Paste it into ChatGPT or Claude and interrogate the lesson in your own words.
You most probably used AI in some shape or form for your business. But let's be completely honest. How much time did AI save you? How much time did that ChatGPT conversation save you? How much money did AI make for your business? For 99% of people, you're actually working more and making less money right now because you added AI to your business. You have a lot of tabs open, and a 30-minute task you used to do manually now takes an hour and a half, three times more, because you're feeding context, building the MD file, fixing the prompts, doing whatever you need just to get certain tasks done.
And this is not just an observation from us talking to business owners, and it is not just a feeling you have when you use AI. Harvard Business Review released a paper in February, which I'll show you below, and they put numbers to this. The problem is not AI itself. It is that you are most probably using AI like a tool. This was me as well, because we were born using computers before AI was cool, and now everything has shifted, digital transformation, AI transformation, and so on. We're used to certain habits, and we use AI like a SaaS, like a tool, like software, like just something we interact with, when the true way, the right way, and this took me time to make the shift, is to use it as an employee. And this is exactly what I want to talk about in this video.
We're going to cover it in three main points. Point number one, we're going to identify the exact problem we're sitting on. Point number two, we're going to do the shift in mindset, the reframe, from looking at it as a tool to looking at it more as an employee, and we'll bring in analogies from biology, how humans operate and do tasks, and from space missions, how we can use space concepts and translate what humans do to machines, because that's the intelligence of AI. And point number three, we're going to think about how we can hire AI, not build AI. And if you want the material, everything you see in this presentation, click the link below in the description and you can download it.
So the problem right now is that it is quietly making business owners babysitters. What do I mean by that? Harvard Business Review released a paper on this. They said yes, there's a slight increase in productivity, but people are working way more hours and they have this cognitive, mental fatigue in their head, and of course it's not giving them what they hoped for, which is to save time. You've seen the headlines, of course. You've seen the YouTube videos saying they saved 10 hours a day with this AI system and so on. Well, we have not achieved that, exactly because we as humans did not reframe ourselves in order to know how to implement or integrate AI within our business.
So the main reframe I want you to take away is: always stay a business owner. You are a business owner, and do not become a babysitter. The reason I say this is because every single week I hear this five to ten times from business owners. They say, "I thought about AI. I tried ChatGPT or Claude," and I'm sure you went through this process, I'm sure you'll relate. So I went through ChatGPT, and there was step number one: I thought I want to automate this one process that we have. Then I said, why not automate the full business? Then I thought, why not vibe-code a dashboard for my business? And then I find a bug, it's not a linear path, and I fix that bug for the next three to five days. Then I didn't like the UI much, the graphics, so I fixed that for seven days. And then I got too busy and away from marketing and sales, the main business functions, which I didn't do much, and I focused on AI for the next three months. And here we are today with less money. I wasted a lot of time, and now I need to go back and do my marketing and sales. It's a rabbit hole it takes you down, because those business owners became babysitters and did not remain business owners.
The solution, of course, is to treat AI as an employee, and this is exactly the methodology I'm going to show you, and it's very simple. Here's where I make the relationship between space and AI. When you launch a satellite into outer space, in space you have the sun, you have gravity, you have a lot of things happening, a lot of unpredictability. So the satellite itself needs to understand, in order to conclude the mission it's going for, the different disturbances or external nudges it's going to get from space, without actually knowing what those nudges are or at what magnitude. So the satellite senses those nudges, decides on the information it has, acts and fixes itself, then gives itself feedback and concludes with the result: this is what I did, am I achieving this result or not? This is a very simple logic, and it's the same logic humans go through. I actually explained this this year at one of the conferences, a space conference, one of the big ones.
So we'll go into details, but it needs to sense, decide, act, and then get feedback for itself. And this gives you a slight perspective: the satellite is not acting just as a tool where we click buttons, but more like it's owning an outcome it wants. For AI itself, we don't want tools, we don't want to give it a prompt and have it just give a result back. We don't want a straight-line movement, that's a 2000-to-2010 mindset we have toward software. With AI we need a slight reframe.
So the old way of thinking, and this is how the majority of people getting into AI use these tools, is that they prompt it, "hi," AI says "hi" back. Or "what's the colour of Earth?" and AI gives an answer. So it's prompting like a tool, going back and forth. "What content should I create today?" AI gives an answer. "What do you think about this hook?" AI gives an answer, and so on. But an employee, on the other side, you set an outcome for the AI. As a business owner you'd say, "I want 50 leads from content," or "I want 50 sales from content marketing." You own that outcome, and you work toward it and get feedback until you achieve it. That's the slight shift you make in your thinking: set outcomes, don't just prompt the tools and get answers out of them. That's the first part I want you to understand.
So, as a human now: if you give me a task and say, "Hey Adham," or if you employ me, if I become an employee in your company, you say, "Adham, I want you to take care of the ads, I want this many ads, this is the budget I have, and I'm expecting 40 leads per week from the ads, can you work that out?" I say, "Yeah, sure, I can do it." So what do I do? I sense: what is the current state, what is the future state we want, and what's the gap between those states? Then I make judgments, I decide what tasks I'm going to take to achieve the future state I want. Then I start executing on the decisions, the judgments, the plan I made. And then I check: did we achieve those 40 or 50 leads per month or not? If we're at 30, there are certain things I didn't do, so it goes back, I sense, I make a judgment, I decide another set of tasks, and I act on them. And this is the loop that keeps going. So that, essentially, is an employee and how an employee would think.
A nice analogy is the AC. When you turn on the AC, you set the temperature, say 20. Does the room temperature quickly go to 20? Say you're at 15. No, the thermostat still shows 15, then it starts sensing, deciding, increasing, and if you keep it on auto it raises the fan speed and the cooling until it reaches 20. What happens when it hits 20? It calms down, because it hit the target. It looks at the gap, and it closes it. Similarly with cruise control in your car. You're at 120 km/h on cruise control. If there's a slope and you want to go up and hold your speed, a gap appears, because you need to push more on the engine. So cruise control fixes that gap and keeps sensing and deciding until it reaches the target. This is how we did it in satellites too: we have a target.
In business, as a business owner, you have certain targets. You want 20 leads a month, 20 clients a month, 20 pieces of content a month, I'm just saying random numbers, but you for sure want something. So this is a feedback system. There's a gap: you have a current state, say you're at zero, so the gap between 0 and 20 is 20. The controller here, the AI, decides on certain things, acts on them, then checks the result and sends that feedback back. Say after deciding and acting, it turns out I have 15 leads, not 20. The master told me we need 20, so the gap is 5. Let me decide some more things, act on them, and see, now it's 17, so the gap is 3. Let me decide more and act. And you get the point.
Here's a nice example. This dotted line is a human. You say, "I want to target 20, this is my target." Usually what happens, because we have emotions, we're humans, things happen around us, we have responsibilities, the business is not everything to us, and we get distracted, we have feelings, sometimes we don't feel like working. So with a 20-leads-a-month target, you do the work, then you slack off, do the work, slack off, "I'm super motivated, something nice happened," then slack off. It's not consistent, and most people don't achieve the target, because it takes a lot of brain power, knowledge, and self-awareness about yourself and the business, the constraint, the problem. But if you give it to an AI running these loops, and depending on the design you have, the AI can react, remember, and anticipate. React: it can be very patient about achieving the outcome, or it wants to achieve it very quickly. Remember: it can never forget, or it forgets a lot of what happened in the past. Anticipate: it acts in the moment, or it sees things coming in the future. So depending on the design, you play with these parameters and specify the personality, or behaviour, of the AI to achieve a certain outcome. You keep playing with the parameters until you achieve that outcome and it becomes a stable system, locked on the target you have.
And this is a very important point. When you start this on day one, because the gap is too large between your desired state and your starting point, between 20 and 0, what happens? It sucks on day one. You see maybe 25% of the full process working, because it's learning and the gap is too big a leap. You know the saying, take one step at a time. So over time, with this feedback loop, with it remembering things, reacting, and having a certain behaviour, it improves. An analogy: if you bring an employee with a strong, photographic memory who never forgets, versus an employee who forgets immediately and doesn't remember what they ate yesterday, they have different personalities and they'd achieve the outcome in different ways. It's the same thing here. That curve depends on the personality you give your AI. I hope that makes sense as an analogy, and that's how I'm approaching AI right now, to make it compound and work for me.
Now, if you want to score this, say you brought in an AI system, or you have one, or you're building one for your business. Let's take a Dubai real-estate brokerage that added AI. They say, "We're AI-native, AI-first, we added agentic AI," and it's a chatbot on their website or Instagram that answers in Arabic and English. So you take four stages, a rubric or checklist for any AI system, to make sure: is this an employee, or just a tool? Point number one, does it sense? Yes, it reads the typed questions in Arabic and English and gives answers, a back-and-forth. Does it decide? It answers a few FAQs, yes, but it never qualifies based on the budget or whether this is a real lead or not. Does it act on the knowledge? No, it doesn't, because it doesn't book a meeting, a call, or a viewing, it doesn't push to a CRM, it doesn't follow up, so it's not taking the extra step. Does it self-correct? Does it understand this individual, or a pattern, how many leads we get each day, what's happening, the patterns for the business? No, it's not self-correcting toward the outcome you want. So this, essentially, is not AI as an employee, it's just a tool, a chat widget you've added. You haven't really added AI to your business.
So in summary, what we talked about today: the problem is we're super busy, going down rabbit holes we're not supposed to, because at the end of the day we're business owners, we want to make more money and more sales, and we're not going to be tech nerds lost in AI. Point number two, always hire AI as an employee, not a tool, and make it run loops, sensing, deciding, acting, and self-correcting, which is the feedback phase. Earn the automation you have through this checklist. And don't say "I added AI to my business" when it's just a widget, because this is exactly what I look for when businesses tell me they have AI. If it's not sensing, deciding, acting, and self-correcting, you're not building AI, you're not adding agentic AI, you're just adding something that wastes you more time and doesn't achieve the outcome you want. You want to employ AI that compounds over time, where you, the founder, are no longer the bottleneck of the business.