. They gave them more and more tasks left and right that was never agreed on. They want to be great workers. So obviously they take on, they make over hours, and this is how you go into a spiral that you cannot get out of because the whole business assumes just this is the normal. Also, the whole team is already on a stretch doing it, and then the moment people stop working, they get out, they leave the company, or they become too expensive, or whatever.
And what I did when I went into this team, I started collecting the projects, and we ended up collecting 150 projects. These five people were working in parallel on a global scale. So this means also whenever there was evening for us, it was morning on a different part in the world. You see, I'm I'm sharing this with you because I want to see anybody who has anything more complex or, you know, when people say we are too restricted with the tools that we can use, forget about tools. All this, I was not allowed to use any tools. What I was able to whitelist back then was Asana, a project management tool where I simply collected the projects. If I wouldn't have been allowed, I would have used Smartsheet. It was back then. And if this is not allowed, man, then I use an Excel sheet just to write this down. Because what this allows me to get the overview, I was then able to block projects, which led me then to talk to the management and say, "Sorry, but you have to tell me what we need to prioritize on this." And everybody now easily understands if you have five people working on 150 projects in a firefighter mode, this will keep failing. And this is where then we identified business-critical projects that really bring return of investment versus optimization projects, which was the vast majority where a student came into the company, they had a great idea, and they offloaded it to our team because they left the company again, and all these things were grown over 15 years. And I came in and disrupted the thing be just based on pure facts and numbers. And if you create a map like this, you would be shocked about what a broad understanding you suddenly get across the board, no matter if it is management or your own team members, what not. And now we can do the same for our life. So we map out also our life like this. That's also possible. And you get a much better understanding all the different things and moving parts in your life that are going on. And that's where I realized it applies to all the things. We have to zoom out and get a holistic understanding of what is going on before we zoom back in and then go optimize things.
And as I worked in one of the richest companies in the world, at least for this industry, I went through all these things like Six Sigma, Agile thinking, a lot of trainings. And it ended up always with just buzzwords. These trainings were just given because it is a company goal. But nobody really understood why they are these things that an external trainer told them to do. And we never looked actually into the actual business processes that give real tangible changes if you optimize them. And this is one of the things. And before I took over this team there, I already increased team performance by 60% as an employee inside a team where all I did is creating a single source of truth. And that's what I want to get next into another reason why AI is failing. Yes, we are still here for AI because bear with me, you will get to this point. You see already here because once you have laid this out, you realize there are individual parts and here it was not me doing the work. I delegated it to my team who are experts who are best in doing this. Instead of me struggling my way through to make this. So delegation is much closer to automation. And AI in my opinion sits between delegation and automation. Why is this? Let's quickly think about this for a moment.
If we think about delegation, this is delegating work to another person inside a team or to a freelancer, you name it. In the end, I'm handing over work to a different person. And this is again where people struggle. Solopreneurs, it's the best example. If you see creators on YouTube and they become famous and grow very fast with the idea they have, they think they need to hire now people, and now they have five people around them, and they want to delegate work to them. But then they realize they have no idea what they should actually hand over to these people. So they constantly end up in meetings talking about a lot of things. And I will show you in a moment what this looks like. But this is the problem with delegation. So this is handover to a person. Automation was in the past always something that is very repetitive that has no variations in there. So you were able to create automations with a vast amount of tools. It was developers doing it. Then you had the no-code solutions like Make.com and Zapier and so on. And that's something I used a lot when I launched the Paperless Movement business and now my ICORE. I had no need to hire people because I automated so much in my business because I was able to identify the repetitive work in my business. And AI, people think now it is equal to automation, and it is not. It's also not better than automation. In fact, it sits right in between there. Why is this? Because if you give AI the same input, you will constantly get a different output. You can ask AI the same over and over again. You can give AI guardrails. You can give AI instructions and all this. And yet you will see there's always a variation in output. This is not happening with automation because this is based on pure code and mathematics. While here, we cannot predict what's really coming out. So we cannot ensure outcome will be always the same. And this is one of the reasons why people say AI is not working. It cannot help me. However, it is actually an advantage that there is a variation here because automation was always prone to fail because the moment a slight thing changed in your process, the whole automation failed, or you had to update the automation, or you had to have several versions of the same automation for all the different use cases. And that was a lot of time investment and money investment if you had consultants doing this that many companies realized it's not worth doing for business process optimization for the normal workforce. AI is now compensating this. It has a certain degree of automation which is in fact now delegation, and it can autonomously work on the task you give it. And then you need to be happy with 80% of the output. And that's the point where people say, "Well, if it is just 80%, I'm not happy with this. And it makes no sense to hand over work to AI if I don't get 100%."
But now is the moment where we have to pause for a moment and now think again about what the alternative to this is. The alternative is to hire a person who at best can perform three to four hours focused per day. The rest of it is you need to give this person shallow work and things like that because the energy levels drop. There's a variation each day of this person. You know, had a hangover from the day before, maybe, or, you know, some issues in a private life, or suddenly needs drops out because there's an urgency or something like this. So the consistency of 80% output of a human being is much less than having AI. So even if people say, "But you rely too much on Claude or on Gemini. What if these servers go down?" Yeah. Then, in my case, a whole team of AI agents will go on holiday for the time until they are back on. But here, I was relying on one person, and if this person dropped out, then other people kept working or compensating for this. However, this is only possible at scale. If I'm a solopreneur and I hire a personal assistant who works along with me all the time, and I did this too when I started the business, and the output again, the type of work that you hand over, I would say it was at 60% quality that I received back, and I had to always come up with the remaining 40% doing it myself. We had so many CEOs and successful people who say, "I'm overwhelmed. I'm hitting a ceiling constantly. I'm hiring now a personal assistant helping me to manage my calendar, helping me to make my task list, and so on. Give me prioritization." This is where we always want because these people hire another person, and then they realize they just added another layer of communication on top of the things that they had in their mind previously. Things that they are not cautious that they are doing on a daily basis, that they don't need to talk to somebody to organize the calendar and things like this. Now, they had to communicate with this person. So, there was a constant back and forth on, "Okay, is it worth making the schedule this way?" "Yeah." "Okay." And then this person drops out. And the 4-hour work week, maybe you read this book, too. Sounds exciting that you had some a bunch of people on Fiverr and some people in the Philippines and cheap labor and so on, and you can just sit on the beach and enjoy life. This is a lie if you don't know how to optimize and also compensate for any dropouts of these people. Otherwise, your business just stands still.
And if you consider this quality of work compared to AI, it's already improvement. So because if I consider one person costs me, let's say this person works full-time, and it's just doing, you know, administrative work, and let's say it's really cheap, okay, and we, I'm not even talking about all the additional things that we have to pay on top. Let's say it's $2,000 per month for 60% output for a person who works five days a week and only half the day really focused where there's really good output, and I pay $2,000 for this. And I consider here getting 80%. And let's not be unfair because I see people already in the comments, "How can you say that?" Let's say both give me 80% output, right? And I consider here that I pay for Claude $200 per per month, and I just have one specific use case that this AI takes away from me so I can focus on my expert work. These numbers no longer match up. There's no excuse anymore.
And this is where people start warning of the huge job loss that is happening, and that's where people are not aware. I see people out there, even productivity gurus, who say, "Don't worry, we have time. The AI is just hype," and so on. These are people who have no idea, first of all, what AI is already capable of autonomously inside your business with real output that saves money compared to having people. And in many cases, they also have no idea what it really means to work in a real business environment on high stakes with competition around you, all these things, pure stress all the time. I was working 10 to 16 hours each day in this company in corporate because I wanted to get up the ladder and all these things, and you just burn out. You cannot sustain these things. And if you never experienced this and back-to-back meetings, how and how this is, then you have no idea what it means to implement systems like this in a real-world scenario. So to me, automation was never the danger for these people, but AI is.
And now let's talk quickly about the type of people who are in the most danger. And those are the people doing low-effort work, but it is still time-consuming. So, for example, data entry into Excel sheets, right? Or you copy-paste word documents or whatever. All these things we delegated previously to people. And who does the delegation? Experts, because experts get a lot more than this. So, let's say here delegation, and then we can have here an expert. And if you think I'm making these numbers up, no, they are real numbers, at least in the corporate I was working in. And this is the expert work, okay? So when I took over this team that I talked about in the beginning, there were experts sitting in meetings wasting time instead of focusing on their area of expertise that they get paid for. So me as a team leader, I looked for what can this expert delegate to a person who is much cheaper so he can actually do more of what he is the expert in. And that's how we need to think. And you see, I can now just ignore this AI and automation. I can already optimize my work inside my company by just realizing this. And I show you now, uh, going back to this very first ICORE map here. If you go down here, here is where I thought about the principle of single source of truth. And here is the exact example that not only me was suffering in the beginning, but then also my team later on. So my team members, they went to a meeting, to a project meeting, they sat there sometimes three hours in a meeting, kickoff meetings, and things like this, and were stuck in a meeting where they could have done their work they get actually paid for. Afterwards, we had people in this business sitting in a meeting just taking meeting minutes, which weren't even experts. This is what we just talked about delegation. There was the wrong thing actually delegated to another person who is not an expert, trying to take summaries and meeting minutes out of this meeting, who sends then an email to the experts what was concluded, which was half of the thing was wrong because obviously the person taking notes never had a clue. So these experts had actually their paper notebooks where they wrote down those things. So what happened here? Obviously, they archived this email immediately because they just had this meeting. So, why should I look it up? Then one week later, thousands of emails later, during a week, and I'm not joking, this is real life corporate. Go into the comments and share with me if this is you. And you know exactly what I'm talking about. This person already forgot what was said here. Nobody looks up this email because if they start searching, they are overwhelmed anyway. So, they reach out in a direct message or calling this other person again who just was in the meeting asking for the questions. There's a back and forth and arguing until two weeks later, you have exactly the same meeting again that you already concluded two weeks ago. And this is not something that I made up here. As I showed you, this map is from 2020, and I was just mapping out my experience working in corp right there. And this is something we see over and over, no matter what business people we are working with, and this is constantly happening. And to break this circle, that's where we define with ICORE productivity system end to end because if you understand there needs to be a single source of truth, that's something I sketched out here, right? Where I thought, "Okay, what if we would have a single source of truth and you just have a meeting for clarification, for making clear what's going on, and then you can just update the single source of truth inside the meeting. Everybody has seen how things got updated. There's no need for meeting minutes afterwards. There's no need for follow-up meetings because it was agreed on and everybody saw where the things got updated." And then I came up back then, something a term I'm not using anymore, personal single source of truth, where I considered that I have a personal knowledge management system where I update my own insights. Okay, personal knowledge management is doing exactly this, where you write down your own insights, your conclusions. Like, "I hate this person." I was sitting in a meeting, I'm not writing this into our business single source of truth. And what people also misunderstand is talking about single source of truth. People always think it's one tool that I mean by single source of truth, and that's not right. It is the individual task inside a tool. This is your single source of truth where you add information about this task. If you don't want to put this information right in the task description, which makes sense, you might have information scattered on SharePoint, on OneDrive, on Google Drive, and in big corporates, there's a mix of a lot of things. You still can define a single source of truth where you keep track of the task, and you just link out to the different external resources. And the moment a person needs to work on a task or you sit in a meeting to discuss the task, you are just one click away to open the relevant information. And again, that's just a small insight again, and a huge boost in efficiency if you do this for yourself, but also for your team or in the whole business where you boost efficiency. And again, there is no AI, there is no automation involved. It's just streamlining your business processes that are surrounding you. I ended up in this company as a business analyst. On top of this, I'm not going into deeper detail. I left the company in 2021 to teach this worldwide in different businesses.
And now we come in the era of AI, and we see the same problem over and over again. People wanted to automate this since the get-go. How can we get meeting minutes immediately? Then we had the AI meeting notes where now each meeting is tracked, and you collect all the meetings, but it makes no sense because in the beginning of a meeting, there's confusion. In the end of the meeting, there's conclusion. And you don't want to go through the whole process again. And if you give the whole bunch to AI, it might get confused with what was discussed, arguing conclusions. And if you have a three-hour meeting, how do you expect AI or this poor guy, you know, making meeting minutes here, who has no clue what's going on, how do you expect that they will be able to create the right summary, extract the right information from this? So, it's great that we are able to track all this, but now we need to use AI, and that's the advantage now to extract the information that is relevant for you. And this is not working without context. And that's the next thing we want to look into.
Now, I talked a lot about general productivity. This has nothing to do with automations or AI on top and so on. But you see already how human work can be optimized to get a most efficient workflow going. This is business processes, meetings, but this applies also to production lines and all the things. It's the same thing. Now, we talked about delegation. We talked about if you don't see the worth delegating this work to a person with $2,000 per month because you say, "This is so much money, $2,000 per month." And you don't realize, but if the expert needs to do the same work, it is $20,000 per month. And you make the calculation, you know, obviously, it's just partial. So you will see if this guy works half the month on just work about work, it's $10,000 per month. Now, it's worth offloading for $2,000 per month. And the same applies for AI. People always expect they hire a personal assistant or anybody helping in the business, and this guy needs to do everything now correctly, and it's impossible because we have experts, and we have people supporting these experts. And that's how you need to see AI. And that's why we believe instead of forcing in your business AI to the people and say, "Okay, everybody's using AI now. We need to use AI too. Here, you get a subscription to Claude or whatever. Start using it." They will fail. Because most of people in business don't know what they are doing the whole day. They just doing what they are told to do. And then they complain about if you keep switching and many things. So if there's a lack of clarity what people need to do in your business already, and there are people complaining, this is not only prone for toxicity and downward spiral, it is also failing with any type of automation or AI.
What we think instead is education for these people first of all, how they build their business processes and understand what they're working on daily basis. And then the team leader should be clear first of all what the team is capable of, that they are not working 100% of the day, but only 50% because energy levels. And if there's a lack of clarity of business goals and the projects that result out of this and the tasks that result out of these projects, this is again part of ICORE, then you will never get to the point where AI will work. And this was always the case for any type of productivity tools beyond even automations. It was the same for project management tools, and it is still the same for personal knowledge management tools, and so on. And now we are here with AI, and I'm explaining you the same thing that I explained already years ago. And by the way, my co-founder Paco Cano, who built four businesses in two decades managing over 70 people, and he doesn't have a personal assistant because he is exactly working according to ICORE as I do. We just know this is working because we are doing it every day, and we are maxing out AI, and previously we maxed out automations and so on.
And now let me briefly talk about AI in general. People in businesses and generally the people worldwide, they still treat AI like a chatbot. There's a back and forth, they get some insights, but this is it. If they, for example, use AI to write their email responses, they are not happy because it's not their tone of voice, which makes total sense. Why is it the lack of context? I mentioned it previously. Context is the other missing part. And that's the point. If you don't give AI the right context, it cannot help you at all. So context matters. And this again matters for your people as much it matters for AI. And what do I mean by context? Context is the relevant information that is specific for your business or your personal life. If you just use a chatbot, AI will just use the general information it has from the web and it was trained on. And therefore, it is such a vast amount of information that you get generic answers. Then the more you use these chatbots, now there's some type of memory in there, so they learn over time, which is nonsense. Instead, we need to treat AI the same way as we treat our team. And how do we optimize now human work, and therefore we will also optimize AI work. This is where the basic folder structure comes up that applies not only to AI but also to human beings. And this is where you probably have seen that so many people talk about Claude MD, and you have to have agents, and you download skills and plugins and all these things. Forget about it. It's so much more simple to use AI. If you know how to manage a human team in the most effective way possible, you will know how to build your AI agent team. And I will show you now the basic structure, and then I will talk about the examples that I'm leveraging AI on a daily basis in our own business.
So we have a folder, let's call it PKA. PKA stands for Personal Knowledge Assistance. You might know PKM, Personal Knowledge Management. We think now it's the new era of PKA, Personal Knowledge Assistance. PKM was always a manual work to do to manage your knowledge. There was just a slight layer of automation. It was possible. But in the end, to connect the information, to link the different insights, you had to do these manually. And you had great tools already with the launch of Roam Research back then. Notion was the next one. Obsidian came up, and you learned about backlinking and how to connect your information. But all this was manually done. Now we enter the new era where AI takes over this management of linking the information together. And that's so powerful because this alone is just worth to me $200 per month for the Claude plan. But I see many, including my father, who's leveraging this too. It is really something you need to understand what you need to offload to AI that makes it worth it. And knowledge management is one thing out of this. But here in this video, we don't talk about just personal knowledge management. We talk about real work operations that AI can take over. And yet, it is still based on these principles that I show you in this folder structure.
So in this folder, we have subfolders. And in these subfolders are Owner's Inbox, Team Inbox, and then for this video, we call it AI Team Inbox to be clear that this is just for me personally as an expert leveraging AI, delegating work to AI to do the work for me instead of a human being. So that's why I keep using PKM because work, this is my personal work area, and I'm not sharing this with the team. I'm just optimizing my own work the same way other experts organized and optimized their work already previously. BKM, which is Business Knowledge Management. And in here, we have subfolders. And these are just examples to start with. SOP, Standard Operating Procedures. If you don't have this in your team, and people constantly reach out to you asking for information and how to do things, and you say, "Man, I told you already for the sixth time, why, how this is done, why do you still ask?" Well, maybe it's time to create an SOP for it or work instruction so this person will never reach out again. Then the general knowledge, like the business identity, but also the expert knowledge, right? Let's, let's actually say expert our knowledge, which is your own knowledge, right? In this example. And I think we can keep this. We could now talk about workstream and so on, but we want to keep things very simple, and this will go already a long way. Then we have a folder, Team. And in this team, I have two agents. I call Nolan, which is HR, and I have another team member called Pax, who is a researcher. And on top level, there's a file called Claude.md, who is Larry. And those who follow my channel know what this means. Those who use Claude know what Claude.md is, but it still applies to any other AI model too. In fact, it is just a formatted text file. That's just the explanation of the identity of the AI agent working for me. So this is the root agent, who is just an orchestrator, delegating the work that I put into AI Team Inbox box or that I provide through the chat, and understands who is in our team to work on these tasks. I will show you in a moment how this looks in my own business. But this is essentially the basic setup. So you see here, why is this a HR agent? Well, this is now the perfect setup for scaling your AI agent for any kind of work that you might consider that is needed. So this Larry knows that if there's a task that nobody inside the team box can work on, he will reach out to Nolan, and Nolan will reach out to Pax to research for the best person possible to do the job. So they research for real experts worldwide who are best in doing this, and what are the different key indicators that make them the expert in their field. And then they hire it, which in fact is just creating a new folder in here with a new expert. And then Larry can actually delegate this work to this expert. And maybe you see it, this is how you would work in a business too. If you have a specific use case nobody in the team is capable of, but you see it's crucial for the business that you do this work in your business, you look out to hire somebody who can do this work. And this is exactly how you do it. So once you hire this expert, in order to make him stay aligned with how things get done in the business, you have SOPs. So this is where all these AI gurus out there overcomplicated things with prompt engineering and skills and so on. No, you just need to have a simple file where Larry knows that he needs to look up the team roster to understand who is in the team and what are their expertise. Is there anything who could do the job? If they call Nolan, as the example, for example, he needs to hire somebody. He looks up the standard operating procedure, which is in here, to know how to hire somebody, that there needs to be a folder created for this new expert, how the expert needs to be written, how Larry needs to know about this new expert, and so on. All the things you don't even need to think about because Larry will do it for you, and the different experts too. And this is where you grow the folder with the different SOP for these experts. And all this is readable. There's no code involved. If you have this already available for your team, you can leverage this now to train your AI team to get the right context to do the work in the most efficient way possible. And this expert knowledge or company knowledge, this is the general knowledge. So why do we have this extra folder and we don't provide this to the experts themselves? Well, because maybe two different agents need the same SOP. You don't want to have duplicated information in these different folders. Guys, this is something we teach for decades that this is the single source of truth principle. We want to have one single SOP that two different people are leveraging. So this means I only need to update one file, and these two will always use the latest information. If I start scattering the information in different folders and so on, people get confused. What is now the latest version? What is the thing? And that's where things fall apart. This is the very basic setup.
And now let's go all the way back to the very beginning of this because now we can build inside the team but also for AI agents this feedback loop too. And that's called persistent memory and actually self-improving AI agents that nobody so much talks about, but it's actually pretty simple to set up too. You just need a logging. So there are many ways you could do this. You can give these agents their own PKM system, which is just a file where they log any sessions. Okay, but let's not make this too complex for the AI agents individually. Let's just talk about Larry. So Larry could have another file, let's just call Larry's Journal.md. And this is session logging. Okay. And that's where people say, "Uh, I set the things up as you showed me, and I use the same prompt, and then I have a different outcome and so on." Guys, this is why I make this video to make you aware. 80% and always different outcome. So, if you understand what you're doing and why you're setting these things up, you wouldn't complain that the things are not exactly working as I showed you because this is just not possible with AI. But if you have the right foundation and structure, you will have your AI team always in these guardrails. And this is the same here. If we would give you a template and say, "This is how teamwork should work," and you just need to use these buzzwords, as I mentioned previously, right? As many consulting companies do, you just have to follow these 10 steps and everything will be amazing. No, it won't because it's not applied individually to the specific use cases. And this is why all these things fail because the expert, the team members, they don't think about their own situation and get these foundations right. And I showed you with this, right? Even in huge companies like this, nobody thought about this. Just pause for a moment and think about the things that we are actually doing and how things are connected. And if you do this, you will be able to apply on top this optimization that we are talking about. And the same applies now for AI.
Too. So if you do this, you need to train your AI based on this. But if you don't know what is the outcome you expect, it's impossible for you to train the things. And that's why people, and that's very dangerous actually, that people take everything for granted what AI says because AI gives you the feeling that it is perfectly the outcome and very confident that this is the right thing until you point it to an error and then it starts apologizing. "Oh yeah, you're absolutely right." This is what happens if you don't have no clue. And this is why a person in this situation will be never able to create AI for them to work and launch a multi-billion business. It will be always the expert being able to leverage AI to improve their work and delegate work so they can focus more and more on their expertise. Those people who are in danger with their jobs can now become experts themselves and learn in a very specific area and things like this and delegate the work then as well.
So we talked about the persistent memory and the positive feedback loop and this session. So here, every time I end a chat, a conversation, and by the way, I always try in one conversation to stick to one specific topic with Larry, who delegates it to the team, and when I have the outcome, which usually is then provided inside the owner's inbox. This is where I can review anything that they worked on and I agreed on this. I say, "Close chat," which is just a command or prompt that says, "I will let Larry then lock the session," which just means that Larry writes down his conclusions from the sessions, his insights, the back and forth we had, things that he did good, and things that he did bad. And this is now locked in here in a chronological order based on dates. So in the next session, he will already check this and understands, "Oh, something is going off rail," or if he's not certain how to do things, he will go in there and check the things. But this is just one part of this persistent memory because this expert knowledge and more folders are in here where you expand the general knowledge available for the whole team over time. And this is how you grow the expertise of each individual AI agent too, and this is how everything self-improves with each session. And you see here, this is just local folders. There's no mention about code. There's no special things going on or that you need to download something from GitHub or anything like this. And I showed you already in another video how simple this is. And I give monthly workshops where we build this from ground up too, if you're interested to follow along live with this. And this is the basic structure of any agent team that I would build. And before we now end the video, I will show you now this in action with my own team. And in order to show you that this is really just based on folders, I show you the the folder of my Larry version that is working inside the Myore business. And this looks a lot more complex than the folder structure I showed you. But this is because it's grown now for several months where I extended this over and over. It started with the simple principle that I showed you here. And then you will realize that you can expand this more and more over time. The important things here is you see the BKM, you see a clot.md. If I open this up, you see here, "Identity: Larry the Smart Fox." And now, why why is our fox? Well, for those who don't know Larry yet, you'd go to our website and you can go to "Team," and on this page, you see the whole AI agent team that is running in our business and helping us on a daily basis. And you see on top, there's Larry the Orchestrator. You can even go in there and he, you know, see the insights, who he is. He has his own identity and so on. All this is not essential, but I like it to be more connected to these agents. And I know perfectly who I need to reach out to and who Larry is reaching out. And if I'm now looking in this, this is Larry essentially. It's just this file here who, if I launch it, identifies, he knows the team architecture. He delegates the work. He knows, you know, what he should never do, that he should never work on these things. He always should delegate work. If this is not possible, he should reach out to Nolan to hire more people. And if we open up now the team here, you see there's a lot more than Nolan and Pex. But you see here is Nolan, and there is Pex. These are the two core agents together with Larry who expanded this list of team agents to work on different things. So an example is Iris. She knows all about our brand design. Now I can leverage this. So we can actually have Pixel to create thumbnails based on the brand design. We have Charter who is able to create infographics and so on, who are all based on this brand design. And this is why all these things don't look generic. In the BKM system, they have access to a lot of references, right? Where I said, "Hey, keep this in mind for Charter, for example." In BKM assets, here, diagram references. And here you have negative and positive. So this is something they created that I didn't consider good. Also, I think probably a lot of people would say this is good, right? But there's also instructions why I don't consider this good. And then in positive, these are the things that I say, "Okay, these are positive examples." This was not pre-created. This was created by Charter, by my AI agent, and I said, "Save the positive and negative in the BKM system for later reference." And now she always can go back here to the references.
I told you about the logging system, right? That they log their sessions and get their insights. But I goes one step further. There's actually a QA agent, Quality Assurance, who will always double-check the work of the agents, what what they have done. So she has a general understanding about the rules and guardrails. And whenever, for example, Charter created a new diagram, Vera will go there and double-check if this is what we need it for. So there's a feedback loop before it even reaches my owner's inbox. And here's the owner's inbox. There we go. And here's an archive. See, whenever I close a session, all the things that I had in the inbox get archived in this archive. And you see here, there's a lot of things that we did, guys. This is all going all the way back to February. This is all in the owner's inbox and it is archived as a history. But it's not only archived in here. Whenever there was something necessary for the team, it got extracted and expanded the business knowledge management part here. And here are just working documents that are still there. And this is the owner's inbox. And then here, in this case, it's called Larry's inbox because I just need to share it with him and he will delegate it to the team. It doesn't matter in the end. It's not about if this is Larry's inbox or team's inbox. It's about giving it the right definition what this is for. It's not the name, it is the functionality behind this folder that you need to provide to the team so they know how to leverage this. And then you build this over time. And then we have in the BKM the whole understanding of our Icore book, all the Icore journey course lessons in details. Then they have access to nearly a thousand articles from Paco and video transcripts from me and the podcast episodes. Then two years of coaching transcripts where Paco and I were coaching business owners over time. All this knowledge is a business knowledge management that our AI can now leverage to become really useful agents inside our business. And therefore, this is not all stored here. This is where our AI team has access to our learning platform where you guys can sign up, where you get access to the videos that I make with expanded details. So if you want to have more details about this video, you will have it here with a lot more. You can access this for free. And then inside the community, we have Jex, who is our community manager. You can see here, he's our community manager. And this is something he, I trained him over time. He creates the announcements. If there is anything new, he writes the announcement. He has the connections between the different things. And now he is actually called by so many members. They prefer to reach out to Jex because they get instant response. And he is so comprehensive in this because he not only points people to the right articles and videos that help them to move forward, but also different conversations inside the community that are relevant for the specific conversation. And this is amazing. Something that Paco and I couldn't do to have this full scope of all the different conversations that are nested in different places. And also not possible to hire a community manager for this to have this insight too. And this is where Jex is really brilliant now because he has the knowledge about Icore and all the insights. He has also the knowledge of the conversations. And this is why you see, just having this is saving already a position for community manager that we need to hire. I could go on and on and on. I can make a dedicated video about all the things. So, this is so powerful where I just think it's crazy if someone says AI is just hype. These are just people who are closing their eyes and gathering other people around who are scared too and they say, "Oh, it's fine because this guy says it's fine." Guys, it's not fine if you worry about your job and you have the feeling after this video that you are in a position where AI could replace you. We are here to help you to take advantage of AI. If you're a business owner, I hope this video helped you also to understand it's not about buying a bunch of AI licenses and give it to the people, but to actually think about what's going on in your business first and empower the people to build their own AI teams to work with them.
People also ask, "How can I now share this folder with my team?" You shouldn't. This is for the expert themselves, and the output of all this should end up in the business systems, the project management tool, the different documentations, and the databases, and so on. This is where the output of the expert's work comes in. But the expert's work will get amplified if they use AI properly. And in the past, it was the same for teams where it was clear how team members can help the experts to thrive too. So for developers, they had junior developers, and they do the minor work, and yet the senior developer had always to review the things. Same here, AI cannot do it all for you, but it can take away a lot of friction of the things that you had to do. And in another video, I showed you about daily journaling, how powerful this is. I have endless examples that I can share with you, and I make a dedicated video about all the examples of AI what's doing for me in a business. But I think this video would go just too long to show you all this, and that's why I will stop with this video. I think it was just very important to me because we see this over and over again to bring this point across that it is neither AI nor software nor automation that is failing. It is the foundations in your work and in your business that doesn't make these things work. And this is what we provide with our Icore journey, where we cover everything productivity end-to-end, digital note-taking, PKM, personal knowledge management, task management, project management, and already for years, automation came last. And think about the circles: input, control, output, refine. The refine comes in the end. What do you want to refine? If you have no clue about these things in your life, once you start realizing what's going on in these parts, you can think about automation. And now we have AI like a pro also in the Icore journey as a course that help you to thrive in AI, which is also last. So we have so many people joining each day. So we are so, so thankful. When we go to members, let's see. Man, we are reaching 4,000 members already. This is amazing. And we have all these people joining due to AI, due to the AI videos that I make. But once they come in, they realize, "Oh my god, there are so much more that I need to learn." And now you think, "But I don't want to learn all this." Now, the thing is, you will see the moment you realize about single source of truth, where do you take notes? How do I retrieve information? How do I manage my day? How do I plan out my day to actually achieve my goals? And all these things, you will speed up your work so much before you even think about automation or AI. You will close friction, you get clarity, you finally feel fulfilled in the end of the day because you know this is exactly what I planned out and that's what I achieved, and I can compensate any unexpected events. And I showed you in the beginning of the video all the experience that I went through, incorporate our co-founder Paco, with multiple businesses. This is a business-proven methodology based on general wording. What is a goal? What is a project? What is a task? If you start the first session in the Icore journey, you know digital note-taking. The first lesson is, "What is a note?" People consider they know what a note is, but they don't. And that's why we have the growth assignments here, where each lesson has a growth assignment. And that's the uniqueness to this learning platform because you can go through the whole Icore journey just answering these questions, and you get so much insights out of these questions, questions to understand the gaps in your productivity system end-to-end. So that's why we always recommend that you go through this journey with all these questions answered, and then you perfectly know how to go back and see, "Oh, well, there's a gap." And now you can go into the vertical content, which has a video, a TL;DR to really get you up to speed. But then you can really dive deep because these lessons are really going deep into each concept and workflow that we are teaching. And this is where you can really master productivity end-to-end, which will take time. But if you focus on the gaps, you will speed up along the way. You don't need to stop working. You will keep optimizing. I was in a high-stake environment in corporate, and I had to build the plane while I was falling. And this is where what all our people have. That's what Paco and I need to do in our own business constantly. We are all in the same boat here. And this is why this community is so amazing with our weekly coaching sessions, where you guys come in. And this is just amazing to talk about exactly what I just shared today with you. And if you're interested to join us, you can join for free. You can have the Icore journey kickstart course that gives you the insights, and you have access to all our additional learning resources for free. So feel free to join us. You're always welcome if you're not a member yet.
And one more thing that AI does for our community is the events recordings. We talked about the meeting minutes, right? So if we look, for example, on a simple weekly coaching session, how many people of you been in coaching sessions and you might get access to the recording or the and the transcript. This is where we put it a step further again. We not only have the transcript, we also have a comprehensive summary below, and this summary has timestamps. So I can always click on this timestamp and it comes to the position in the video. And there is also, I can filter by questions. It identified all the questions. I can quickly jump in the section of this questions. It also identifies the different tools. So it jumps wherever we talked about Obsidian. Boom. You can watch this here. So this is where AI thrives too for us because this is impossible for a human to go through this without hiring a dedicated person just doing this each week. And then obviously Jex is aware of this and points you to the right sections. So this goes a long way. This is where we really see now how people are able to leverage the knowledge that we gathered over all these years inside our community too.
Guys, what a video, very intense. I think it was a bit different from what I shared in the previous videos, but I think it was a really crucial video to share with you so we really get a different perspective on this AI hype. It's real. It's happening. It's changing, and it will just accelerate. I don't want to make you afraid. When you know what's going on and how you do it, you will approach this differently. There's no need for fear. There's just need to be prepared and to leverage the things that are already surrounding you instead of closing your eyes until it's too late, and then you constantly in reaction mode because you missed out. And that's what we here for. And if you want to dive deeper and you haven't watched my previous videos, then sure to subscribe to the channel, and I'll catch you up in the next.