📱

Get Our Mobile App

Take your business learning on the go!

Download on the App StoreGet it on Google Play

ИИ-агенты — главная находка 2026: 4 бизнес-кейса (Claude + Gemini + NotebookLM)

Дмитрий Попов | COMANDOS AI – Элитные AI команды27:19

Transcription

Team. Hello. Today I will show you another four business cases. These are really cool cases that exist in absolutely any business. And it doesn't matter if you have a business or if you are an integrator and will set this up for a business. This is super necessary. Look at what we will cover today. Today we will cover how to ensure that a meeting goes well, and you immediately get a transcription from this meeting, identify all tasks. Distribute tasks to responsible parties, and this is all on full automation. This is the biggest pain for any manager, especially one who doesn't like attending meetings. What will we do second? I will show you how you can very quickly monitor the reputation of your business or something, see how the situation changes, see what is going well for you. What is going badly. If you have a business, especially in different cities, this is super necessary for you. What will I show you next? I will show you how to organize a hiring system that will accurately select vacancies, select candidates, evaluate candidates. And all this will happen on full automation. And then I will show you a very cool case. How you can create incredibly accurate commercial proposals for businesses, and they will trust you just from the commercial proposal because they will feel that you have invested an incredible amount in it, but in reality, you will spend very little time on it. As always, I have prepared an incredibly detailed step-by-step algorithm for you. Everything you need is here. Just take it, copy it, and that's it. We will do all this now. And I was also told that they want it to be like before, an instruction in a format. And I will share all of this with you. So, no more words, let's go. The first thing we will do is settings. I will show you how to do it at the end. Let's move on to the most delicious, the most useful part right away. Look, when you conduct meetings. It doesn't matter. Let's say you are conducting a meeting. You may have many different meetings. They can be of different lengths. Or a very good practice. For example, if you are a manager, you have some systematic coordination meetings, and you don't want to be at them. But you want to know exactly what is happening there. This is for you. You have some video meetings. You can record them on a dictaphone and it doesn't matter how you record them. And the first thing we do. We need to prepare a video transcription. If you conducted it, you can just take the transcription here, download it, and that's it. Here it is. And look what you do. For our work, we will need Claude. Look, the cloud version is important. Because it can work with your computer's file system. And I will show you now. This alone can replace several people in your team or create a separate team. We have a chat here. We have a co-worker here who performs tasks. We have here. Where to get this desktop. Well, everything is very simple, of course. Go to the website. And download it here. What is your operating system? Once you have downloaded it, install this application on your computer, and then you will just need to create a folder that you will work with. I will show you everything now. Look, you just create a folder on your desktop. Well, here it is, I called it for deletion. Look, here is our first case. Here is the meeting. Here I upload some transcription of our meeting that we had. It doesn't matter how long it was. You can make it so that everything is extracted automatically, but that's not what we're talking about now. Look how it works. You just upload the transcription here. What do we do first? We just go here. And here is our first case. Everything is very simple here. Everything is written out for you. It says upload your transcription, when done, mark it as done. And then we have a system prompt here. You can copy it and we will just go and paste it into Claude. Well, before doing that. Look what I have prepared for you for your convenience. You can fill in your data here. Where is your folder located? The folder path, so that the system prompt is immediately here. You see, it substitutes the path and the company name, and everything. And you substitute and create a very convenient control panel for yourself. And before that, what will be needed. Claude also has an extension. We need to install it. You also get it from the website. First authorization, when you install it, you will see synchronize here. You synchronize, and that's it. You will need this tool, and I will show you why we are doing this. We open the first prompt, copy it. We go to our desktop version. We open co-worker. What is co-worker? It's essentially, it automatically launches processes here and performs a lot of work automatically. You can systematize a huge number of things, and here we just select our folder. Our folder, I pasted the script here, essentially, here with one command you can program, with one command you can call. But now, so that you understand how it works. We press it, and it starts working here. That is, it launches processes here itself. It has progress. It extracts all documents itself. Look at the pleasure. It does the work. That is, it analyzes the context, extracts, launches processes itself, calls its correct skills. You see, correct components. You will have all this. You will go through all this. You will be able to repeat it. I will show you at the end. It does this as sequentially as possible, on full automation. Press one button, everything is ready. You can even schedule it by time. Look, it finished working. It generated two reports for us. Look, open the first report. You see, it created a full report. Yes, but let's see what it looks like. Here is the report of our meeting. Imagine, your meeting was two hours long, or your employees conducted coordination without you. You just open it and see key decisions, problems, risks, next steps. You looked. Okay, great. Understood. And what about tasks? Who should do what? And it immediately generated tasks in a table for us. Look what it did. It immediately prepared a table with tasks for us. We can also open it with one click. That is, it appears directly in our folder. Here are our tasks. Here is the task to send the contract, date, deadline, in progress, priority high, started. Look at this pleasure. And all this is configured with one click according to the recommendations I give you. Just go through the guide. How do you like this result? Give a fire emoji if you like it. We will. Let's move on to the next case. The next case. I believe it is super useful for any company that strives to develop. It's a reputation tracker. Look what happens. A company has some reviews, and they are published very quickly. For example, one review is posted, and you need to react to it immediately, either dispute it, or something else, because other people come in. If you have, for example, a coffee shop that is chosen by maps or something similar. They look and think, I won't go. And by the time you get to this review, you can lose a huge number of customers. I will show you how you can analyze all this on full automation. Look, I created this table. We will look at this using the example of a coffee shop. Here we have some dynamics. We collected previous reviews, and we need to see how the picture changes and how to do it. We just take the link to the table you want to fill. We return to our, which will be set up here, to the link to the table. Everything is inserted. Here is our folder inserted. That is, we have made all the settings. We go to our reputation tracker. The first thing we do is just copy our prompt. Indeed, everything is already substituted everywhere. I copy and return to our desktop version. Everything is exactly the same here. We open co-worker, create a new chat. The folder is selected, and I just paste our system prompt and press it, and that's it. And then it will do it itself. It has already created a plan for itself, it is studying the context. Look what is happening. It is now launching the browser itself. It opened it. You see, this little yellow thing appeared. This means it opened the browser. But now, if we look, if you are doing this for the first time, it says, can I launch it? I say yes, you can. You can choose it forever. I say yes, you can. So it launches the browser, and it will start working in it now. It looks, it opens reviews itself. That is, you can automate absolutely any work. Imagine, it wrote a script for itself. It is now conducting an analysis, selecting everything, extracting. You see, it's yellow. All of this. This means it is analyzing reviews. It is now going to reviews. You see, our company. We can have any number of sources on full automation. Press one button, and it flies. Look, here you can see, it created a route for itself. Search and collect reviews, classify reviews, open. That is, it set tasks for itself. Here it performs them. Here you can see, if you are interested, what it is doing. You can always see what it is doing. It is opening, now it is taking screenshots. It wrote for itself, it takes screenshots and extracts data. It is directly visible on the cards, and it will now group everything. And while it is working. Let me introduce myself to those who don't know me. My name is Dmitry Popov. I am 43 years old. I have a loving wife. I have a son. I study a lot. Incredibly a lot. I am part of various communities, clubs. I study a lot and teach. I have been building ordinary, solid, real businesses for 17 years. I built a chain of grocery stores, sold the company, an advertising company. A financial engineering project, a construction holding. We built more than houses here, became leaders in our segment. Such low-rise houses. I love to travel, of course, and I have been professionally consulting for over 5 years, working with businesses, and I calculated that I have worked with businesses, and the results were. Growth up to 3%. That is, I have seriously improved in all of this history, because I know how businesses work. I know how to build a management system, how to build processes. In addition, I am on the board of directors of 5 different businesses. That is, I have incredible experience. And I spent over 2 hours studying artificial intelligence. How to make it start bringing you additional money, save your time, and increase efficiency. Let's see what we got. Well, look at the pleasure. It filled in the table. You see, it added all new data. We can immediately see what is happening here. Processed, processed. Look, it's visible, requires a response. That is, we can immediately make it notify us about this. Super convenient. Well, how do you like this result? Imagine, you just launched it, and it did everything itself. But we are moving on. We have collected reviews. We mark that this is done, and we move on to the next point. It is also completely ready. We just copy it and go back to Claude. I gave it the task to analyze, I say, analyze the reviews in our table, create a new sheet, and it is now planning its actions. It. It reported that the tasks are completed. I also want to show you. When it creates a path for itself here, you can always click the plus sign here and add a comment to improve this or that. It's very convenient. But let's see what it did. Here is our table with reviews. It created a separate tab for us. Remember, we told it to create a separate analysis tab. Look, we go here. And here it creates an analysis for us. Average rating. That is, look, it writes for us, five points for this week. Last week three and six plus and the dynamics are awesome. Look, it says top three problems: rating does not match, price increase. Top three praises, seven reviews praise us. Imagine, you are a manager, or you are setting up such a system for a business where reputation is managed very precisely. That is, he just opens it, looks at such a report, looked, okay. Where are the reviews, which ones? He goes here and immediately understands everything. You can react to reviews lightning fast and thereby greatly boost your business or your client's business. How do you like this approach? If you like it, give it a like, write comments so that as many people as possible see this video, and we move on to the next case, and it will be even cooler. Here we mark it as done, and we move on to the hiring system. This is the biggest pain. The biggest pain is always marketing, sales, and personnel. And to do it very accurately. I will show you an incredible life hack now. Look what we will do. We will connect tools now. We will create a separate notebook in Notes, and I will show you an approach that absolutely no one shows. When you create this notebook, you can upload all your hiring approaches here. History of your company. You can upload many books by specialists here, etc. I have prepared documents. Look, we have documents on our company culture here. We have a small description of how we do things. The more documents about your company there are, the more accurate the result will be. What do we do? We just upload these sources and call the notebook. We have created a database. What's the trick with the notebook? It never lies. It always works with this database. But we will go further now. We will go to Gemini. And I will show you a life hack now. Almost no one knows it, or very few people. Look what we can do. When we get here, we go to chat. We press the plus sign, and here we can use our notebook as sources. Look, I press it. I press, I select. And now all the sources that were there, it takes them for its database, and it will write vacancies with incredible accuracy, understand audience portraits. I will show you how it's done. And if you have multiple notebooks, then you can. Attach another notebook, for example, a transcription of a meeting with candidates. It doesn't matter, but I will remove it now. We only have ours. And look what we do. We return to our dashboard. We mark it as done here. We have prepared. Now we take a ready-made prompt. Look, we take a ready-made prompt, go into Gemini, and paste it. That is, it has a database. And look what this prompt does. It creates a model immediately to correctly identify a candidate. What is the problem with any business? It doesn't know who is needed, what they will do. How to evaluate, whom to choose. Look, a team of experts and HR. Use documents, mode, ask questions, start each position. You position, tasks. That is, it asks us who we need. That is, it interviews us, and at the end, it will give us the result. Look what we do. We charge it, and it starts communicating with us. But. Look, team, what a pleasure. This is the problem with any business when it hires. It doesn't understand what the person will exactly do. The person comes for three months, does something, and after three months we realize they are not suitable for us. The person leaves upset, the business is upset. And all because. When we say hire a person, and the HR says, "Okay, I'll hire," because you need to ask questions. It's inconvenient. But look, we have a hiring system here. It asks us questions. Look, it asks questions. The first question it immediately asked is who do we need, what position? I tell it, we need a customer service manager. It says, "Okay. Next question: what three specific results will you expect? Be precise." We answer. It says, "Okay. Who will the manager report to?" That is, imagine how deeply it digs? "Okay. Why is the role open now? Is someone leaving or what?" We answer it. "Okay. Who will definitely not suit you? Name three red flags." Imagine how accurately you can find people with such an approach. "Okay. Recall the best employee you have ever hired." And it immediately adjusts. That is, it is guided by our company's criteria, by my values. Next. Look, it says in your task. It says, "Well, this is too much for the market." That is, it clarifies. That is, we change the numbers. It says, "All good. Next, it says, 'Now I will create a vacancy for you.'" Look, it creates a vacancy for us, red flags, a matrix. And we will now take this document and upload it to Claude, and it will do what previously took several days for one or two people. But it will do it all in just a few minutes. What do we do next? The first thing we do. We change the salary range. We go to our dashboard. Remember, we have settings at the top for yourself, hiring system. Here we set 75,000. I have substituted the data that we have worked through. Then I go down. And look here in the system prompt. Pay attention. It says here for Claude from Gemini. I just copy all of this. I go back to the desktop version. Here I take the entire response so as not to confuse anything there. I just take it, copy it. Here I pasted the prompt from our instructions. And here I just delete this and paste what Gemini sent us and charge it. Look, it immediately launched the browser itself for work. Yes again, it will now work with HeadHunter. And. It says allow, I say allow. So that it can start working fully, so that it will now go and do it itself. Look, it went to collect vacancies on HeadHunter, evaluate them. But this is pleasure. This is simply incredible. This is simply unreal work. Plus, all this needs to be evaluated, structured, in general, pleasure. Let's see what happens. Well, just look at the work it has done. It analyzed the Excel file, uploaded 20 resumes, found as my my said, and I want to draw your attention. Look, 566 resumes were found. Imagine the amount of work. How much an HR or a recruitment team would have done. Let's see what it did for us. It made. Here you see the table. Let's go into it and see what happened. But isn't this delight? This is simply incredible. Look at this pleasure, what a table it made. That is, it found all the managers for us, wrote down all the salaries. And the last place. You see, it made us a heat map. And separately, it made us a shortlist. Look. Whom we choose, 77 points, 75, it immediately evaluated, evaluated according to our criteria, yes, imagine. When we answered it, it understood who we needed, how to evaluate them, and look how it distributed everything for us. We can immediately ask for a link from it. But how do you like it? Imagine if the process in your business or in your clients' businesses looked like this. How much time do you think this can save? If you like it, give it a like, write comments so that more people see it. Of course, I will share all instructions and settings. I will now tell you how to set all this up. Here we mark it as done. And we move on to a very cool, very awesome case. Commercial proposal. Many businesses make commercial proposals in a maximally templated way. Whatever happened, happened. But I will show you a completely different approach now. I will show you how to generate commercial proposals using the example of companies that produce workwear. I chose Techna. Simply because I think they are awesome. Imagine that a business is involved in workwear and needs to send a proposal to another business, and it must be one that they cannot refuse. And look at our algorithm. The first thing we need to do is to have our products in our folder. Look at our products. I have already prepared our products here, just the products we sell. And let's say we want to sell these products to some company. Let's say we want to sell them to Gazpromneft. And look, whether your commercial proposal will be successful or not depends on what you know about the client. And the first thing we need to do is conduct reconnaissance on the client. What are their strengths and weaknesses? What do they need? Where do they work? And I will show you how, based on this, you can create very cool commercial proposals. What do you do? Well, of course. Here, in our preparation, you can prepare your data and leave the existing data. Then I go here and just copy our first prompt and return to our Claude. Everything is classic. We go back to Co-worker, upload our system prompt. What does it say? Find information about Gazprom company. Sources: 2GIS, HeadHunter, Yandex. Because sometimes employees can badmouth and say, "I worked there, the clothes were constantly cold, bad, etc." And this is a reason to make a powerful commercial proposal. Yes, everything is substituted here, everything is as simple as possible. Copy. Paste, charge. Then you will be able to do all this with one click for full automation. We are now creating an incredible commercial proposal. I will show you how not just to make it beautiful from a PR perspective, but how to make the pictures attractive. Let's see. Look, it conducted a deep analysis. Studied everything. Looked at the company. Imagine, a client comes to you and says, "Do you have winter jackets for builders?" You say yes. You just upload the name of their company. It collects all the information about this company. And you prepare a commercial proposal that they cannot refuse. Look what it prepared for us. It prepared a full reconnaissance file for us. Well, of course, it's Gazprom. What they do, where they work. What is the staff size, how many people? What kind of company, working conditions, according to which they work. Digital transformation, industry pains, sites. That is, it completely. Look at what it collected. It did an incredible job. That is, for you to collect such information about each of your clients and make an individual commercial proposal. You would never have done this because it's too much work. Yes, but here with one click, just think about it. Well, we are moving on. It will be even more interesting. We return to our prompt. What do we do? We copy the system prompt. Next. That is, the previous reconnaissance is done. Let's mark it for order. Next, image generation. We copy this system prompt and return. And we just upload it here. And look at this point. For Claude itself not to be able to draw beautiful pictures. And we will use nano-banana for this. And you will see what we get. You saw, we uploaded the picture of our products. And we also uploaded the logo of our client. Yes, it's very fast. We could have told it to extract it itself. It would have extracted it itself. And now look what we tell it. Take three scenes from the report, realistic corporate style, so that our employee. Let's see what comes out of this. Well, look, it finished working, created three pictures for us. Let's see. I'm doing all this in real-time with you. Let's look at the pictures. So, this is a small size. Let's make it bigger. Look, it took our jacket, took the Gazprom logo. Here it is a bit blurry, and the prompt needed to be worked on, but in any case, we are doing this in real-time. That is, look, it recreated a real site, recreated the clothing, applied the logo, says, "Here you buy from us. This is how it will look." Just imagine if the client receives such a proposal. Yes, let's see what else it did. So it also made it like a foreman. You see, our jackets are hanging, beautiful. That is, we can show how our clothes look in any conditions, what kind of Gazprom it is. And here is an employee sitting in our jacket. That is, look how it integrated all this into the infrastructure, into the interface. Well, what pleasure. And based on this, you can create a truly cool commercial proposal. That is, look, to generate these pictures. We had a connector set up here. It's called, you see, nano-banana. Here is the file system, to change folders on the computer and nano-banana. Yes, I will also show you how to set this up. Let's move on, generate a commercial proposal. We mark it as done here, and we move on to the next step. This is the final commercial proposal. Everything is simple, everything is super simple. Copy, return. Paste, send. This is how regular processes, regular management, can be built now. All this can be maximally automated. What previously required a team, many people. This is incredible. Don't play with artificial intelligence like a chatbot or just drawing pictures. If you draw pictures, make super accurate commercial proposals, some offers. It should earn you money, save you time, and increase efficiency. If not, then you are doing the wrong thing. All the instructions I am talking about now will be completely free under this video. Don't try to copy anything. You will get it. I will now show you how to set up this entire apparatus for yourself. So, team. Well, look, it has finished working. Look, it says structure by pages, cover, preview. But let's see what happened. Open. But look. 83,000 employees. Workwear that works in -50. Picture. I did nothing at all. I just wrote. It collected everything itself. We know your specifics. We know where you work, shift work. How many employees you have. What this means, you need workwear that can withstand extreme cold. Our solution: Technaavia workwear. Just imagine if you send such a commercial proposal to each of your clients, where you integrate things into their business, and they can model it all. Imagine the response you will get. It can be done in your style, in your brand. How do you like the results? If you like it, give it likes, write comments, and we will take one more strong step now. We return here, we will do the most difficult thing now. What we have is we put a checkmark here. Everything is there. Put it. Now what do we do? Just copy. Return and paste. Look, we are saying, read the commercial proposal that we sent. Check how well it corresponds, how well it will resonate. Personalization. Is there a feeling of a template? At least two facts about Gazprom. That is, we are doing an additional verification step, that this commercial proposal will reach where it needs to reach. And we send it. And this is how all systematic work looks. Imagine, previously several employees. Who, if you are in such a business, or you know someone who made individual commercial proposals for businesses? Almost no one. They put a template, pictures. That's how it looks. But to analyze a business like this, to immerse yourself in it, to collect information, to generate pictures based on that, to integrate into the business's life, and to make a proposal. Just imagine, if he sees such a commercial proposal from you, he will think, damn, if he put so much effort into the commercial proposal, what will happen next? You can simply tear up the entire market. Look, it conducted an analysis, personalization, yes-yes-yes-yes, concreteness, language, design, next step, result: 14 out of 14. No corrections had to be made. Everything. Now you can take this commercial proposal and send it. Well, team? How do you like the result? Give it likes. Write comments if you like it. Well, look at the pleasure. I am simply delighted. So. How to set up this whole devil's machine? Look what I have prepared for you. First is our interactive dashboard. We have completed everything here. Put a checkmark. Everything is fire, ready, done. Take it and use it. Here you can substitute your own data and work with it. There are already commands written here, but since you asked to do it in the old familiar format, I have prepared this format for you for the old-timers, say, take it and use it. It's the same as in the interactive dashboard. Here at the beginning, it also talks about settings, how to do it. But I thought that perhaps you need it to be even more detailed. And I have also made another guide for you. You will get all of this for free under this video. How to set up everything so that your Claude works on your computer, can change folders, create pictures, generate commercial proposals. That is, how to give it all these tools. Yes, look, I have written everything here for you. What to download, where to go, what to get. Check how it works. I have written all the codes that need to be entered somewhere, commands to be entered. What needs to be replaced with your name, I have also written. But if this is difficult for you, what do you do? You just click. If you have, go to Claude's chat, paste it here. And say, "Help me set this up. Guide me step by step." And it will guide you step by step. But that's all for today. We have covered four cool business cases. Thank you for your attention. Take everything under this video, absolutely free. If you like this, and want to improve even more deeply, there is a lot of useful information in my Telegram channel. I am waiting for you.