It's Here! The Second Part of the AI in E-commerce Podcast with Bartek Bilicki.
We are pleased to announce that the second part of the podcast with Bartek Bilicki, dedicated to the use of artificial intelligence in e-commerce, is now available on our "Przelewy24_Official" channel. In this episode, Patryk Harasimowicz and Bartek Bilicki continue their fascinating conversation, delving into the practical applications of AI, tools for workflow automation, and the future of the industry in the face of dynamic technological development.
What will you hear in the new episode?
Bartek Bilicki shares his insights on where artificial intelligence will be most commonly used in 2025. He emphasizes that even fields like visual art, which seemed resistant to automation, are already being transformed by AI. AI agents can perform both simple and more complex tasks, such as sending a drone.
We will learn which tools will help automate processes in e-commerce, especially in the context of product segmentation. Bartek points to the OpenAI and Google ecosystems, which are developing towards assistant-agents. Tools such as internet search (Surf), calculation tools in AI models, and internal ChatGPT tools like Canvas, which facilitate working with text and code by showing changes in a more readable way, are discussed.
The world of Gemini also offers many possibilities, allowing for conversation with the chat to obtain information or create documents after appropriate permissions are granted. OpenAI, in turn, has hidden functions, e.g., the ability to connect a desktop application with Notion, which allows ChatGPT to access current information from an active page.
For those who want to build more advanced solutions, Bartek recommends tools such as Zapier.com – the oldest solution for integration between various applications. Simple automations, such as publishing posts on WhatsApp after a new blog entry appears, can be enhanced with AI that shortens the message content.
Another discussed tool is Make.com, which visually facilitates the creation of diagrams and integrations, as well as N8N (n8n.com), which is a compromise between capabilities and ease of use. N8N allows for building automations using AI, and can also be run on its own server, which is important for companies with high compliance requirements.
The podcast also touches on the application of AI in Smart Homes and emphasizes that artificial intelligence is moving beyond text, connecting with other devices and services.
AI in e-commerce practice
Bartek answers Patryk's question about whether AI can help in adding thousands of products to a store, assuring that it is possible – AI can process various data (PDFs, scans, photos, CSV files from Excel, links to Airtable), and even suggest popular products based on Amazon's best practices. The entire process is supervised by a human, in accordance with the Human-in-the-Loop methodology, which allows for controlling the AI's actions at key stages.
AI can also become a personal assistant, accelerating daily work, e.g., creating emails based on voice commands. In the context of e-commerce, AI is invaluable in generating product ideas, especially in creative industries, e.g., in florists.
In marketing, AI can support reactive marketing, based on various data sources (blogs, portals, Google Trends, Google Alerts) and automatically generating social media posts, including graphics and infographics. Bartek mentions the lower costs of generating visual content using AI compared to hiring graphic designers, while emphasizing that graphic designers (as well as lawyers and programmers) are still needed.
AI also enables animating photos, creating three-dimensional product visualizations (e.g., using PhotoRoom), and generating audio content, which speeds up and facilitates the creation of marketing campaigns.
AI in everyday tools and legal issues
Many popular applications, such as Canva, Excel, and Word, are already infused with AI, making it easier to generate graphics, create formulas, and analyze data. Google, with its numerous applications (Gmail, search engine, calendar, Docs, Spreadsheets, flight search, shopping aggregators), integrates AI in many areas. Google Analytics also has AI-powered analytical solutions, facilitating data segmentation. Microsoft's CEO rightly noted that all interfaces are unifying into a single conversational one, changing the way we interact with technology.
The important issue of copyright in the context of AI-generated content is also raised. Bartek emphasizes that the problem is significant because many AI models have been "trained" on human creativity, leading to accusations of piracy. He indicates that there are two approaches in this field – one that assumes that copyright to generated content does not belong to the user, and another, more lenient one, allowing the use of generated content with a possible mention of the AI model used.
Call to action
Bartek Bilicki encourages listeners not to be afraid to enter the world of AI, regardless of their field. He emphasizes that all tools are becoming easier to use. As an example, he presents his personal health project in ChatGPT, where, thanks to constant interaction with the chat in a dedicated folder, he receives personalized information and recommendations. He encourages closing areas of professional or private life in such "projects" in the chat to build context and receive increasingly better answers.
Where to seek knowledge?
Bartek recommends using academies and educational materials offered by AI model and application creators. He also invites you to his website evernex.pl, where he conducts trainings and courses focusing on AI agents and providing a comprehensive overview of what is happening in this dynamic area.
Artificial intelligence presents challenges, but also unlimited possibilities. Let's not be afraid to try and use its potential.
Full transcription of the conversation
Conversation with Bartek Bilicki about artificial intelligence in e-commerce.
Patryk Harasimowicz: Where will artificial intelligence be most commonly used in 2025?
Bartek Bilicki: I discovered that art, visual art, is the most certain and stable for me. I founded a visual art company with the conviction that AI would not imitate it. However, my thesis was overturned. Now an AI agent can not only add simple things and perform tasks, but also more complex ones, such as sending a drone.
Patryk Harasimowicz: Bartek, regarding the segmentation of all these products, thank you for explaining how we can achieve this. However, could you advise which tools will help us automate this workflow?
Bartek Bilicki: I've already mentioned that the entire OpenAI and Google ecosystems are moving towards assistant-agents, so we'll be seeing more and more of these tools. A simple tool, for example, is searching for information on the internet. OpenAI was trained on data up to 2021. It was funny to ask, for example, who the current president is, and ChatGPT would stumble on that. That's why another tool, internet search, or Surf, was added. In addition, new models began to appear. It turned out that the AI model couldn't handle mathematics, so in some models, a tool was created that uses specific functions for calculations, instead of relying on model hallucinations. ChatGPT itself is moving in this direction. For example, it gained an internal tool, Canvas. When working with text or code, previously each subsequent version of the text would appear below. On Canvas within ChatGPT, we see in real-time that only fragments are changed, which is more readable. That's a cool, next tool. There could be more such tools. Entering the world of Gemini, there are many of them. It's fantastic that I can talk to the chat and ask it, for example: "Tell me what emails I received today." By giving it the appropriate permissions, I will receive such information, or I will create a new document. This is the direction everything will go. Returning to OpenAI, it has a lot of interesting, somewhat hidden functions. If someone works on a Mac, they can connect the desktop application with Notion or Notes. Then we launch the Notion page that interests us. We can have current information about products, questions, answers there. Then ChatGPT will connect to that one active page, not to the entire Notion, and we will get more information. These are glimpses of this agent-assistant approach. But if someone really wants to build this, we enter a different world. We enter the world of tools that have been doing this for years, only AI has been added over time. I mean solutions like Zapier.com, which is the oldest solution in this family, it allows for integration between different solutions. It's a very cool tool. For example, by examples, because I know it's a certain abstraction. We can connect and create an automation where if a new blog post appears, it should be displayed on WhatsApp. This is a simple automation. There's no artificial intelligence here at all, but we can insert artificial intelligence in the middle, so that when the new post appears, it shortens it for us on WhatsApp. And that's the difference. Similarly, in this formula, but more visually and nicely, it's easier to create diagrams, the Make.com application works, which for the last few years has been the king of this segment. Make.com dictated the terms, it has the most integrations, it took over companies that also dealt with integrations. Previously it was called Integrator. It's quite easy to learn. If a new record appears in our Google Spreadsheet, then create a new post on LinkedIn. You can do something like that. And again, we put artificial intelligence in the middle, which, based on one line of text, creates a series of posts on LinkedIn for us, and even generates graphics. Wow! But finally, I left the solution that for me is the best, most useful, it's N8N, which is n8n.com. This solution, I believe, is a good compromise between possibilities and ease of use. There is one big block there, because it's all based on certain blocks, nodes, blocks. I call them blocks. A block called Agent AI simply has different inputs. Sometimes I say, these little hands. For example, we can connect to this Agent AI via Telegram, and the output responds on WhatsApp or sends an email, and along the way it also checks our availability in the calendar or calculates some complex integrals, but that's incidentally with Wolfram Alpha, a dedicated application that typically deals with mathematics. In short, today AI agents are mostly clicked, not coded. Of course, there are also dedicated frameworks, tools for AI programmers, such as LangChain, a fantastic solution. But if we want to do it quickly, but also bulletproof, N8N will really work in most cases. I also want to say that it's good for companies that have compliance departments, which care very much, more than in small companies, about data issues, and have additional requirements. Here, as far as N8N is concerned, you can run it on your own server, in your own server room, and rely on models that comply with compliance guidelines. So we could connect what we need here and what requirements we have, so N8N is fully friendly for both small and large companies. The costs are also low, so I think it's worth trying and checking these automations yourself. I use automation and agents interchangeably, because in essence, this world today intertwines. The essence is similar, we simply arrange blocks, indicating where artificial intelligence should appear. Sometimes these are short workflows, automations, and sometimes they are very long and complicated. As I said, I could talk about this longer, but maybe I'll add one more thing, it's very important, because someone might ask: "So, does it always work like this, that I have to talk to it?". Not necessarily. It could be that we connect, for example, a smart home, and for example, a sensor, I don't know, smoke, light, any sensor, we connect it to such an agent, and based on something that happened, for example, it calls us, or sends an email, or even calls the fire department. Anything. Simply, the world of artificial intelligence has begun to go beyond text. It has started to connect with what is within our reach, what we already used, what only has internet access.
Patryk Harasimowicz: I understand that this is a type of application that has a graphical interface, which allows us to modularly model the workflow of needs that we ultimately want to achieve. Is there a need to acquire programming skills for applications like N8N, although on the other hand, we can support ourselves here with artificial intelligence, right?
Bartek Bilicki: Programmers still, I say still somewhat sarcastically, programmers still have a big advantage. Logical thinking, a kind of structural thinking, the ability to translate a certain vague, general concept into processes. So programmers will do best here. That's the truth. It's not like someone who has never designed or modeled such things will immediately do great, no. You have to learn it, but you don't have to learn complicated commands, and that's the point here. That's the difference. So you can definitely go in this direction even with little experience. And for example, Make.com, which is not yet visible in N8N, but Make.com has an internal assistant, just like in other applications, which can click these blocks for us. So we are simply going even further into this artificial intelligence. It's already leading us by the hand so much that we'll soon lose the ability to walk. So yes, everything is getting easier and easier. Make.com is also a good solution, but Make.com doesn't have such a nice plan that we run it on our own server, which is very, very important for many businesses.
Patryk Harasimowicz: Full agreement. Bartek, you've already started presenting the potential applications in business processes. Our listeners are e-commerce fanatics, and perhaps, since I have you here, let's help them use artificial intelligence in their business. If I were a seller and faced such a challenge: I have 1000 products to add. Can the use of artificial intelligence, an AI agent for example, help me with this?
Bartek Bilicki: Absolutely. The beautiful thing is that each of us has had moments in life where we had to manually edit countless files, countless records in a database or an Excel spreadsheet. So we start by not needing a standardized database from the very beginning. So the answer to your question is probably multi-faceted, and everyone has a different approach to it, but what I notice right from the start is that, for example, I can throw, I don't know, five different PDFs into artificial intelligence, two of which are scans. Add four photos. Add 500 records generated in CSV from Excel, and also link Airtable, for example, from which I have some product ideas. And then also write to the artificial intelligence: "And generally, let's also come up with what popular products are now, based on Amazon's best practices." For example, and artificial intelligence won't complain, it won't say: "But what data did I get here, I have a mess, I don't know what to do with it." No, it will simply generate everything, organize it, probably make some mistakes, but it will do it in a flash, it will do it in an instant. And what we would normally shovel, so to speak, on this computer, moving files, checking here, analyzing, automating, artificial intelligence does immediately. And then we are essentially just supervisors of this process. So here it's also worth mentioning a very popular methodology of working with artificial intelligence, which is called Human-in-the-Loop, meaning that we generally delegate certain things to artificial intelligence, but we still supervise the process. And we are at certain stages. This can be done in a simple way. For example, we have an agent, I won't go into how to create this agent for now, but let's assume we have this agent, who is supposed to send emails. And we are afraid that he will send these emails incorrectly, or to the wrong person, or to some, I don't know, president's friend, he will send the wrong prices, whatever, something we'll have to undo. So we create, we place such a guardian in the form of ourselves, or in the form of an employee, who just has to confirm that it's a good email and click. I've seen such applications. It's also cool that someone, and this also supports sales in a peculiar way, that I found this application on LinkedIn. Some company CEO was saying that if he's driving and someone tells him something that needs to end with an email, he just starts his agent, who says: "Send an email, write an email about something to such and such a person." And the chat will do it in a formal, informal style, in Polish, English, Chinese, however we want. And at the end, he either listens to it again in the car, or just keeps it in drafts, and when he's at the location, he gets out of the car, then he calmly checks it and sends it. It speeds up work incredibly. The agent becomes a personal assistant who is always and everywhere with us. So why not use it without it? But of course, it requires some general configuration. Returning to e-commerce and those thousand products, just adding, inventing, because some industries are based on ready-made products. For example, it would be stupid for the chat to invent a new iPhone that doesn't exist yet and for us to list it – that's pointless. But if we have a flower shop and we lack ideas on how to arrange these flowers, we give it information that we have roses, violets, some hyacinths, and we lack ideas on how to arrange them. And this works perfectly in such a generative collaboration. So suddenly, until now we had 10 products, and now we have 1000, well, why not? So at the very beginning. Then marketing itself, well, we could probably talk about marketing for the rest of this podcast, because a lot is happening here. AI can help us in reactive marketing. We can indicate which sources are interesting to us. These sources can be typical blog sources, some portals. These can be insights, we can be sensitive to Google Trends, Google Alerts. All these things that we knew, checked, and relied on so far. Similarly, Agent AI can also rely on this. And based on these countless sources, at some point it can trigger itself and write a post that automatically publishes, or again Human-in-the-Loop. So it turns out that, for example, we ourselves post twice a week, and AI additionally adds, I don't know, four or five times, and we have such a nice spectrum of different posts. Graphics are also generated incidentally, if we need an infographic – fine, no problem. Incidentally, I will say that OpenAI handles infographics better in English than in Polish, but it still does a really great job, better than many a graphic designer. Unfortunately, with a heavy heart I say this, I would also prefer to outsource work to physical employees, but the truth is that if I have something in three seconds, and practically for free, you could say, because the costs of these tools are usually, we joke in the industry, that it's always 20 dollars. Yes, because usually it's 20 dollars or 20 euros, or 15 dollars, or 25, but more or less that's the monthly cost. Well, let's compare that to the cost of, I don't know, a graphic designer. And I still know that graphic designers are needed. I want to emphasize here, my brother is a graphic designer, fantastic thing, he also knows that lawyers also have a future, and programmers too. But the point is that today the entry threshold is much higher, and this can also lead to a certain drama in the labor market. Because if today it's hard to become a junior, because junior work is done by AI, then who will become the mids, the seniors in the future? Returning to e-commerce, there's a lot of work that can be done in marketing. We can animate photos. After all, photos sell our products most often. We can make it so that, I don't know, a book suddenly becomes three-dimensional, we rotate it, or there's a product, PhotoRoom, which is called PhotoRoom, and this product generates very interesting things for us, these are no longer mockups, these are already generated photos, actually generated visualizations of various products based on people, based on, I don't know, some elements that we would normally have to go and hire. For example, we need a Harley Davidson. Until now, we had to do it, go, drive somewhere. And now, of course, we could also, say, hire Photoshop for that, but now we have to know Photoshop. We only need to know how to properly put it into words, and it will probably do the job. So here, precisely in generating visual and audio content. I myself generated, organized such a small event, a hackathon, an event for programmers, where we focus on solving some challenges. And I edited the video myself in CapCut, but I generated the sound. The advertisement itself, well, it wasn't a radio hit, but the music was catchy, a bit funny, a bit cheesy, but it went viral. If I were to commission a new song today to some band that has to do it by 10 PM, because I need it for today, it's impossible, I simply wouldn't do it. So this marketing, in short, has so many colors, so many possibilities. And then we go into further, further branches of sales, where I don't even know where to start, really. Because, for example, even such a large part of e-commerce as inventory management, based on, I don't know, some logistical movements, whatever, well, artificial intelligence is also here, but I assume we are closer to small shops. So, if we want to generate content on a blog, just like on social media, we can do it similarly. It's worth mentioning one more thing here, a simple thing, that today we are looking for dedicated applications, and indeed we will find plenty of them, but also applications that we have used so far, they are permeated with AI today. Most applications that are popular in some way, I mean simple solutions like Canva, a company for generating interfaces, for creating, they are permeated with AI today. They create ready-made graphics for us, ready-made visualizations. Even if we knew how to operate something so far, but we used to do it ourselves from A to Z, we had to sit down, think about how to design it, write it. Excel is also an example. Word in any edition, whether Microsoft, Google, or whatever. Today we already have an internal assistant in these applications, who simply does certain things for us. We don't have to write formulas, unless we really had to have full control over every comma in that Excel formula. Generating charts – everyone can do it, but still, when we have to do it, it's a frequent question which one to choose, where, where from these data, whether they are really well formatted. AI does it in a flash, so it's worth paying attention to such artificial intelligence that is already inside applications. It's not a chat that works rather externally, that advises, but it's inside various applications. And here, the dark one is the king of this type of solution, because Google itself has 1000 different applications. Some of them no longer exist, some exist. But if it has 1000 applications, most of us know Gmail, Google search, but there is also a calendar, there is also Docs, as I said, all these Spreadsheets. But we also have a flight search engine there, we have shopping aggregators. When we start analyzing what this Google has in its stable at all and where this AI already is, it turns out that, well, maybe not ChatGPT from OpenAI, but maybe it's better to use Gemini. So I personally use several, because every week a different solution steals my heart. So I know that we are constantly revolving around e-commerce, but e-commerce is a very broad topic, and it's worth saying directly here that e-commerce in every aspect can have AI in such operational, executive, but also strategic ways. If we have to make a decision, Google Analytics also has strong solutions for analytics and segmentation based on artificial intelligence today. And once we had to have such knowledge on how to click this difficult solution, however administrative, where to go, what to do, how to configure. And now more and more often we have an AI assistant on the right side, in which we write: "I need to extract a segment from my Google Analytics, help me do it." And it usually already does it, or shows the path, or will certainly do it in the near future. So it's true what the Microsoft CEO said, that all interfaces are unifying into one conversational one. And yes, we will display various things there, charts, some graphs, some diagrams, various visualizations, we will, I don't know, have a live camera view, whatever, but the main interface will no longer be file, edit, something. It will just be: "Write what's on your mind."
Patryk Harasimowicz: It sounds terrifying, but positive because of its meaning above all. In our last episode, we had Tomek Ronowicz from Mohito, and we were wondering about artificial intelligence in e-commerce. We assumed that first of all, you need to take a few steps back to understand the data, and this data helps us interpret it in various ways, but above all, to boost the possibilities it offers. Bartek, a question in the context of application for our listeners who run, as you mentioned, also small sales stores, because we also see these trends in the application of artificial intelligence, for example, on temporary platforms, where from the moment the first click to set up a store begins, there is already an operator who allows us to set up the store that we would like to return to the Internet to start selling. You mentioned graphics, PhotoRoom, right, which allows us to change pictures, for example, to adapt a T-shirt, maybe put a hat on someone. Can such applications, when it comes to using some specific frameworks, be used quickly?
Bartek Bilicki: Gemini is the best here, for now. Rather, OpenAI is great at generating or processing graphics. I have family photos, we want to convert them into some cartoon-like ones and hang a picture, it will do it perfectly. Really fun. Coloring books for children. I'm giving more such examples from my family garden here, but also infographics, we have these things. However, Google Gemini gives more possibilities to showcase specific products in a different configuration than we would have done ourselves. Or this example, and there's always a delicate "but," that it's not always for everyone, and especially if we want to do it ourselves. If, for example, we have one T-shirt that we want to expose in some way, dedicate time to it, then I think Gemini works great here. But if we were to maintain such consistency across the entire brand and so on, then it would be difficult. This function in Gemini is a difficult topic. I remember generating a jacket, yes, a white, puffy jacket somewhere on LinkedIn, you can see it there. The only thing is that I uploaded it, it looks cool, interesting, because it's something new, it's a good example, but it's also a good example of how AI didn't cope with one issue. Suddenly it turned out that the jacket had a hood, but originally it didn't. So this is the visual side of this hallucination, right? That it decided it would be nice if it had one, right? But such a product didn't exist.
Patryk Harasimowicz: So I understand that we still have to wait a bit to achieve the ideal that we would actually like to see in reality. But there is already potential.
Bartek Bilicki: Yes, I am a fan of these general ecosystems and you could say I specialize in them, but there are dedicated applications that are already appropriately prompted, appropriately automated. Certain parameters are checked there and they work a bit better then. And such a solution is, for example, PhotoRoom. This is a commercial one for such mockups, but generated by AI. There are more solutions and they are constantly either appearing or collapsing, so it's such a hustle, such an interesting moment on the market. But it's also worth mentioning Midjourney, which is the model that first conquered the world when it comes to AI in the visual world. I remember I was running a company that dealt with art then, and I founded this company with the thought, I had a hunch, that AI would enter the market very strongly here and that it would sweep away many sectors. And I decided that in that case, what was most certain and most stable for me was art. It's visual art, AI certainly won't fake that. And in essence, my thesis was simply overturned, without even fighting, it was simply overturned. I did tests myself with graphics, with art. I remember generating Marilyn Monroe in the style of Tamara Łempicka, a wonderful Polish artist. She, of course, doesn't have such a painting, but those people, and I also uploaded it to LinkedIn. And those people who studied art history, of course, had a full understanding that it wasn't it, that there were some nuances, but people who knew Tamara Łempicka's style, but not by heart, were fooled by it and thought that, damn, it was already generated. And this is also an interesting point. I'll add here, I know it's a bit of a digression, but these digressions are only here and now. This led many lawyers to wonder, in the context of art: should style be subject to legal protection? The style itself, the line, the proper approach that we feel when looking at the works of various artists, yes, we can recognize the works of, I don't know, Van Gogh. They are different, but the style is similar, and perhaps it should then be protected in this era. So here we have an impact on every single field. But Midjourney is worth mentioning because a new version just came out about two weeks ago. Two weeks ago, as we are speaking, a new version, which has a "consistency" option, and this consistency is another so-called killer feature. Because what's the point of generating cool things if each one was different and didn't match? Consistency is very, very important to keep it within our branding. Whether this is the consistency that marketing departments would like to have? Time will tell, but it is clear that companies behind artificial intelligence are already working on this problem. So I would recommend Midjourney for such marketing fun.
Patryk Harasimowicz: Oh, Bartek, you touched upon a crucial element, reaching into the letter of the law. Of course, we wouldn't have enough airtime to cover that topic. But how does it look currently in 2025 regarding copyrights and the use of generated content and images?
Bartek Bilicki: In principle, yes. Firstly, I am not a lawyer, so it would certainly be worthwhile to obtain information directly from the source. But there are, I feel, two approaches in the market. One is the approach that there is no, that we can play, but the copyrights will not belong to us anyway. There is a certain conundrum here, difficult, hard. And the second, more relaxed approach, on the principle: generated, we can use it, eventually mention that it was generated perhaps by a specific model, and that's it. So here, from a practical point of view, but this is absolutely not a legal opinion and it would be worth familiarizing oneself directly. But the problem is huge, because in fact all these models, maybe not all, maybe some not, but most of those we use today, were created through piracy, through scanning, processing the sum of human creation. There is the broadest definition of culture, that it is the sum, the general creation of man. Yes, culture. So I feel that AI has simply swallowed culture. And now it's being smart about various topics, but these are often our contents. There were many, many situations where companies that create models were accused of authors recognizing their content, or a specific text, or specific fragments in generated images. So, it's worth mentioning that here creators – this is a threat to creators. What to do in such situations? Well, a lawsuit, and fight David against Goliath.
Patryk Harasimowicz: Yes, that's true. First of all, this AI agency, it's primarily opportunities, but also a threat and copyright, we may not discuss this topic. However, let's slowly move towards the end of this wonderful episode, in which we share this practical knowledge thanks to you. What would you recommend to our listeners, regarding not oversleeping this time, in which area, regarding sales, should they start immediate actions? Is it analysis, or is it simply graphic modeling, in these frameworks you mentioned, which area is it?
Bartek Bilicki: It's any development in this area. If someone is a graphic designer, they should get into AI and graphics. As I mentioned, AI is already in all tools today. For example, Photoshop has AI-powered brushes. We can draw some mountain, and it will be professionally redrawn in a moment. So there's no need to be afraid of it. Those who were not afraid first are today pioneers in combining something with AI, whether it's music with AI, or graphics with AI. And those who thought it was, I don't know, an attack on humanity, on our thinking, and were a bit afraid of this topic, well, they still have to catch up. That's the truth, because that's what the market demands. So whatever we need, I would use it. It's definitely good to thoroughly learn the tools that we think we use brilliantly, even ChatGPT, and I'll tell you what I'm thinking about. I have a project, a private health project. I created a project, that is, a folder in ChatGPT. Folders are not in the free version, only in the one for 20 dollars a month, if anyone is looking for it. There is an even more expensive package for 200 dollars, if someone really wanted it, but I think the 20-dollar one does the job. And I created such a folder. In fact, I have a dozen different folders that describe some aspect of my life, some project, company, whatever. And one is health, it's very interesting, because I think it's very universal, as various people listen to us. And now, since this is my data, I didn't have to go to the management board and ask if I could put this medical data there, and I recommend everyone to think about it. However, I shared my medical data within this project, thanks to which every day, by journaling, having an honest conversation with the chat, I get information about my daily rhythm, energy drops, caloric intake, any infections, whatever. I immediately enter it into the chat. And now I am aware that I cannot recommend this, because it would be medical advice, but I just want to say that I closed off some area of my life and it was totally transformed personally for me by consulting with something every day. My wife is a dietitian, by the way, so I have Human-in-the-Loop here, but I started drinking various types of hemp oil, flaxseed oil, and so on, simply pro-health products, I gave up milk in my coffee. So I'm talking about very mundane things that I believe were beneficial, but they wouldn't have worked if I had just been chatting with the chat in broken conversations, as most people usually do. And why am I saying this? It is very important that we enclose some fragment of our, I don't know, professional life, whatever we want to improve, in such a folder. We should appropriately set certain instructions, certain, certain kinds of. Give some personality to this project, upload the data that concerns this aspect of life. For example, I don't know, our company, or whatever. Of course, I know, legal issues, GDPR, and so on, I leave to compliance. I'm talking here strictly about using tools. And then it turns out that I don't have to constantly re-onboard the chat, give it that context. It learns every day and gives me better and better answers. It gives me new, new, new insights. And now let's think, let's say this: the chat is imperfect, it makes mistakes, but in the context of health, I'll say, because it's most visible. If I wanted to improve something in my health, I would have to have contact with a specialized doctor and a dietitian every morning and evening and exchange everything on an ongoing basis, and here I get full feedback. So once again. I am crazy and generally this is my personal experiment. I do not condone, I do not recommend, but I show it as an example of something that totally transforms our cooperation with the chat, not some other chats taken out of context, but let's close some area in a project and then we will get much better feedback. And I recommend doing this in whatever aspect of life we have come up with. But I definitely recommend doing it, and besides, I am aware, it's a complete chaos. So I think there's no need to worry about not getting everything, because even I, fully engaged in this, don't get everything either, and it annoys me too. But, but if we have, if we maintain such childlike curiosity, I think we'll be on this wave. So it's worth taking this topic seriously, because this topic really concerns everyone.
Patryk Harasimowicz: I fully agree with you 100 percent. Bartek, you're right. First of all, to sum up, let's not be afraid to try, let's not be afraid to use artificial intelligence, because as Bartek's last example showed, a private personal diagnostician can be truly at your fingertips and be with you 24 hours a day. Bartek, before we move on to the summary, thank you again for the opportunity to talk to you, an outstanding specialist. In fact, because we covered a wide range of possibilities for applying intelligence in business, primarily for our listeners from the e-commerce category. However, if you were to say where they can draw inspiration, where they can draw knowledge about all these elements you shared with us today, do you know what those websites are?
Bartek Bilicki: We live in such a fantastic moment today that we have access to virtually all knowledge at our fingertips, and often this knowledge is really well-prepared. So, in essence, each of these ecosystems, each of these applications we talked about, these models, often have their own academies. So I definitely refer you there. It's always up to date. I also conduct my own trainings, my own courses on the evernex.pl website. And there I focus primarily on AI agents, because this topic, I believe, is inexhaustible on the Polish market. So if someone wanted to have a full overview of what's happening in this context, I definitely gather and structure it on evernex.pl. Yes, so I think we'll provide links in the description, and I encourage you to participate.
Patryk Harasimowicz: We also encourage you to visit evernex.pl, where you will find a wealth of amazing information regarding the application of artificial intelligence and everything else you shared with our listeners today. So all information can be found in the link below the video you are currently watching. For me personally, artificial intelligence is primarily challenges, but especially opportunities that show an unlimited horizon, where in fact, we ourselves, in terms of our creativity, can set goals, but thanks to artificial intelligence, we can also overcome these horizons. Dear listeners, all links to all materials can be found in our description below this current video. We cordially invite you to the next episode of "S.A. E-commerce" on our "Przez 24" channel. Bartek, thank you once again, it was a pleasure to host you, and talk to you soon.
Bartek Bilicki: Thanks a lot. See you.