Hello, everyone! We're on ToTheMoon. Uh, and I'll show you in our graduation the real case of the real business. I have, uh, a share of one business in mortgage lending. I mean, I had a real challenge to assess the performance, the efficiency of this business, the efficiency of the business unit. How marketing works, how sales work, and how all managers work, how the leather comes, how the advertising campaigns are done. Make such a very serious assessment of all this, yes. Not just to see some kind of phone call or some kind of transcribation, not just getting some kind of report, but getting, uh, a system that is, uh, really artificial intelligence, uh, and, uh, a Not just another little sof or some chat-bot. And not enough of this, he, but, incredibly useful to people, not to get a sense of automation, yes, again, or someone who has a business in commerce, but to people, to understand how to do things. In fact, artificial intelligence is built, which must be thought in the development of certain tasks, which must be thought of in order not to make mistakes and to create a real system, but to create a system. the truth, lay the foundation for the future. To enable you to see and foresee many things ahead of you in terms of your career or your business development. It's about how not just increase your income. It's a completely different mindset. And that is, of course, a real example of how artificial intelligence becomes a real infrastructure in business, and not just a microchat or some microsystem. Let's go see you. So, uh, what happened? There's a real estate business, mortgage lending, house sale. It's all USA, yes, it's a pure US business.
And in this business, sales managers are working, and they interact, there, the co-locators, and, uh, managers, sales managers, the markers, etc. And there's a whole series of different systems in this business. There's a Bonza CRM system. It's a CRM system that is used in the mortgage--- the real estate market in the US. Arrive is a loan system, considering that we are connected to dozens of banks and have the opportunity to give, uh, to seek people with the right credit for buying housing or refinancing housing. I'm a good man. And there's a commercial that's done over Facebook or Google. Well, there are different extra sites, yes. There are systems in the financial sense, there are chat systems, and so on. So, uh, first thing I ever had in my interest when I was at different meetings, asking questions, I had a problem in aggregate data all the time, on the one hand.
And the second story I've always heard that new managers who go to work, my partner has complained very much that they're not good or they don't know what they're doing. And I was very important to check on one side their quality, and on the other, to teach them to work. What did I do? First, I quickly, uh, upped a system that connects to the infrastructure of our CRM-Bonza system. What did I do? I've created a draft in Codex. But I had Codex in parallel, and Claude. Codex is 5.5 High, Claude, mostly four--- 4.8, there, Ultra Claude or Max. It's just fixed, as we know with you, uh, Wednesdays, but here we'll show them again. It's a very easy story. There's an app for Mac. I went back to ChatGPT and said I wanted to contact the Bonza system. In this system, in Bonza, I came in and I gave myself a PI token. For many people, it might be difficult now that I'm telling you, but it's a good thing to listen, yeah. Because the point here is not that it's hard, but the point is, I want you to show me what you're doing, because it's unbelievable from what you're doing. I came in Codex, I went into ChatGPT, and in ChatGPT asked me to write a task to contact-- to build a request for Codex in Claude to contact Bonza. Oh, that's what API is for to fully examine all the documentation. Many people ask me how I write proms. I don't show it because I think it's gonna be absurd. It's not about ToTheMoon, and you can write the proms you want. The point is, you can just ask that there is a Bonza CRM system, and you want to contact API. You can ask for something enlarged, yes. I understand that I have a skills in the creation of different systems, and I'm more confused about what API is like working. I understand that somehow it's easier to do something, yes, someone might be more difficult to do or someone would do more than a little more. Next, after getting a connection to API, I started turning the system. And now I've turned around at some point in time, and I've turned the system into which the data were essentially down, completely collated like our system parallel. I mean, I'll remind you again, I've done two things. So on the one hand, my job was to copy, for example, all the leather we have, all, all the calls that exist there, all the realtors we have all, There, all the managers who are, yes. There's a volume of records that's been buried up there right now. I'm showing you now, there's a test base, and I'll show you on the test base.
There are a number of things, uh, there, plus-minus real, but to see what it is, yeah. I'm gonna have to... this data. On the other hand, my job was to show up... to create a system that would, uh, generate, uh, and answer different questions. And she'll answer questions not only on the portal, but also answer the phone inside the Telegram-canal. So one of the, uh, the kind of, uh, decent stories that was, it's very important. Once again, yes, it's that I have a Telegram-canal, that is, that the system works in Telegram Canal, and I could ask any question in Telegram Canal, and that I have a slave-to-slave system automatically show me everything. portal. I mean, I see that if people ask the question, uh-uh, "Say, the shares or cryptatuits take banks to the credit to issue a mortgage?" So my system can automatically check it out and can. I'm gonna ask you this question, yes. How did she do that? That's very important. How I got it. The thing is, when I synced up with Bonzus CRM and got all the calls, I got calls to her, I got an opportunity, uh, in this system, uh-oh. I asked Codecs to write a system to transcribate these calls, yes, and Codecs wrote a system of transcribing calls, pretty quickly absolutely. I connected her with the current system in terms of transcribation, and I automatically got phone calls decoding. There's a phone call for 19 minutes, and there's a-a-a-a-a-a-a-a-a-a-a-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the-the- So I've been transcribing some kind of call. So it's important that transcribation models are also available for different, different, there. Russian must be the cheapest, most efficient, super-breed, ElevenLabs. Or even raise, uh, there's a transcribation on your server. I'm telling you again, I'm gonna show you, you know, incredible things in terms of what you can do, uh-huh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh, uh I've been through the phones, I've got, uh, three, four, no matter how many, thousands of these first calls. And these calls... I'm from Codecs asking me to unload as Excel file. He's downloaded it to me in Excel Bayle. I went to ChatGPT and asked, "I need you to build a knowledge base on how people work, how managers work, how the scouts work, how they work, there, company leaders. In terms of, uh, this mortgage system." And I, uh, got it, and I, uh, uh, ChatGPT built this system. So not ChatGPT built this file. What do you mean, you built a file? Well, I understand I took it, created the file, uh, big enough, with all the knowledge of the system, there, the mortgage on the calls that staff were working, yeah. So he built it completely, there, with different details, different circumstances, and so on. I showed you some, again, example of these documents, just like some examples, yeah. I tried to get you some information, translate it, and so on. I built some big document, a big document, yeah, based on some calls. I mean, that's what he built on the phone. Many people find it hard to understand that, in order to build such a document, there was no need to have something special or unusual. I mean, it was enough to unload these calls. Look what I did next. I asked ChatGPT to build me an architecture of all the terms and to create a basis for me to answer different questions. There's a question-response coming from there. These questions are answers, they're working quite interesting. I mean, if you look at-- I'll give you an example.
There's a bot of a certain kind, and in this bot, I have the opportunity,
I have the opportunity, uh, to ask absolutely any W-- and managers have the opportunity. I created, I wrote, "Can I have a rights system inside Telegram Bot?" He said, "Yeah, I could give it to Telegram Bot." "Can I just, like, well, well, well, well, well, he, he did it all. For example, I can ask you something, uh, to know from, um, a man if he wants to buy--- to buy, uh, a house in California, California, and he doesn't have a family. Uh, annual salary about 100 to 20. I mean, this system, I see, checks out different sources. There, she, um, in the sample, she's gonna come in, get some information. And she'll write, if a man wants to buy a house and he's alone with a profit of about 100, 20 years, first you need to check not family status, but a basic picture of where he works, what kind of income, how many debts, how many debts.
I'm on the first charge, which is the payment to be made by the house and so on. Yeah, what do you want to ask him what data he can get? So she got the system. It's not enough to ask for more details, it'll give more information, it's possible to assess the issue, and so on. I mean, you can, you know, like, go on, yeah. So you can do a lot of different things. I mean, it was a very important story for me to make sales managers able to get ready, easily learn at any time. I mean, that was, that was absolutely necessary, yeah. So that sales managers are trained at any time, trained, uh, easily, trained independently, yes. And the second story, which, well, was incredibly necessary, that story is how to make me understand, and what people come to, how they communicate with them, and what happens to them. And I am a man who is very, very profound, fundamentally looking into business and understands that we are going through a major development of a whole of artificial intelligence, and I wanted to make the next story. I mean, I wanted to see that after every conversation, when the sales manager says, the system almost automatically evaluated the conversation that's happening, uh, not just talking, but how would it be to evaluate the work? The manager, she was assessing the quality of the client that was in this conversation. I've been tying up all previous calls from this manager, evaluating the manager, linking all previous calls to this client and evaluating the client. In fact, I did some sort of mourning. I mean, look, I have my final idea, which is very important to always look at when developing systems like this, what is the problem with artificial intelligence that people see a very narrow solution, and people are all. time makes very small things. Someone's gonna say, "What's the transcribation written? What's the Telegram connecting? What was it that was sent to the bot? I mean, I had a mission to meet one of the fundamental tasks. Well, first of all, I had-- what's the fundamentally important task? If I have a big leather tranship, a big customer traffic comes to me, how do I make these clients work in a good way? Quantitatively, it doesn't mean without managers, but it means there's a backup system that could help managers-- managers or service them without managers, that system, yeah.
It's about any business. You're in education, you're in IT business, you're in sales, you're in any service or any service. It would be a question everywhere: how to make the best possible, client service. And here's always the thing, if there's a really-- well, like a system that allows a very quick analysis of conversation or a pros-to-sex chat analysis, and that's, uh, a sales manager support device. And here is a very interesting one, look, the moment we have in our lives, making the task of certain things to be developed, we are based on what we believe it should be. Like, how do you figure out the call? What characteristics should be on the phone, the quality of the manager's job? I don't know, maybe his loyalty, not loyalty, uh, there, his conversation ethics, not the use of bad words or the follow-up of a certain instruction. Now, we're going to have to say that our demands are so limited. That's what OpenAI is and what's Anthropic, huh? These are companies that made a breakthrough. They already contain information when you don't have to figure out how to appreciate the conversation, there, in the area of mortgage lending, or how to assess the conversation in education, or how to assess the conversation in the field, For example, services, tourism services. This system, it has already studied the world, and she knows how to assess the human being as effectively as possible. And I've been doing this a lot of specialization these years, and I've been studying, going into some company, there, taking part in boards or buying or selling some business. I looked in these businesses, I took a straight shot of the data and tried to analyze what's going on inside in terms of the quality of people's work, and what happens in terms of the quality of these clients. I tried to understand, understand and figure it out. And in the first place, I really did make these parameters myself. How to assess the quality of the manager at all, or how to assess the quality of the fox or how to assess the quality of the software. How do you assess the quality of the programmer, huh, the job?
So you're saying, "This means, for example, the outload from Jira, you're gonna appreciate the swing-- and you're gonna appreciate the quality of the programmer," and you're gonna make some parameters for the evaluation. Then I realized that it was phenomenal if you were to know what it was like in this zone, you understand it. I mean, I understand how very serious the sales managers are, I really understand the job-- what is, marketing leathers, or customers, or how CRM systems work or how they work, there, ERP, XRM system, A, I was increasingly aware of the importance, uh, of giving these systems modern, such as GPT five Pro, for example, to give a customer profile. And, for example, I took a sample of thousands of calls there, and asked the system to assess me, and somehow, I'm gonna take some kind of managers on some parameters and oc-- well, evaluate my calls first, and then evaluate my calls. Managers themselves and then evaluate the leather. And if we look, the system offered me a lot of evaluation parameters. I want to show you that, yeah. So when there's a final evaluation, I always said I needed some final score of human evaluation over the period. There is, uh, an assessment of his full talk. There is, uh, a result of comparable calls, and it's a comparison with similar calls of the same assignment and close, uh, longitude, yes. That's a very important parameter, for example, that the system offered me to count, and I was kind of aware that it was, but the question is how to count it. This is, for example, the level of weak 20 per cent- to 5 per cent conversations. It's how much choir-- how well the manager makes his weakest calls. I mean, the higher the system failures. So when he has weak calls, what's going on in them? And a place like-- well, it's clear that there are ratings, and then look, there are criteria estimates. And if you look here, the evaluation of the criteria will always be different from the sales manager, for example, for any man who you're gonna appreciate. You're assessing the man who recruits, recruits, you're assessing the person who communicates logically, sells goods or buys goods, makes purchases, doesn't matter at all. They all have a lot of different parameters, yes. And since we, for example, have different types of communication, i.e., I don't know, real estate purchase, refinancing, general mortgage consultation, like, or investment, investment, The mortgage, and so on, there are different calls, and they still have to be assessed differently. And this is how he's highlighted the quality parameter, we'll show you, yes, here. And he also says there, like I-- we did it on 50 eight meaningful calls, so you know, there's, like, 200-five minutes, you know, that level is incredible. Not enough, he gives you every phone call, he shows you that there are phone calls, but there are phone calls, down there, below the standard, and you can go in any phone call and see inside the evaluation of any total call. But it's not just an assessment of any call. He knows what people talk about, gives you segments of this conversation when some stop-words were spoken, and shows you why he counted, for example, uh, the next parameter. step by step, the agreement is 100. And that means that there's something here, like he's got a description and shows you that description, yeah. Okay, look, it's very important that the initial analysis makes ChatGPT five and five, yes, and it makes some kind of phone call.
And someone asks you, "So, five and five on API will be very expensive, it's not realistic to use." But that's the paradox of the world, which everybody's trying to build systems using cheap models. Your job is to stay in a combination all the time. You have the opportunity to create a certain analysis of data not originally for API, but to open a system like ChatGPT five Pros or what you have, and completely inside, to estimate a certain amount, yes, and, T-critical, they're worth a few copies, yes, a thousand seven hundred calls, that's, uh, ten dollars, there, well, that's a small amount of money, yeah. Now, through the expensive system, do some evaluative things and build a potential model or see what model I can calculate. I'm gonna write it down, uh, extra.
Uh, additional information, considering that we're having special subjects now, uh, every three days. And yet you're not missing it, are you? So the OpenAI released a new model. It's ChatGPT, uh, five points six, and you might not even hear it very much, because it's not likely that it's gonna work out at the time of the release. Maybe she showed up. I wonder if five points of six models have three separate names. I'll tell you about it on the special edition we'll have the other day. But the point is what? Whato, well, there's this soul in charge and the main cool, balanced model. But the fundamental story is what is it, if you look at it? So, uh, why didn't the model, uh, get out of here for everyone? The model was not entirely released to all because, uh, OpenAI sent it to the government for study. And, uh, the U.S. government. I recall that in early June the United States issued a decree that companies have the opportunity to give the State a preliminary assessment of new cool models, uh, to avoid allegedly having to face something. What can't be encountered? In May, the US Government stated that companies were obliged to give models to the State ' s preliminary mission assessment, and that they did not show such cool models as Mi-Mythos, Antropica. Then they said that they-- company somehow had the opportunity. And we remember that case, I was talking about it, and we talked about it a lot, and Antropica's got a cool Fable model, and the government said to block it, at a very serious level. We still have no access to her. And there's five six. And, uh, two days ago, we had a, uh, sub-cate on the weekend channel. And Elnar said it's now dangerous for everyone to produce new models because if the model is available, it's probably not that qualitative, because it's not blocked. And if, uh, uh, well, she's being reported completely or if it's high, that's the possibility that she's being blocked. Well, the OpenAI went to an agreement with the government. Sam Althman has released-- he has made a statement that he is not opposed to working with the government. We'll see what happens. Obviously, it's gonna happen now. Five and six will be available for all the yia Fable, too. This can certainly be considered, on the one hand, that, uh, the State is, uh, as some, some such interaction with these companies would show. It's good for the market. On the other hand, it's absurdly unstable, and I'm with it, uh, I really agree there's some chaos. And when there's chaos in this story, it's gonna lose ordinary people, yeah, ordinary people, and the State's gonna lose. Uh, keep watching, uh, my proo thing that I did, uh, in Claude, Cody and in, uh, Codex. I hope you like it. You write it down. I'll tell you more about these applied cases. I think this real estate case is very interesting.
And look at the other day, uh, special provocacy five points, six ChatGPT, and, uh, the thoughts, uh, about that. Mm-hmm, what happens with the opening or closing of the model. It is of course important to any man in the world who is absolutely, uh, totally, because it seems like a story, but it's real-- it's more important than the closure of borders. Yeah, one of the stories of the troubled people now that they are, it's hard to move around the world. People can't just travel, they can't see each other's family. Now, of course, this is the subject of the closure of artificial intelligence, I think people don't even know the importance of this issue.
For example, we in our system have determined that we will use three models. We'll use four O Minis to initial evaluation of the post-recruitment. We will use five points four for, uh-oh, more complex cases 10 to 15 per cent and use five points for critical situations. But it's important that the system automatically looks at the manager's complans. For example, we have an important story, using all the critical moments of sales managers. When the sales manager reports something that is illegal or false or incorrect or does something wrong. So I'm showing you the test base again. There are leathers. I mean, I'm gonna see if there's any foxes and ice, I had a moment, and can I just get a leed out and see if this fox has any parameters? And there was a story about how I could make this ice exterminated, and, uh, uh, na-- used both old calls and notes and M.S. messages, emails, messages, there, with WhatsApp, with Telegram, with different systems. And plus, uh, some kind of evaluation system existed. For example, he determined that she was in a state of urgency and intent, involved, prepared to proceed, how complete the important information was collected, loyalty, whether there was any loyalty or any kind of loyalty to you. Risk of competition and communication. We're also gathering a different conversation from him, some conversations where we are, everything's still there. So it's clear that there are leathers, for example, with low quality processing, there, and right away, for example, we see that there's a potential for a client's flow, like low, yes, or a willingness to work together is low. Well, for example, how much information about him is collected. For example, we have, for example, gathered very much information about him. I mean, we, uh, analyzed a fairly large amount of data., I'm a system like ChatGPT five, I've got it all figured out what the real thing is, what they need to use them further inside. I've made some system.
Well, I see it's, uh-oh, micro-recorded stuff. She described some methodology, described how this methodology works, described, see, that's the concept of a full-scale evaluation of conversation. I mean, I wrote how we'd appreciate it, like, when we purchase real estate, how the scale will be broken. And I did that information. I'm already downloading this information to my system and I'm saying start--- first, I'm in a position to, by the way, get out of my system, like, 30,000 calls, if they're identified, and analyze. They are not spent separately on GPT-5 chatting and loaded further into this system. If it's clear that many businesses have a million calls, it's a little different, but even a million calls are getting well, it's a good thing to do, and you can build. What's the big story? That my system has allowed me to make an incredible big step. I want you to know about him today. I mean, having actually collected all the calls or collected all the information on all the lids, I had the opportunity to analyze these leaf very quickly. I asked her to build, for example, HTML records, in leather and call forms, and the quality of the calls. And I mean, the system is building this incredibly. Well, that's, like, she's very quick to connect data on lids to me. How many leathers we've reached, how many lids have reached more than three minutes, how many lids have been in there in extra crows, how long it was before the first call, the leed, I mean, how many zero-talking attempts, like, four hundred, so this sample just had 13 lids in a four-and-eight test, that's more than ten times, and people never call. We're up. I mean, some quality of lids, raising, yeah. How many calls, for example, have been made so far. And then I can totally watch every source of all the statistics, see if it's gonna put them in the states and watch all the advertising companies, break them. Someone's gonna say, "We've got a system somewhere. Look, this report, it was collected in a few hours.
I mean, once I've had my system to connect with the data, it's just, in this case, it's one CRM system, it's not even an analyst system, yeah, because every fox has a source, it's me. The system built this report and enabled it to form and form these reports in the next few hours every day and at any time, not only here. She gave me the opportunity to shape them in Telegram. So, to date, any manager inside the company or any manager, it's clear that we have some of this test, it's a small, boutique business. Although we are a big broker in the US mortgage, I don't know what we're doing now on top thirty, I guess. Or, there's a top-up of $40 in the U.S., which is, anyway, it's not like, it's just, like, super huge, straight-out leather volumes. It's a thousand lids a month. It's not, it's not millions of clients there, it's an opportunity to check everything very quickly. But when I say three minutes, there's no question of three minutes. For example, this report I just showed doesn't have any transcribations in it, it's now embedded in transcribations and there's a real conversation. And the real conversation is not a matter of minutes. I was always impressed when commercial departments, sales directors everywhere said that, you should evaluate the call from 30 seconds. I've been saying 30 seconds all the time, it's not a conversation. And now our task is to say now, and this is a real call, and the real call is what's real? Well, that's what GPT thinks it's real or what Anthropic thinks it's real, yeah. And post-processing is a quick call to determine his reality, and it's worth a piece of parameters, it's very cheap, yeah. And when we know he's real, then we have the possibility of lids to assess and show up clients.
For example, a real man came in absolute, but he just doesn't like us or he has-- he's in a tough competition. So he's gonna go, study in other systems, find himself a contractor, like, or he's come to us, but he's got a backup, and in this case, he just doesn't fit a certain program. I'm sure he's gonna wait three months. He's a great, cool client. I mean, of course, it's a completely different level of advertising management, that level of advertising management in the world is less than 1 per cent, less. In fact, he's less than zero per cent, because even a simple leather to "sniff" levels makes up to 5 per cent of the companies in the world. A serious sorin in terms of the type of conversation, the recognition of this conversation and the analysis of the kind I said, the fox, the analysis of the fox, the analysis of the probability of the fox, is unbelievable. Someone's gonna say, "How will he appreciate you buying these clients? How can he score this client?" I said, "Look, we're gonna take, unload all our clients over the last ten years, and we're gonna get all the calls and make some map of certain numbers.
"Traggers or points as you can determine, this client is moving to buy or not, and we'll try to run this map in parallel." It used to be Maba-- ma-mo-mom-mom-mom-mom-mom-mom, yes, and in the machine-learning, you could do that. What-- not do that, but microprocesses do it. Now we're going to be completely in a new era, and I'm telling you it's not a spaceship. That's very important, yes. But to create this, we need to think of future yia, not real, because in the real world people say, "We can't put in a system, because we don't have a corporate rule, we're not. We can put in a system, we can't buy, we can't pay for that sophth, and we can't buy a system because the analysis of this call is too expensive. And we can't do this because I don't have access, I didn't get it from the marketing department, so on." Also, in parallel, I want to tell you that, for example, on the same sales managers, we have built a management analysis system now, yes. And where we have the opportunity to see a conversation quickly, for example, unload some time and look very fast, which every manager has done for a certain period or what he does every day. We have the opportunity to download the analysis of the individual manager and see what the specific manager did. By the way, I wonder if the system is very fast, very fast, it's not just, for example, that I think that this man made ninety-eight calls in a thousand, and there were 60,2 calls more. Three minutes, yeah. He says the middle team, like the one on the test sample, was there, so I'm looking at eight managers, that the average was 100, ninety-eight calls, and he's got a fifth. Or, for example, shows how many calls were-- more than three minutes in this case, yes, because it's a report without transcribing. Well, in a period of time, I'll show you how and how much it's gotten ahead. It shows, for example, the average of 10 percent on the team is the place, yes. And we say, "Ah, that's not the problem with that manager. That he's got short calls. Overall, most of the 10 per cent of the calls are more like that. How many unique clients were there? We see that a man has worked five hundred and ninety-eight clients over the period, the average was five hundred and seventy-eight and so on. There's a lot of data here. Like, how many minutes of conversation a day. And the interesting thing is, look, one of the know-hows, I was talking about him, and I'm the first time I've ever had a system of sales managers, markers, they're the first time in my life, they're letting me assess how much I've ever done.
I'm working as a manager for the day. And we've got a parameter like that, how much work he has. Like, it says five hours and fifty-six minutes. That means when the first call was made and the last call was made. And he says that it's 70% of the eight-hour working day, yeah, for example. And the interesting story that I asked, "If it's a sales manager who should keep calling, how many breaks he had more than thirty minutes?" And he says, "He had a break. Three razors without calls for 30 minutes a day on average. And on the team, he writes, two points eight. Unreal things, guys, from analyst perspective, unreal, just unreal. And the system is helping you build these reports, it's the way you talk, it's the way you write about the work that you have on the day and shows it all every day as it works. Now, for example, one of the very cool cohort analysis guys is almost nobody's gonna look at and what kind of lead the manager worked with. I see, there, I don't know, well, in this test, it's a test sample again, yeah.
In the twenty-third year, he, during this period, contacted one hundred and twenty-one clients from the twenty-third year. And it's clear that there are 100 or twenty-nine customers, and the calls, the clients are actually here eighty-one, and it's clear that only nine clients were effective. And, for example, since last month, there have been 12 of the twenty-one, yes. Previously, there are forty-six out of seventy five. Well, he's got different lid statuses, quality and everything. The correspondence, the reports, the number of reports, the number of cases where he left, the number of letters he was processing fast, the letter-working, the amount he answered. There are leathers he worked with, like he worked, where he made all the lists of all the calls that were long.
The system then introduces the notion of data quality, definitions, i.e. transparent data. For example, there's a complete source, yes, his phone calls with a known length, you know, ninety-seven percent. I wanted to write a video and tell you about these things. I may have been in a very rush. And I'll tell you something important right now. The thing is, this is the infrastructure, and you saw about 15 percent of it. Took less than two months and, uh, my personal total time, well, a dozen hours, right? Ah, it's clear that there are things I need as a business owner. There are things I need as Alexander Volchekov in terms of understanding the code and Clo Claude. And there are things I need in terms of understanding the intellect of intellectuals, how he works, what he can raise, what he can do, and so on. One of the most important conclusions I want to give you today is the incredible conclusions I want to give you today, where there is probably a big problem today, people are trying to create something very much. short-term and very short-term. So they're thinking very little steps. They're trying to get ahead of the task, they don't think forward. They operate on the basis of their knowledge, and often in business areas where they are not fully aware. And basically you're the Mossic and the elephant, right? You get-- you get, he's a mourn, he gets a elephant to help himself, and, in fact, often only creates business problems. Which is why Mark Zuckerberg says, "We've spent trillions of tokens with you for billions of dollars. Let's spend less." People just start spending the tokens inefficient ways. And your task is to think ahead of you, on the other hand, to understand that you do it for different spheres, including in understanding the work of artificial intelligence, that you will not do anything at once. You don't have this task to programme some kind of service. It's not a question. It's a system you programmed. Maybe you won't need it, maybe you'll all work in this chat-bot in the future, but you'll understand how it works. You have a service and a chat-bot somewhere. And yet you must understand that there are areas you don't know. And when you're not in this field, you lack knowledge of business analysis and business analysis. I think it's a fundamental part of the world, and people will make a lot of money on it, and in the future... They're earning, earning, and they're gonna make a living. Those people who understand business analysis understand how to connect data between themselves not to hear, but to the real ones who live, they are interested and willing to see the large quantities of probability. It's like a man's life, right? The life of every human being, it, uh, is incredibly saturated by causes and consequences, by the number of different events and by the number of different bases. It doesn't mean that all the time you have to live in preliminary calculations of everything that's gonna happen, but you can find different options and different possibilities. You're on ToTheMoon. Sign on our channel. Send this video to your friends and before you meet.