Thank you for the wait. DTA, can you bring me all the approvals that are pending? Those are the ones that still need to be approved. DTA, can you give me the details of item two? Would you like to approve and release the next steps of the process? Yes. DTA, please, can you present me the best guess for the forecast for the monthly month? Would you like to have a graphic vision of the data? Yes. Would you like to send a message to the VPI of sales? Yes. The message is, "Good luck." Okay, so a good target is the overcoming one. DTA, could you prepare me a presentation with all the summary and the relevant information for the first quarter of 2025? And can you also add some comments about this period?
Would you like to schedule a meeting with the CFO for the evaluation of the second quarter of the year? Yes. The possibilities are endless.
[Foreign language]
Good morning, everyone.
I'm very proud to be here with all of you.
Looking here at those people, so many people all together to follow everything that TOTVS has to show regarding innovation, what AI is bringing to our life, everything that we can do to improve our daily basis, our company. Again, I'm very proud to be here with all of you. Thank you, thank you for being here. Thank you, our board, our council, our sponsors, our partners, and mainly each one of you, the customers and everyone from TOTVS. We know that today we have a big amount of companies that are here that are not our customers, but they want to see everything that we can bring, everything that we can change on a daily basis for every company in this country.
Again, thank you, welcome, and I really hope that we have two wonderful days with lots of content, and I really hope that we will leave here much better than when we arrived. Today what we will see here basically is an AI topic. We will bring a set of concepts, and as I said, we will try to decode what AI is, how it has been evolving, and how we will be able to make this be something huge for every business. We will start here by showing that the evolution of the management software through years, through the decades, is an evolution that goes together with the AI evolution. Those journeys somehow have brought a huge leap in the productivity, and for me, it's just the beginning of it. We still have to achieve lots of productivity, efficiency in all the sectors, in all companies.
It's very, very important. Let's see how the evolution was between the management systems and AI. In the beginning, in the 1980s, 1990s, what we had on one side were the first models of automation. AI didn't come in 2020, 2023. The concept of it has been on for so many decades, as well as the management systems. We started what we called SIG, which is an integrated system. In the 1990s and 2000, we started having those statistic models, the first systems that would forecast the results. On the other side, there was the RP abbreviation. It was something very effective for the back office. Later on, in 2000, the first decade of 2000, we had the machine learning that we already know. It's a very classic one, the machine learning with the algorithms and evolving.
On the other hand, we had the evolution of the front of applications for management with the CRM system. Then we had another important evolution within this AI world called deep learning. That was the beginning of the generative AI, where we started having a very important advance that we call the neural net. It was the natural use of the language inside it. For the management software, we started having what we call vertical automation level within those systems, especially what we call front systems, and they are very big. Everything was connected to the omnichannel and e-commerce. What we see from now on is already a reality. In this presentation here, what we call here from now on, it was for a very close future.
I said, "No, I want to make an adjustment here," because it has turned into our present. It is already happening for the AI, what we call agentic AI, that they can act in a very automated way. On the other side, we call agentic ERP. The application, the effective use of agents that are doing tasks on a daily basis into those softwares. To be able to have an AI that will impact our routine, we need to build a journey and evolution. That is what we will show here, a little bit of those steps, what we need to get to this place. We will go through the history of what AI is. What we had in the 1950s was the concept that there was the possibility of creating something artificial that would replicate what the human science created.
In the beginning, we had the machine learning, where the machine is already able to learn with algorithms and evolve by itself within what the data gives to it. We went to deep learning, where there was an acceleration of this learning process without the machines. We get here to the generative AI. The difference about it is a democratization of the access to this AI. Through this natural language, through prompts, we are able to interact, and it opens an amazing capacity of taking advantage of this use, which is the use of AI. Today, we have two big models being used in the AI in the whole world. I remember that last year, in 2024, this opening was based on AI, and at the time, if we had this slide, we would have shown just LLM.
Today, we already have the small models, the SLMs. The LLM is a model which is something general. Most of the interactions that we have on our daily basis, like ChatGPT or any other tool that has been very popular, those are the big models. They have a quantity of information that is something huge, like a trillion, trillion of parameters, as we say, and they can answer anything. At the same time, there is a difficulty of being accurate when you go into more specific topics. The cost of processing those systems is much, much higher. It is not perfect. It is not the best response that we have for our challenges. That is why we have the smaller models, which are the SLMs, and they can generate a level of knowledge that is much higher, much deeper.
When we speak of the use of AI within the ERP world, the possibility of SLM being more important and something deeper is something that is much, much better. That's what we will see through the day. Of course, when we speak of AI, we are speaking of governance and security. Today, there is a concern about it. As I said, it's not only not sharing data, like data that belongs to the companies, and if you use them in an appropriate way, you are exposing a very important asset of the company to the market as a whole. You also have the risk of using data that you shouldn't be using. Again, the concern about security, governance within AI is something that is much more important.
Again, we will talk about it, and TOTVS is very prepared to help any company regarding this safe approach. When you are in a safe environment with the government, it makes the whole difference. When we look at AI today, as we saw the word today within AI, are the agents. Even more, we have the agents and multi-agents. What are they? An agent is the capacity of turning a task into something that will be autonomous. You will have a system that will enable you to take decisions regarding this in an autonomous way. Of course, you will have different levels of autonomy within this task. I always use a frame that is very simple to understand the path that the agents will take.
The more critical and the more complex a task is, the more difficult you will have to have this level of autonomy because this task still has a processing cost, but mainly because it's a critical task. The critical task will generate a risk of error, and the cost of this error will overcome any kind of saving that you have. The other way around is the same. Tasks that are not so complex, for example, the chances of having a level of autonomy within the use of agents, like can be total, for example, and the speed of those, the way they will be implemented will also be in a higher level.
We are working here at TOTVS to have both models with more complex and critical things to gain productivity, always associated to a less level of autonomy. We will help the user, and in less critical tasks, a level of autonomy that will be higher. It is always important to bear in mind AI with what we call the human intelligence. AI has to be together with the human intelligence. We are not speaking of replacing people. We are speaking of AI and agents that will actually turn them into superhumans. The execution that they have today will be much bigger, will be much better.
The expansion of what we call the human capacity of carrying out things does not only improve the tasks that you have, but they also turn the management of the team into something much more powerful, but with a very important detail. Who is in the front of the usage of AI on a daily basis and with all the daily processes? For sure that you will be ahead. You will have a gain of efficiency, productivity, and you will bring more profitability as well as sales opportunities. It will be, it is a run. It is a marathon. If you are in the front, of course, that you will probably win.
What is most important with AI, with the agents, is to have the capacity to identify concrete issues of our day-to-day, concrete issues about the company's process, and within this process, these concrete issues to correctly apply AI. AI should not be a rocket science. The practical application of AI is what will make the difference. We do not need to create our model to be faster or in the front of this AI run. We just need to have the right understanding of the AI functioning and how we can apply this in the process we have already in our companies, expanding incredibly the team's capacity and consequently our productivity and effectiveness, and giving more time so we can focus on what is really important and strategic in the core business of our company. TOTVS is adopting this AI revolution that we call our position.
What is the great competitive advantage that TOTVS has in this AI environment? It's the knowledge we have in every economy segment. TOTVS has more than four decades of history present in all segments of economy, in different company sizes, in all geographies, whether in Brazil or Latin America. This gives TOTVS a level of knowledge and access to a data mass, making TOTVS a great expert in the creation of these agents that will make the difference in your companies. Count on us, count on TOTVS to create these agents and to make all this capacity available to you.
Of course, if we are working and gathering all this data and putting all the data in this functional knowledge of every segment, if we are creating agents that you can use to get efficiency and productivity, this means that you have more time, more possibilities to focus on what you do best, your core business, the focus of your company. Last year, exactly one year ago, we brought some concrete examples that were already being applied in terms of AI use in our day-to-day. These are five examples. We have these applications running in several segments and types of processes. Just to recap here, these are very interesting data. Today, we have a number of companies, including the SMBs, people using AI in their day-to-day, 50%. Today, we have more than 50%.
Look at this number on my left side, amount of companies that are really extracting the real value of AI is very low. Only 6.5% of companies are able to do so. We are here exactly to take the next step and help each of you to make this percentage, to put this percentage at the same level as the one at my right side. Basically, what we are talking about with an agent and automate and make tasks automated in systems, you will hear more of this expression, task as a service. Many years ago, we had the SaaS, software as a service, exactly to show the new model that was replacing the traditional model of license plus maintenance when softwares were offered in the cloud and as a service. This brought an increase in productivity and efficiency from the providers.
When we create AI, when we create SaaS, the leap will be even higher. Just like we have SaaS living together with the maintenance and licensing model, we still have for many years the SaaS model living with SaaS model. They are not excluded; they are complementary. Today, TOTVS has already created what we call this big AI agents ecosystem. We have already, and we will show some examples of this capacity to create agents to all segments, to all company sizes, to a great number of processes and modules within our system. Let's start with the video where we show our agents store. This is just a marketplace of AI agents within TOTVS, but this is a reality. It's available today, and it can be accessed through our cloud, our cloud platform. Here, this is a showroom with a simple browsing.
When we go to our agent store model, we start to see that we have access to enable a series of agents in our day-to-day. Again, this is available today to our clients that have access to our key cloud. It is important to come to our TOTVS cloud, you have access to this agent store. The second video is what we call the DTA Studio, which is a control panel of AI agents. It allows management and the configuration of the agent command. The purpose is to provide some intuitivity in the use of these agents, and it also allows to have chats with natural language. It is the connection and integration of the agents with the AI applications within TOTVS. This is not a dream; this is already available to you.
The last video I want to show you is what we call voice command for AI agents in our WhatsApp. We have this capacity of integrating these agents with our WhatsApp. This is very easy when we saw in the opening of this event, gives you the possibility of interacting and generating a series of information. This will give you an interactivity in the expense management for the financial department that will be much more assertive and accurate. When we continue with this voice command, we see the capacity of you to command in a very practical and safe way or secure way. Again, data security is important. This is an example of a payment to suppliers. In practice, this does not add value. So clearly, it's a task that can be automated and can be used by an autonomous agent.
Again, these are agents that are really working and available to you. Basically, the message we want to convey and that I want to convey is that the future, the AI future, is already in TOTVS. I said last year, and I will repeat, if you think AI, you think about TOTVS. I will invite some VPs, the TOTVS VPs, to come to the stage and share with you the use of these agents, Gustavo, Marcelo, Abelard, and Apendino, please.
[Foreign language]
Good morning.
[Foreign language]
It's always a pleasure to be in this giant stage with so many people. It's a great pleasure to share with you this story. Let's share how we are going to do and deliver this that Denis presented. We look at this conversation, and this is so much aspirational. It's so far away.
No, we see that this way is not so far. These are organized steps that we provided the solutions, and you maintain your process in-house and make the decisions to reach these results. A very fast parallel here. What we are seeing here, the importance of the systems will continue. In the last mile, to execute something, you have a technical asset to do this: an ERP, a segmented app, anything that is available at TOTVS Solutions Portfolio. Another important step is to have this in cloud. Why? Does it only work in cloud? In practice, we can provide you all these resources, especially data organization, when we are in cloud. If you have in cloud, we have key cloud to help you. Is it enough to be in the cloud? No. I need a clear organization.
Many challenges, not all data are available in the management system. We have other external sources that are complementary, so it's also important to have a data platform. But Gustavo, okay, I have the data platform, organize the data. How can I have access to AI models, LLM models, or develop practical things in-house? We have a platform for that. In this case, it is the DTA platform, Digital Trust Advisor. This is the digital version of the Trust Advisor. We can facilitate the access, a framework, so you can access AI, and this will go finally to the agent concept. This is a very interesting cycle. It's a retrofit. It has a feedback. You have the agent; you can find out a new functionality that you can develop in your applications.
It's a journey that will continue to evolve as the tools are used, and you go deeper in your learning. How do you do this in TOTVS? Many people ask, some clients. We talk about this with some clients. How is this journey? How can I organize myself? These are some examples we did in TOTVS, a hybrid structure. Sometimes we do centralized things, sometimes decentralized. In our case, policies, best practices, governance, what we can do, what we cannot do. We do this centralized, spreading all over the organization. We have people researching according to security, privacy, how this will facilitate the daily lives. We do this centralized, spreading to the organization. The use cases, just one single area of the company has enough knowledge to treat all cases, all situations where this tool will be useful. In this case, we have a decentralized approach.
Every product and process owner will generate knowledge that will feed the platform, will go through the data platform to be spread all over the company. Finally, we adopted the experience of using multi-models. We are not in one single AI model or execution. The platform helps in this sense. According to the situation, we can choose the best model, the best application for everything. How does it work? Let's see some more detail on this component. When we talk about the data platform, we go from information collection using models that we have already in several products TOTVS has, up to the organizing of this data accessible, ready to be used. It's a great challenge in this. It seems simple, but if you don't have a good data governance, the fuel to feed AI will be characterized.
We need to do this together with you in this platform. To use the AI term in the data platform, we see some AI mechanisms to use pipelines, surveys, and searches to ensure the best performance in data use. The second step I mentioned, the DTA platform. What's the main idea here? It is to bring observability, governance, know the models, the models that were applied that had more success or less, not only to apply the technology, but to manage it, to follow, to have a follow-up, what's been successful, how you were using this. You have a framework to facilitate the observation and data application. Over the presentation of TRACS AI Arena, you can try in practice what this component means, and you have also a roadmap to explore this.
Just to give you some numbers, we have more than 100 projects focusing on the use of the platform and their components, always using these concepts: live in action, execute something live, content and knowledge, generate content and knowledge, and build and code. I will give you some examples over this day and in other presentations. We have things already happening for this. Also, when we talk about AI, both in the products, in the portfolio, and operational efficiency, you can look at this as a tool or a process point of view. Denis said we are preparing the cloud to do the whole arrangement of it, which is T-Cloud, and you will probably see those changes to make sure that we have the data, the governance, and the security of it.
You will see many things going on within the T-Cloud to support the model that we are mentioning here. Of course, the intention here is to have this network, which is the agent store, and our plan is not to have only this. We want to have you with us producing content and feeding this store and all the agents. Here we have some practical examples and videos to show what the DTA agents are. The idea is to build your own agents. We have here the availability with code, low code, or even with no code. The example that I show here is the creation of an agent with the example here to evaluate documents from LGPD to see if they are sensible documents, if they are according to the law.
I am just asking for the creation of them through prompts and/or through codes because some people prefer codes to anything else. I have the agents that are available. I am dragging here a document to be used, and asking the agent to check this document. I have just created an assistant to look at LGPD, and I will ask it to process the document, and then it will bring all the information regarding this. The idea is to facilitate this, and once it is created, it can be published here in the store. There is available information about it, the resources that are necessary, and you will be able to consume them through the store. You will organize those agents through what we call DTA Studio. What are the agents that are available? What is the history, the questions, the topics?
The traceability is very important to compose the process, to understand what we are doing, and we will make them available through DTA Studio. At last, some examples. Within those applications, we have an example of payroll in Proteus. We have all the available agents. As Dennis said, we are always betting on the human intelligence, and the idea is to give power to our users to use them. If you want to learn more about it, to understand more about the technology that is behind it, go to the tracks, to the AI stages, and then you will see in all the trails everything that is already being applied, and people will reinforce more about it. Now I'll give the word to Marcelo, and he will share some real cases with you.
[Foreign language]
Good morning.
I will bring here the DTA that is applied to the things that we do. I will show three cases. The first one is our cost cockpit. You use the AI to close all the manufacturing costs because we know that this process, the cost closing, is something that is very slow because you need lots of information from the ERP because it was not built the way the information should be there. You will have a new interface to do the cost closing, and the DTA will help you through the process. You can ask questions. Here is an example. You are asking DTA a question about the asset, how it was used in the process of the manufacturing, and at the same time, it gives you some insights, like what was positive, negative.
We are working not only on the AI application, but as well as on the journey of it. Then you can gain productivity. That's with the AI, with the human intelligence. Another example that I bring here to see how your inventory is with the support of AI, there is a new dashboard that consolidates all the information that is in the products. I'm giving some examples of DataSoo, Proteus. We are working a lot with the interface of it, so then it's not so connected to the backend, and then you can benefit from the modern things. With this example, you will be able to change the graphs with the support of DTA.
Everything that Gustavo mentioned here about technology, you already have an application here, and everything is available in the versions that we will be launching here in the second half of the year. Here is the main cost, so you do not need to leave the screen, go and talk to someone. You have everything here. The DTA is checking if the person that is asking for the information has access to it. Remember that ERP controls a whole company. We know because we have done some tests, and the first question is, what is the salary of my boss? If you do not control this well, then the AI will give you the response because we do not know if this user can access it.
It gives us a security that is very important to all the users, and then the information will be according to the permission that you give. For example, the inventory area just sees a piece of the purchase department, and it can even say, okay, you do not have access to this information, for example. The next example here is how we will get the AI and make it available so you can have a simpler process. Here there is a functionality that is simple but very powerful, and you will get any document and include this to an automatized process. You all design a process, and a human has to validate it if it is not a picture on the beach or a driver's license or a passport. It will enable you to have DTA validating it instead of a human.
Here you are uploading a driver's license that can be used in the health department, construction, education, in the material receiving process. Another nice example here, DTA supports in the construction with the RPA information. It shows how you do the process of this and then what it should be doing. In this case, for example, an engineer, a bricklayer will be able to have a corrective action during the process. I was, for example, building a wall, and now I want to change something. I have to ask for something. Instead of leaving the interface and then logging into the system, get a mouse, a cell phone, at the same time he's interacting with it, he will be able to see all the information, the next tasks. Everything is integrated.
We did not want to be disruptive, and then we enabled our users to see all the evolutions of the product by gaining productivity. AI will have a very big purpose. Now we will talk about Winthor, that is the new face of it. I invite you to go to the booths of it. Here there is a nice example of how you do the register of a new product. You know that when we register a product, that is the key for a good management of it. Because if you do not do this well, all the information will be wrong. All the reports, like sales, profitability, everything will be wrong. Here we have shown an example that by using your own register, DTA will help you with this process.
Here there is another example that's called Stock, and it is helping you with AI doing the sorting out of a certain product store, and it will show what you need to sell. It will help the customer with the decisions, so what kind of promotion he has to do, when he has to do it. Another example is a Salesforce automatization software, and then it's helping the salesperson to sell more. Here he has an interface with WhatsApp, and then he asks DTA some questions like, what is the day today, what is the target for the day, what is the route, if there is someone that hasn't bought anything yet, and then you will put everything together, and you will bring productivity so the salesperson can make his day easier and work more with this human relationship.
You can do the order, you can see what you should do during this week. Everything that you do when you call people, we want to give this DTA too as something that will bring more productivity. Good morning. Bom dia, pessoal. Good morning. That is my second universe here. It is a pleasure to be here with you. I was talking to Vivian, and I said that this world has grown a lot, but it needs to grow even more. There are people here who are trying to look at our presentation, and we said here that without data, there is no AI. In the RD, we do not have this step of the cloud, which was a very clear message given here by Dennis and Gustavo because our products are right in the cloud, but the steps are very important.
A crucial part of it for the outcomes that we have when we go ahead is to organize the sales process for our customers and organize all the base data oriented. From this on, we are able to advance and deliver many things with AI focusing on the salespeople, aligned with what's shown here, focused on marketing people who operate our products. We also have a huge difference between performance and delivery of results because we have the heavy user and then others that are still in the beginning or are still in the middle of it. The AI has as a main goal.
I will talk a lot about it in my presentation during the afternoon, so I really hope that this place here is crowned the same way as it is now, but it has as a goal to turn everyone into a heavy user and turn salespeople into super salespeople. I will show here an example that's our co-pilot of the CRM product that is aligned with some functionalities that have been shown here. We can operate our product with voice and feed it through voice in a very simple way, just between a conversation between the salesperson and our AI. It will be a co-pilot that in several steps for the use of the product, it will make sure that we will have accuracy, quality for the use of the product, and automatize many steps that before we needed some manual inputs from the user.
If you could access CRM through WhatsApp in real time, like in a meeting, I will show how it works. Hello, hey, so I've just left a technological meeting, and I met Andrea Gomez, the manager of Tecnova. Now you have the name, Andrea Gomez, role manager, company Tecnova. You can continue with the register here, and now give me an assignment, so then you will send a follow-up email to him next Tuesday. It's one of our features for co-pilot, that's the voice operation. We have many others. We record the calls between customers and salespeople, and then from this time on, we recommend the other steps. I will show that with more details in the afternoon.
After this, after so much information about AI, you must be asking yourself how you can buy an AI product from TOTVS, and now we have Bettino here who will explain about it. Thank you, everyone.
[Foreign language]
Good morning.
[Foreign language]
I am the salesperson who Dennis or DTA charged or demanded, sorry. Before, I would like to change into Spanish. We have people from Spanish, Chile, Colombia, Mexico, Paraguay, Uruguay, Peru. We have people from the West here. As everyone knows, we have operations in the whole Latin America. We have many people from those people, people who came from far. We have customers from all the states, from Brazil, and we have people who have arrived just today. How can I have everything you are showing here in my company? How can I implement AI? How can I get here?
The first step is to migrate the solution version to the last version. ERP and HR must be with the last version. They could use that. 80% of our clients are already with the last version. This is incredible. It's not common in the ERP and HR segment, but we still have some clients who are not in the last version. This is a special request to go to the last version so you can have access to everything we are showing with the version. Now we ask the TOTVS Cloud. This is a common question. Many of you are just getting knowing TOTVS as a proprietary cloud. We have cloud independence. We have 40%, a little more of our clients using our cloud solutions, but we still have a lot to evolve. Another special request. We have specific tracks talking about our cloud.
It's not a $10 cloud. You have your governance. We can manage your application. You have several advantages in accessing our cloud. These integrated agents to TOTVS solutions will only be accessible from TOTVS Cloud. We have data security, data processors, so it's important to migrate your solution and take it to the TOTVS Cloud. Also, I'd like to bring some cases from our clients using our cloud solution. The Fini company, everybody knows Fini Candies. It's a Spanish company. The only plant outside Europe is in Brazil, and they run TOTVS solutions in cloud. I don't know if Javier is here. Madeira Madeira. Everybody knows Madeira Madeira. It's a company starting online, and Madeira Madeira runs TOTVS solutions, more than 15 distribution centers all over Brazil. A totally digital company, and they use TOTVS solutions. Fadel Transport, they run TOTVS in Brazil, Uruguay, and.
[Foreign language]
Logistics company. Fadel runs our solutions. They are the largest distributors of Ave Drinks, Castelo Alimentos. Castelo is 118 years old. They run everything with TOTVS. They are very well known by their vinegars, but they manufacture many other things. Sebrae, we know what Sebrae does to Brazil. Sebrae also runs TOTVS in our cloud. Sebrae had productive and performance gains that were giant. I'd like to show some logistics points because logistics is highly complex in terms of cloud. You have orders coming and going and trucks coming and going.
These are some of the clients. We have more than 10,000 clients running in our cloud. We have native cloud solutions, and if we add them, we go even more. We have opportunities for clients to migrate to the cloud. This is important. If you are not in TOTVS cloud, we cannot add the AI agents that we are bringing to you. Finally, it is to organize your database. To organize your data, you need to have an ERP, HR, a complete data organization. For this, you need a TOTVS solution with a back office, marketing, and HR. If you think AI, you start from TOTVS. This is another tip to you about the universe. The universe increased over the years. In our land, we have several stages called the code.
If you want to break or reach your target, we have all code stages of this size. If you come to the universe, you have HR, you have fiscal tax area, and we have a lot of content. We used to have no logistics but we increased our size. We have fancy design, and this is the plenary. The plenary seems to be giant, but it's just that little corner up there. We have another side with the people world, code no code, so it's a giant event. I ask you, go to the other parts of the event. We will have debates about AI, which is very close. It seems far, but it's not. We have AI debate and a specific container AI lab, how to create agents. Clients help with this morning that is creating clients for re-establishment with our AI agents.
We have a big event on the tracks. We have events on both sides of our pavilion. Also important, this part of the tax reform. We have a crowded event. We have a great demand. We have a specific stage and a specific track on tax reform. We have many clients leaving for the last time, the adjustments of tax reform. So pay attention to this subject. We know how Brazil is. We find some difficulties on this, so we have a specific tax reform track. These are the solution space. It's a big space. And we have the opportunity of a social opportunity. It's not the iOS of your smartphone. iOS is where we have our gift shops addressing the training of low-income people. 7:00 A.M., we have a line in our gift shop.
We have many, many nice things, some sweaters, a T-shirt with TOTVS new logo, so many, many nice things. To include many people back to the stage so we can close.
[Foreign language]
You are on the right day when you are beside those who are working with TOTVS and create solutions that solve the problems of many companies in Brazil and abroad. We are together. We are close to those who plan, serve, deliver, care, educate, sell, construct. We have technology and successful histories. We are efficiency. We are present. We are partners of a Brazil that is always running, increasing, growing, and doing. Companies that move the country trusted TOTVS because the Brazil that makes it makes with TOTVS.
[audio distortion]
Recording in progress.
[Foreign language]
Good morning, we will start Investor Day in a few seconds.
[Foreign language]
It's a great pleasure to have you here in this 2025 edition of the Investor Day.
As a tradition in this universe of thought, I'd like to welcome all of you here in person and those following online. Those who are here, please go to our tracks, our content. It's a very good opportunity to have to be in contact with everything the company is bringing. We prepared a morning, a very nice morning with a lot of content. We have our executives opening venues to talk about our strategy. The executives are talking about the trends, cloud, AI, hot topics in the market, apparently bringing the perspective of some themes on this revolution. Vivian, our HR VP, is bringing novelties and how the company is preparing itself for the future. I will back to a Q&A session for those here in presence, in person, and those remotely. I invite Dennis to come to the stage.
[Foreign language]
Thank you for your presence.
[Foreign language]
To talk when we have less people, when I came up in the morning, lots of people standing at the back. It was scary. I will be very brief and basically I will convey two messages in this opening. First is what we call our AI vision. What I'm showing here, we are already talking and showing to many of you in our interactions in these last weeks and months. Uma visão mais completa. And maybe here we will have a more encompassing view of many subjects we've been talking about. Basically, the first thing is to convey this message. I will going step by step to finally reach this AI vision.
The first step is to show you that the enterprise software, the management software, is made of two dimensions. One, that is the best known for those who are not experts in this subject. Many times the client recognizes this first dimension that we call the technological or architectural view. Every time we talk to someone who is not an expert in this theme, talking about the threats, for instance, disruption in the management of the world, is just like talking about this first technological architectural dimension. It is important, of course, but we have a second dimension that we call functional dimension. We have no doubt that the functional dimension is heavily much more important than the technological or architectural dimension.
It was funny because I was sitting over there with Avelar, and when Marcelo was showing the screen of Winthor, one of our platforms, Avelar said, "Wow, this Winthor screen was evolved a little bit in terms of layout." I told him, "Yes, the Winthor screen is the concrete proof of the first screen I will show, because Winthor is the main product for the distribution market." Why? Because functionally, there is no other product in the market. The level of accumulated knowledge of this product is incredible. It's specialized in this segment. Keep it clear in your mind. You do have technological and architectural. They are important, but in the management software, functional is the weakening. When we go to the new technologies that are coming, and today we are talking about AI, but we have already talked about cloud.
We talked from going from DOS to Windows. We've talked about so many changes. What we see, and this is due to the importance of functional being more important than technological and architectural. These new technologies, they increase the addressable market because they do not change the functional level of the application. They change the technological and architectural level. When the functional is preserved and AI, in our view, is not changing it, they become an increase to the addressable market. Maybe this is one of the main misunderstandings that the market has in terms of the size of opportunity and why TOTVS is able to preserve and keep this growth performance for so many quarters, regardless of the economic scenario.
[Foreign language]
there is no company in the market with the accumulated level of knowledge and data in this functional dimension for any segment in the economy, for several different sizes of companies, just like us, TOTVS. TOTVS is, for sure, the company that concentrates the largest expertise and data quantity. In an AI scenario, this gives us the possibility of winning these fights, not only in agents' creation, but better agents, more effective, more accurate, and everything together becomes a better cost of this AI application. For those who saw the opening, saw the difference between LLM and SLM, the more specific and functional knowledge you have, the better is your capacity to create a smaller model, an SLM that will give you this competitive advantage. This is what we believe.
In our view, in this AI world that is already a reality, what we see today are three big groups of winners. On one side, those that we know better, the manufacturers of equipment for this AI revolution, NVIDIA is the great exponent in this. We also see providers of computational capacity, and these are the usual subs that were the winners in the cloud world: AWS, Azure, Google Cloud. Also, in our view, they are winners in this market. This third group of winner companies is where TOTVS is. We are talking about specialized software developers with an incomparable level, with a quantity of data and expertise that are unbeatable. With this, the capacity of creating agents that will create this addressable market. One year ago, when AI started to be strong in the market, not necessarily this third group was recognized as a winning group.
Just the opposite. Many people thought that SaaS would disappear. For us, it's very clear, and we showed this. SaaS will, far from ending, but many companies that generate SaaS will be the main winners creating SaaS. What should we do? What is our role? Where should we look at? We should continue expanding our functional expertise. This is a great concern of ours. When we look at an opportunity just like this one we are negotiating with Linx, we are talking about expanding a lot the level of knowledge, the access to data, and functional expertise, in this case, in the retail market. We need to continue developing the capacity of agents development. You saw the agents we have today and the tools we have already created for development within TOTVS and outside TOTVS.
With more knowledge and competitive advantage we have, we are not able to create all agents and tools in-house. We will have more people developing and working with us. Of course, the third step, a crucial one, and we reinforce this in the three messages to our client. We need to continue mastering our application and master also the environment where the application runs. The application is totally controlled. The telemetrics we have today and the level of APIs controlled in our application is in a totally different level than what we had some years ago. We are going towards to ensure that most of our clients will be running in our clouds. These are essential movements.
[Foreign language]
what we are saying is something like this.
I like to have a background that I could design here was not possible. Basically, this is what we are saying. When we see the total cost of a property in the point of view of the client at the origin, we said, "We have the application with this continuous cycle in the center, and this has always been the original addressable market for TOTVS." This application was always running in an environment that we call the deployment environment of the application. This always existed under the point of view of the client, but it is a dotted line. It's not a continuous line because it wasn't an opportunity for TOTVS. The customer, the client, spent money on this, but not with me. We always had this more external layer that we are calling human agents that impressive.
Means the client, the user who is operating the software, whether it is in the financial department, accounting, tax, production, all users are there, whether they are employees or a VPO. For the customer, it always has been the TCO of this customer. For us, this is the addressed market, only the circle that you see here in the center. When we had the cloud, we had an amazing change. At first, there was a lot of concern that the cloud would enable, would bring more disruptive competitors, and the cloud will activate this. Into a proxy, it was the other way around. Basically, there was no new competitor in this market. We had the same players, but then this addressing market grew a lot.
The dotted line here then turned into a continuous line because then we had the capacity to offer the application within our cloud. And our estimation here is that this market of the cloud will be two times bigger than the application market itself. As we already run, we have run customers in our cloud for more than one decade, so we know how important this multiplier is in our customers. In my opinion, it's only the beginning of it. When we look at the AI, the conception of the cloud, even though it's within our application, our specific cloud, it will probably grow a lot in the next years. Of course, that for the TCO customer's point of view, and here you don't capture the same way we design it, the TCO will be lower for him.
When he runs the TOTVS application within our cloud, it will be much cheaper than running this application in this on-premise environment. It was a very clear, it was a very gain-to-gain situation. When we have AI and we go forward, this agentic AI or ERP, what we have here is the final step for those technological changes that will make the TCO lower again. For us, instead of having this as a threat with the appearance of new competitors that do not have this expertise, those functional data, what we see here is the opposite. Again, the expansion of this addressing market. Now we look at the, like what, how much the, like the people spend on employees, VPO, and then we look at it and say, okay, we can get those pieces from this market.
Here, instead of having two times, we are speaking of an opportunity that is 16 times more than with those users, than with the application itself. The ERP will replace all those users. Of course not. No, we will not replace them completely. When we speak of 16 times, if I capture 10%, I am already speaking of a market that is two times bigger than the application that we have. We have opportunities for like many, many, many years ahead. Of course, that if we execute everything that we need to, at the end, that is our view. There are risks, yes, of course. In our perspective, the main ones are the normal ones that we already have in any company.
[Foreign language]
Secondly, we need to try to get to what I call the core of TOTVS. That is an amazing company.
It has a very rich history. Even the name itself says everything. TOTVS means in Latin everything, everyone. Many times we are not so clear about the core of our DNA. What do we do the best here? I really wanted to give this message in a very clear way. The core of our DNA is to work with this SMB. TOTVS has many big customers, one-third of them that are in the market. Yes, of course, because TOTVS does many things for many people. What do we do here? The ideal profile for customers, the presentation that we had here in the opening was addressed to the SMB. It was done for this kind of audience. Of course, when we say this SMB, if you ask SAP what SMB is, the answer will not be the same that TOTVS provides.
We have to spend like two minutes on it. It's worth doing it. We will explain how we see SMB. We are speaking of companies that usually will profit from BRL 100 million to BRL 2 billion a month. We have customers that are above BRL 2 billion but have very similar features. There are companies that are below this number, and they also have very similar features. This is our ECP. That's the customer who we are used to working with. That's where we have our speech, the way we arrange the company, the profile of the product fits into everything. For this customer, I always say that for this kind of customer, to be this trusted advisor, to be this partner, this reliable partner, it really makes sense.
If we have a company that has the ICP at the top, that's what we call large enterprise, then the trusted advisor doesn't fit into it because those companies are not trying to concentrate those partners, the tools, the technological tools into just one partner. That's the other way around. Our example, which is a large enterprise, every time we start having this level of concentration in just one partner, it doesn't matter who he is. There is a light, there is a yellow light, and then we start getting concerned about it because in our customer profile, it's the opposite. This guy needs some support. When he finds a company that he really trusts, that he knows that it will be a healthy relationship, that is looking ahead, then every opportunity that he has, he will want to concentrate the solutions into this partner even more.
That's the space, that's the position that TOTVS has been building for so many years. It means that our strategy can be summarized into one simple sentence. What we want at the end is to increase the relevance that we have into our customer. We want to be more present into this investment in technology, not only in technology, but with this technology that will be feasible in our customers. We can translate this like this construction, this position into a metric, the take rate metric. In this case, the take rate is what the customer spends on me, like a month, a year divided by the revenue. It can be the month, it can be the year. We are working to increase that, or the take rate, like all the expenses that the customer has on us. Here we have some interesting numbers.
When we have here our data, we are speaking of 15,000, 50,000. We have here this percentile, which is 90%, the average for take rate. Again, what he spends on us divided by his revenue monthly or yearly. The average here is 1.14%. When we go down here, 75, this number is half of it, 0.59%. Then when we have the percentile of 50, we are speaking again of a decrease that is more than 60%, like 0.21%. Probably we will have questions about it, and I will anticipate one of them. Is it possible to get to 1.14%? You probably have some differences according to the economy, level of sophistication of the companies, the needs. You have some situations where we are not the only partners.
There are several factors, but of course that we have here some passes that we already know, like some journeys that have been taken that show us an opportunity and a possibility of expansion for this take rate in a very amazing way. It is not that we need to learn how to do it. Actually, it has been like many years that we have this meshing because it is the cross and upsell for the ones who are already customers. Yes, of course that it is important to have new accounts here. Most of the people who are here are prospects, people who are not our customers, and they are very important because we still have a huge opportunity. 25% of our sales are due to new purchases, but at the core of our company is the cross and upsell.
Now I'm going to the end, and when we show here on one side, we have our perspective of AI and the other our core for the ones who will work, like the ones who will work for who the SMBs are, the way we are like our position, the place we want to have in this relation. I will tell here a quick story. We started our life, as everyone knows, in what we call this like management, like our vertical applications, CRM. As everyone knows, we have evolved in those last years to build two new business units, the RD and then Techfin. We opened a new field of opportunities and The Economist of last week. There was a headline about it, and I posted it on LinkedIn.
There was a nice piece of information regarding this topic about this merging of this kind of world in big companies, and the comparison was between SAP and Salesforce. SAP always domaining the back and the other domaining the front. Most of them, like both solutions were there and have been together in a friendly way with very well barriers. Today those barriers are disappearing, and this relation is not so friendly anymore. They see both as a potential opportunity. When we created our old business performance unit, we had seen this opportunity. Actually, we saw this in 2019, and it had already seen because in DCMB it makes more sense because of the concentration of just one partner. Somehow it was a pride.
We are even provoking this, the economics, to have something specific about TOTVS to talk about someone who has been successful for six years differently from those other companies. Again, that's our reality through three business units that we have here. The takeaway here when we are focused on SMB, when we see an opportunity of take rate as we see here, when we want to occupy a space inside a customer of a trusted advisor, increasing our relevance. When I look at this endless horizon of time, we don't feel that we have any kind of limitation here, any kind of possibility, technological possibility that really makes sense to our assets, to our expertise, and things that fit into our customer can then be some interest objects. What are they? It can be a distribution platform.
That's the sale machine of TOTVS, a massive client base, 140 of what we have, all sizes of customers. Of course, that a high-end RD investment capability, that would be something amazing in our Brazilian market. It's not something that we need to learn because we've never done this. Table that we are showing now, it has been updated. It clearly shows that TOTVS can do it. It's already doing it. In those 10 last years, when we look at the GDP, we see, and that's not the real GDP. We are looking at numbers that are between 6% and 80%. When we look at the growth of the market or the RD market, we are looking at two or three times that more growth.
When we look at the performance of TOTVS itself in management, we are speaking 1.3x the development of the own market. For the RDs, four times in the last five years, we are prepared to take advantage of those opportunities to carry them out regarding this AI perspective and address those new markets for continuous growth. Thank you. Let's continue here, Marcelo and Gustavo, right?
[Foreign language]
Good morning. I hope you are enjoying the content, but it must be cold weather.
[Foreign language]
Opportunity, if you want to buy.
[Foreign language]
It's weather. It's a good reason because it's cold.
[Foreign language]
In this context of AI, appearance, the functionalities of RD, they continue with the segmented use. They are really relevant. It's important to emphasize that this is an evolving portfolio, a still growing portfolio. This is very important to our business. When we are in a client, we see an innovation, a functional opportunity, a change in practice of habits or a change in legislation that leads this portfolio to be expanded. Until some time ago, we didn't have a vertical for rentals. We have industry to making products and selling to consumers. Today, instead of selling the product, you rent the equipment or you rent that asset.
It's a vertical that came, and we transformed this into a product becoming very successful in our base. The evolution process continues in a technical asset. When I use an AI agent for a certain functionality, when you need accuracy, repetitiveness, a very accurate result, you have a software to do this. Whether it is the domains HR, ERP, the health, education solution, examples that Marcelo, we go with detail. Another important thing, in addition to the application software that he uses, is the cloud subject. We have two big points: improve experience and use for those who consume native cloud solutions, migrate clients in legacy system to the cloud universe, and enable all these data issues. In addition to this use of software, our cloud is dedicated to our portfolio. You have a peripheral RBT around this as a service.
.[Foreign language]
As a structure as a service. Basically, we have the largest Brazilian cloud, especially when you talk about a purpose determined and specific. As important as application working well, complete, functional, and understanding, and taking this to a cloud environment with other systems to work the process of reduce costs, I reach the data issue. This is fundamental to our strategy to help clients organize their data, help clients to have governance over it, and allow them to have the proper use of the best field that can be used for AI. If you do not have ready accurate data that are available and controlled, you will benefit much less than what AI can give you. When we are talking about business context, it is important to know the context of the company, the context of that segment. This will be relevant when I am producing an AI functionality.
In addition to the data platform, we deal with a great diversity of products, portfolios, models, technologies, and part of our role, especially when we talk about the SMB, even having the needs. They have their business needs, process needs, but many times they do not have any sophistication. So we built a platform that will help the consumption of AI and SLMs. We authorize a model. If a model is more efficient for a certain use case, we use it with no concern of a different model, always looking for synergy, best application, better use of the technology, addressing the business to enable business with no technical bias. We are always keeping an eye on what is happening. When we had DeepSeek, everybody was concerned. Two, three weeks, we had already used the use of DeepSeek
[Foreign language]
We can also see the traceability on the use of this technology.
[Foreign language]
This is available in the cloud. I help the client. I have a platform to use this data as a field to use the AI functionality. Why not give the client a framework so he can use the agents that TOTVS is developing, but also to build their own agents? When you have a platform at this stage, it's very interesting because we give opportunity for an entire network. It can be an agent built by the platform or by a partner or an agent built by the client himself.
For those who attended to the previous presentation, you see the agents we have, the agent store we have available in the cloud. It's easy to be enabled to be used by our clients, bringing the value we wish to bring for the effectiveness of the company.
What we are saying, we are also working to have a sustainable and supporting journey. In addition to the consumption of the portfolio, the client has the support, whether with the services we have packaged or functionalized services. I'll give you an example. We develop sustaining packages for tax reform. Those clients who need support in the use of the solution and want us to follow up the process, for instance, we have a supporting package. It's a recurrent presence in the day-to-day of the client.
[Foreign language]
We also evolve and improve the use of AI. For instance, IPS integration tool is fundamental in this journey of data connection. It can make transitions and translations, and we continue to invest in this. Likewise, we offer AI in our portfolio, but TOTVS is a consumer of its own products. In the company, we use the products we produce. This is very important. Everything we do, we consume in-house when it makes sense. Evolutions in the portfolio that will be sales leverage. They are also leveraged to increase in-house productivity. We use our own solutions. Also, we are with an eye on what is happening in the market for coding, testing, content production.
We also use market tools with our portfolio so we can consume tools that will help in productivity, speedness, things that we want to give to our users. This is a show what I said. This is our IPS screen. We see here the codification, the coding, objects being tested. This is an example where we produce content with an AI agent, a person with no deep knowledge of technical fields that will be unable to do an integration. I can use the code through the prompt. I make a request, and the tool generates the code to be used. Another practical example. Within the Fluig, our productivity tool or platform, this is a BPM production. I need to create an approval flow for sales.
Instead of design and program, I go to the prompt and I request, "Please create an approval flow with three approvers," and the flow is produced. If you say, "It's okay," it goes to the environment to run the approval in the Fluig of that company. Finally, analytics, in addition to data generation or tactics, the participation of an agent to help you understand the data. Also, you can ask questions for that evaluation. Anytime you have data for high volume, very detailed, and when you have an assistant supporting you that helps you navigate and address the data, it's very useful. This is an overview of what Dennis mentioned about AI strategy. We continue to invest in software application. This will be relevant. We still invest in the use of our cloud.
If you visit our stands in the afternoon, I will comment on this, what we have. This goes to a data platform, data fuel for AI, and AI being used as an agent feeding the cycle. It is a journey where the client receives value, control, management, and productivity. Okay? This is what I have to share with you. I will invite Marcelo to continue.
[Foreign language]
Hello, people.
[Foreign language]
I will bring here how our UX for the future will be like. I think it is a nice topic. I will start here with Datasul. Datasul will not die, okay? If it will die, I am not aware of it. It is still strong with a very strong database. That is how it looks like.
If you have the opportunity to see that and talk to someone, then that's the RM. You will see that both products are similar. We want to prevent the back end of the product, but we don't want to lose this essence of the product. If we look at what Gustavo has shown here, the visual components are very similar. If you are using or any other, it's the same component. Here there is a nice feature. Here is the screen of Proteus, and then you have a... It's something that is very interesting. We have just launched, and the customer can get any information and report and start working as an Excel without removing them from the system. It is very important to control where the data of your company is. Winthor is our product that is focused on distributors and retailers.
It is very good to sell well. Visually, it has a lot of room for improvement, but it has a unique screen. All the colors, the boxes, everything is very similar to the other products from TOTVS. We want our customer to increase his take rate with us. We want him to have the same experience with us. Regardless of the product, we want him to have everything the same. We will now talk about the core functionalities. It is very important because a customer who uses the core will have a lower long-term rate. This is a very poor field here, and we will see that it will improve a lot in the following months because we have the customer, the buyer, and the industry.
We are working on this vertical, and we are expanding this to work the commercialization of the coffee that has a set of processes to work with the sugar cane and alcohol. We've improved our human resource portfolio. We have the assessment and performance. We have information. We have recruiting. We have spots, like online spots. We have many things to improve this portfolio. We will get all the AI functionalities and put them into classical products such as accounting. For example, AI for accounting enables the accounting person to spend less time on what he does. For example, AI follows all the movements and identifies all the standards and identifies the users. For example, the industry 4.0. Most of our data are industries, either Latin Americans that have a lack of solutions.
We want to get our products and make the deployment easier, especially for the industries, because they will capture the need of integration. When you produce, you ship, you have the inventory, and then they lose a lot of money. That is why industry 4.0 is very important. We've invested in health, and health is a vertical that we've been working on in the last years. Now we have an expansion of this new design for hospitals. We are not aiming at high complex hospitals. We are aiming at small and medium hospitals that are part of this chain in Brazil as a whole. It's a new vertical. The nice thing about it, we have to apply AI to concrete cases, for example, construction. We have as a main goal to help the construction worker to be more productive on a daily basis.
We don't want to create new interfaces. We don't want the user to leave and then go to a new prompt. We want him to be more productive. Like we have the construction, the distribution, like a CRM of documents. And now you can manage your delivery with the AI and understand what makes sense and then what you put in a warehouse with products that have a better turnover. Another thing that is nice here, Dennis showed the three specialties. There are many opportunities when we have just one thing, for example, e-commerce. When it's very well integrated, for example, you want to where you want to ship from, how you do the rates. When you have shelf in the shelf life, so many things that are interesting.
Another thing that we are investing in, which another company that we've acquired, which is Madeira Madeira, we are working in the capabilities for the branches because it's a new way of selling. It's a new way of doing it, of implementing it, and also a new way of having a relationship. It has brought new capability, and we've been doing this since January with a lot of success. I want to bring something here, something that is feasible, some examples of AI that we've shown. A very important takeaway here is this simple functionality of reading documents. Gustavo's team worked for the whole platform of TOTVS. What is the main secret here? I don't need to have anyone thinking of a process that has to be validated, like has to validate any kind of document.
When we work with DTA, it will enable us to have all the journeys regarding information that comes from the customer. There will be an option of drag and drop, and then you put it in here. From that time on, everything that the ERP controls and all the inputs of the data, DTA is providing this. If you have a check-in for the hospital, it will be the same component. If it does not need to create a new one, we have a scale that is very nice. Here for the Winthor, it is another nice thing. People worked here by creating the image component. It is the same idea. For example, a component that sees the documents does not understand what is a can of Coke. It is a different learning here. Now we have this into the product.
Any thought person who wants to recognize an image, the component has already been created. It is in our internal DTA. We have shown here the external one, but now we have the internal one, and now we will have the solutions. Another interesting thing here is that this tool combines the best scenario for a TOTVS customer. It has all the information. Who has sold, who has bought, what do you need to sell? It brings all the intelligence part for a real-time analysis. You can recommend, okay, you sell lubricant, like grease for the cars. You would go from A to B, but on the way, there is a new store that you have never stopped at to buy a product.
He goes to the base, and then there is all the data from the external base, and the customer will see all the opportunities for the sale conversion. He will not have to leave this interface. We have the costs here. We redesigned the whole cost journey. You see the bars here. The whole process before, the customer or the user would have to go to many product screens, like he would remove things from one place to the other. Now it is very important, and the closing will be faster. Here we have the general information. You can get the data, and the DTA will help you with it. You will be able to get this, put this into another report. We have the voice commands here. The most important point here that is probably in your mind is, how does TOTVS see this tax reform?
It's an opportunity. It's not a threat. This tax reform will last for many years, and all the companies, all the software companies will have to keep this old compliance for the new and the old one. This accounting, like the fiscal and accounting part, are very big in our country. We will have to stop with the old and then start working on the new one, but you have to keep both products. Most of the cities in Brazil were covered by our products. We are following everything, but the customer has to issue invoices. The operation continues. That's what we've been doing here. That's our effort. Also, we change the way we do, like we ship everything.
For example, there are things that we will be using only in December, but we are trying to bring all the customers to accelerate the adoption of this tax reform. Okay, the customer says, "It's only January, but then November, I will call someone from TOTVS or the support or any kind of service, and we'll hire that." No, it will not work like this. We are trying to anticipate this process because there are some decisions that change the business model. For example, the distribution center that is spotted in another state will not make any sense to him anymore. He will have to change everything because of the tax planning or any kind of advanced accounting and the service that we have created. We've worked on the hiring services for the hiring of the services.
He can, for example, hire some specific things that he will have a long-term deadline, and TOTVS will help him through the whole reform process, and he will have the support every time there is a new tax. We have not done any kind of a rupture where the customer will have to start all over again, but he can also hire a primary service that is closer to his operation. We can help with the customizations that will fit into his products. We have to surround him and use this as an opportunity. We really believe that it will help us with our presence in our country, and we will reinforce that we are a trust advisor and help him with this navigation and not bring any kind of negative impact on his business. That is why we are anticipating everything that we can. That is the takeaway message.
Now it's Avelar. Over the years, we evolved to a complete solution, going through all clients' steps, starting with marketing. Today, we have a pre-sales product and sales product, exact sales, and RD CRM. In the engagement steps, we have the RD Conversation, very focused to WhatsApp, evolving to a conversational platform. We have the Lexus, the e-commerce integrator with Shopify, the largest e-commerce platform in the world. Of course, with AI in everything. I will talk a little bit about this, but basically, it's present in all of our products. Marcelo mentioned a spoiler of what I'm presenting here, but basically, a layer, a surface layer that it was called front office, very close to the client. We have marketing automation and CRM solutions. Basically, they organize the client sales process oriented to the client. Okay?
In a deep layer in the back office solutions, we have a universe of business, a price history, inventory data, transactional data important when we combine them with the upper layers. I will comment the facility when we segment the RD solutions, but in every segment, we have business data in this back office layer solutions. When we combine with the upper layer, when we bring it to the surface and give to a salesperson, we create unique competitive advantages, decreasing this border between back office and front office. We are just in the beginning of what we can do, in the beginning of capturing value in terms of we have in representativeness of revenue when we combine back office, CRM, and automation marketing solutions.
In spite of being in the beginning of this value capture in this RD and management, but in other fronts, you have already evolved. RD is not a single product solution. We have it in all journeys. We are evolving with our products with a higher ICP. Danny showed the pyramid for SMB, for TOTVS, but the RD is the same, a little bit different. We call SMB a smaller client than the TOTVS SMB, and now we have representativeness in our client base in revenue. It is a bigger sector now. What is really cool is the evolution we had with AR, the distribution channel for RD. We opened the day, and 30 minutes later, I had my people from my team celebrating sales being closed with RD products. We are already capturing a lot of value in this sales channel, distributing CRM engagement with RD.
Basically, today, we have a great amount of products, proprietary and partners' offers, all this model that evolved to serve a higher number of clients in this one-for-all model, online and offline, regardless of the presence the client has. How are we going to evolve in this timeline? We will get more opportunities when we look at the segment in a specific way, when we meet the specific needs of a certain segment with our marketing, CRM, and services solutions. Educational example. The entire journey happens around the student. You price according to the student. The sales process is the renewal of a subscription. My standard software does not meet the needs of the education segment. It needs, it requires a flavor that is inherent to that segment.
We will work hard to combine both things: a one-for-all platform with a proper template, with the segmentation meeting the needs and opportunities that many times we left behind. Going forward and giving examples of a case, this is not future. The segmented opportunities are real. We have relevant cases working. We have a dealer that we integrated the ERP. We brought lots of ERP data, including the traffic department, and I have a conversational integration between the salesperson and the client, bringing information, data available in the traffic department and what the dealer has using our product as a background. This is the opportunity when we have a service layer for the segments and products that evolve for templates that will be more integrated to the back office solutions that TOTVS provides. Now, concerning AI, this graph tries to answer how we see the AI evolution within RD.
We have four stages, very similar to what Gustavo presented. In management, he put the functional, but we have agents that are very important. In RD, we have previous steps that are very history. I don't know if you watched Dennis's presentation, but in RD, we have a very important mission that is to transform every user into a super user. It's a different; it is a medium or heavy user, users that use more advanced models. We will make AI to reduce this gap, to deliver more value to our clients. Step one is AI as a resource, several features that will deliver more accuracy to our users. Step two is the copilot, where we start to give recommendations for the use of our products. We get this interface. We have the agents. We mentioned, we talked about this.
We increase the confidence we have in AI, and we deliver, we give to AI human tasks and the autonomous agents. The last stage, agents capable of controlling other agents. We deliver the complete operation. It is soon the inversion in the use. Today, the user does a good use of a product to generate more leads or sales. Suppose we can invert this logic and with a target of sales or visits. We create the content, the channel that will be used to talk to the client. We deliver this lead. We execute all the steps to reach that objective. This is the stage four. This is a list of everything we have with AI within our RD. In pink, we have some good examples of AI in stages two and three. Most of them are as a resource.
An example, we record a call between a salesperson and clients, and we deliver a report with sales pitch adjustments with the entire history of the salesperson with the client. We can recommend the next steps, and we execute them. We start to enter into the copilot and agent order. We have good examples of stage three with service agents replacing steps in the sales process. AI in RD is a reality evolving very fast. We have new launches almost every week. It is hard to close a month without a new AI feature. Two very simple videos. At the left, a video of our pre-sales, virtual pre-sales salesperson being executed by an AI agent interacting with the client and putting in this salesperson agenda a meeting between the two. What a pre-sales would do to feed the agenda of the salesperson?
On the right, with an organized database of our client, this is the type of problem we solve. We train our service agent in several different contexts to behave as a human in the client's interface. Today, it is a reality being used by several clients. Talking about the future, I don't know if you are familiar with the way we commercialize our products, but we have basically our products for the station market, SCE, and we have all the plans. As our client evolves in the use of the product or increasing the size of the company, he goes to more advanced plans with AI. We will revert this logic and have one product for every client. It's a logic where we start to organize in plans depending on the needs of a user with a prompt interface.
I will mold the product to the client's needs with all the features. When you match this with what I mentioned with the AI evolution stage, it's clear where we can get to. We can go from a prompt interface that understands the user need. It operates with an agent's interface. We've been working for some time to capture this opportunity, and this is part of the competitive differential that sustains the concept of RD today. That's it. Now we have Mauro.
Olá, pessoal.
Hello, everyone. Good morning.
Queria falar pouco aqui hoje para vocês. É o que está sendo construído hoje.
I will talk about what is being built today in tech scene, things that have already been built, but will impact in a very positive way on our customers' lives.
The way we will change the way our customers access banking credit products or payment, as well as the receivings. We are speaking of the ERP banking. You have the ERP banking, and you have heard a lot about it recently. It's the symbiotic integration between banks and ERP. In a way, that will improve a lot the efficiency and the operational part of our customers because they will be executing a bank inside the ERP. There will be no need of the creation of some things, for example, leaving the ERP and searching for financial solutions in an app or on the website of the bank. It's an efficient operational journey, very efficient. Besides that, the ERP banking is not only the journey itself. It's the usage of data. And not data to give the best credit.
Okay, yes, it is not only for the granting of it or in order to reduce losses. Also for in order to have a better level of it. But imagine if I have information access authorized by the customers, and I know the rate that the customer has for that loan, the due date for that loan, the volume. Eu tenho, I have a competitive differential in my offer. I can get to the customer with the right rate at the right time, with the right volume, in an assertive way. That is the ERP banking. The beginning here of this construction, we had to talk to the Anel, the financial. It was a very different journey from what we call contextualized journey. You will hear a lot of these words, those two words. And what is the difference between this evolution, the one that we have now at Techfin?
When we spoke of financial dashboard, it was the access that we have into ERP, a place where we would have all those payment, receiving, and credit transactions. In this new journey, it's different. We are in the ERP. We are there when the customer uses a certain product solution, then he will be there operating the financial products. For that, it's important to understand very well the persona that is using Datasul, Protheus, or any other to understand where the financial product offers will be into the ERPs. It has to be centralized in the customer. Even more, we have to be very agile to do our findings and react fast towards all the corrections of it.
When we have this symbiosis between bank and ERP, you need to be able to have a very highly scalable process because I will have many ERPs connected to Techfin in TOTVS. For that, I need to serve all the changes in a very fast, safe way and at a low cost. Everything has to be aligned within this value proposal. All the stakeholders, our customers, our suppliers, the stakeholders, investors, everything needs to understand exactly what the goals are because it is an integrated solution in a simple, precise, easy, efficient journey according to the perspective of cost because it is functional. Of course, we will save time and the costs, and we will have a transformational journey for our customers. That's what the ERP banking is.
Here there is an important point that differentiates ERP banking in this business model from others. When we speak of ERP banking, we are speaking of an acquisition cost that is low. Okay, let's think of TOTVS. I have today 850 salespeople getting in our base, the opportunities. Inside the Techfin, I have only 30 people. My acquisition cost is very low in this ERP banking process. It happens because the highest cost that I have for the financial system is the acquisition of customers. Additionally to that, I have a journey that gives value to my customers are charged more. The ones that have been following us in the supplier know that we have a reward for an efficient journey. That is what ERP banking is. It is the supplier.
That's what we will do with our customers, and we will have a reward for that. We will have a low default. When you have these things together, you have very good profitability. Today we have here at Techfin two business units, one that we call BU TOTVS. That's where we will have all this ERP banking process integrated to the ERP product, Datasul, Winthor, among others. We have the other that is called BU Multichannel. If you know our supplier, you know that we are a diagnostic company. I have products for other ERPs through this multichannel platform.
[Foreign language]
speaking of this new model,
[Foreign language]
we have here that like through this one year and a half of Techfin, we see that we have a new systemic architecture that has to be very efficient. We also have this contextualized journey, which is one that is integrated into the customer ERP daily basis, not a different financial system. We will have several products initiated inside our ERP, starting with the Protheus. We have more than 15 products among like credit, receiving, payment into the journey of the customer. The credit products that were developed by ourselves in Techfin, like those are the core ones for us. All the credit products will be developed by us. Payment and receiving products will be through banks.
We will look for the best services in the banks to offer them to our customers. In a near future, we will be able to have those payment and receiving products in our system. We have here at a certain time, not very far, when a Protheus customer opens the system, he will have integrated a digital account with some credit limits that will be approved, credit products for payment anticipation, for bills, for PIX, for shares. Like PIX, TEDs. There will be a range of products that are inside his ERP available for him, a bank inside the ERP. The customer will be looking at all the payments now in this screen, and he will be able to parcelar, simulate, to share, to do everything.
If he's looking at everything that was received, he will be able to choose all the receivables and anticipate and liquidate everything that will be conciliated in his ERP. Now we don't have it. It doesn't exist. There isn't anything like that in the Brazilian system. It's a very important innovation. It's a transformation. It's a new way of dealing with financial products. What is the ambition here? Who are we doing this for? When we speak of TOTVS, we have here the four main ERPs, again, Protheus, Datasul, Winthor. That's where we will have, and that's where we'll have the ERP banking process deployed. We also have customers that we don't call them full ERP, but those are customers that have some TOTVS products. We'll have financial solutions for this kind of customers. What about the new acquisitions?
For the RD, that has digital journeys, and we'll have financial solutions for those journeys. When we look at the TOTVS BU, when we look at the multichannel one, that's where it's with the supplies, we will also have a very important growth through some agreements that we've been building, especially with Itaú, Itaú Bank, for the credit granting, distribution of suppliers in Itaú's base. This opportunity of distributing this kind of product can take this to a higher level to the supplier, which is our multichannel BU, 100 companies, 110. Those are the biggest ones in our country, and we can reach 300 or 400 companies very fast through distribution of this product in Itaú BBA, besides long-term products that we've just launched.
For the agribusiness, that today represents 20% of all the portfolio of our Techfin, and it has a growth potential with the arrival of Itaú. Speaking of the numbers here, the ones that we had in 2024, we had a transaction of BRL 12 billion, BRL 13 billion in payments, and almost 1.4 million customers with approved credit. It's only the beginning.
[Foreign language]
What are the solutions that today in TOTVS BU are already available for our customers? This working capital, any customer today can have this working capital, our product that is called More Business, that is the supply product applied to TOTVS basis, and then we can approve our customers' credits and sell more receivables anticipation and integrated solutions as well as hybrid slip. It's already available. It's already happening. I will close here with a testimony of a customer.
I think I'll check if I can rewind it.
[Foreign language]
No, it's not possible to rewind it.
[Foreign language]
I'll see if I can do it.
[Foreign language]
No sound.
[Foreign language]
There is no sound, right?
[Foreign language]
Okay, it's a testimony of one of our customers who uses PIX, and here he mentions about the importance of having PIX deployed and conciliated and the value that this kind of integration of a payment product in the ERP brings to him.
[Foreign language]
We see here that he is very happy.
[Foreign language]
Thank you. That's all.
[Foreign language]
Now I will call Apêndino.
[Foreign language]
Competing with lunchtime, with this cold weather. I'm hungry, and I promise I will be brief. I had to put on my winter jacket. This is a history of TOTVS distribution and some themes related to the takeaway that Dennis mentioned. In the last Investor Days, we presented what we've been doing with distribution. I have a very strong RR increase. Very few people believed that we could grow in RD solutions, and this growth rate is really strong. We also went through a consolidation in our franchise network. We were very successful in this process. Very few lost, and we consolidated to grow. We expanded our portfolio, and we are in a continuous reduction process of the CCO and the NPS involvement. We will go into details on it.
In the last Investor Days, we presented what we are doing with the data science to prepare our clients' journey to receive AI. We are using AI in our commercial area. Much of what Marcelo mentioned is used in-house for our commercial team, and we also invested in the data science part. We started with a tool called Empodera. It's a complete kit of solutions and entire white space. As we say, without data, there is no AI. We started in 2020, 2021, with a data area very profound in the distribution areas to build this database. We went to the management, tweaked it and the take rate and increase of shared offering. We also advanced as of 2020 with the digital delivery that reduced the cost. Much of our service is done remotely.
We created Trust Advisor to take to our clients our three business units. Our distribution area works with the clients with RD management and also ERP banking solutions. Our distribution today is multi-BU. We explored our distribution, and we expanded it to meet all, to serve all business units at TOTVS. With this, we opened 19 new sales offices in Brazil. We had a big presence in Brazil, but we opened new ones. We increased 1.7% for cloud sales, 1.7% for HR sales. We increased in support. We called client support a level of support with more sophisticated clients. As of 2022, we increased 8 percentage points our NPS, so our recommendation index of our clients. Now we are with a new process in implementation service and sales to include AI in all internal process at TOTVS. We have an incredible scale gain.
Much of what we build in sales areas in TOTVS, we invest, and we are rapidly copied, but we advance a lot in the creation, access to the clients, building of assets, creation of automatic and final presentation totally automated. This is a scale gain in internal areas of TOTVS. What we see with this is are early going to dismiss people. No, we can serve a higher number of clients. This is a series of activities we are achieving, integrating AI in the commercial TOTVS area. I do not know if you were to speak with me, but Dennis mentioned we have a giant opportunity to take our clients to our cloud. A very big opportunity. A giant opportunity. Many people ask for SAP. They make a comparison with the global competitor, SAP. What is the difference between our cloud models? I do not need to re-implement the system.
The SAP clients need to re-implement their system from scratch. In our models, the client adopts. It's a simplified adoption model. TOTVS model will allow. Imagine the client will only be able to access AI agents with the clients being in the cloud. We'll have a very rapid adoption of cloud. It's a giant opportunity for TOTVS in terms of cloud transition. What we also see is that the NPS is increasing day after day. Since 2018, our NPS increased more than 40 points. It's very nice to see this NPS increase. It's a process of recommendation from our clients, and it brings opportunity. Only 0.6% of our clients has the application management services. This application management services for the client when we manage the so-called new clients, we manage the upgrading versions and so forth.
Only 1.4% of our client base has the private support, the highest level, 24 by 7. It really brings a high potential. Marcelo mentioned this, Gustavo as well. We have a huge potential to sell these solutions. An important factor, one client with Prime has a higher NPS than the current TOTVS NPS. We have a sales opportunity and opportunity of increasing NPS with this AMS and the Prime support, added value service that we give our clients. AMS and Prime support is a huge increase of growth opportunity.
[Foreign language]
A little bit about take away. Many people. Take away do clientes subir. What makes this percentage to increase this percentage of the client experience to go into TOTVS product?
Every time we increment our portfolio, I add cloud, Prime services, HR solutions, HE, XM, and AMS, and vertical solutions, our take away tends to increase. The cross-sale opportunity, if you look to percentages, is still very high. Even in those clients, we have a very low %. This cross-sale window with additional solutions is very big. This increases the take away when we bring this new portfolio to clients. Except for solutions that are vertical, all the other solutions, they are in all clients across the board. All clients can buy an HR solution or a cloud solution. These are solutions we have a massive opportunity in our client base. I'm not bringing it, but Dennis mentioned the potential of the addressable market with AI. If I have a take away level to expand without looking at the AI agents opportunity.
Our solutions are very high potential because we see the expansion of the client base with the solutions we have today. When I have the AI solutions available for client acquisition, provided the client is in the cloud, we expand even more. It is a huge opportunity for us. A little bit about our distribution.
[Foreign language]
Is our distribution design today. We have a specific unit that serves the large enterprises. Yes, we do have clients with a revenue of BRL 20 billion or BRL 45 billion or even BRL 90 billion of revenue. We have clients across this field, but our great focus are the SMBs. We have a specific unity called large enterprise. We have the SMB units serving the clients in the digital and field services way.
We are using the digital machine to also serve the SMB clients digitally. This is gaining a great scale. I always like to remember that we have specific units that serve the public sector. In the case of Brazil, we have operations for the public service, basically indirect management. As I mentioned, Sebrae is one of the big clients. They run TOTVS, DataPrev runs TOTVS solution. It is a public sector. The S system runs TOTVS. We have a vertical for public service. We have a vertical for healthcare serving big accounts to mid-cost. Rigidor is a TOTVS client starting with HR with 100,000 employees. Rigidor is a very big client in the healthcare sector. We have thousands of healthcare clients. Another vertical serves wholesale and supermarkets. The largest supermarkets and wholesale firms use TOTVS solutions, almost 500 clients. It is a specific vertical.
We have a vertical serving hotels and also a Latin vertical. We have offices in Mexico, Colombia, Argentina. We expanded with a franchise in Paraguay and Bolivia. A vertical operation with all this digital and field sales model. When we see an opportunity to make it vertical, we will do to expand our depth in the market. We verticalize healthcare. It works in supermarkets. It works. This is distribution AI. As I mentioned, we have AI being applied in distribution and internal operational costs. AI is being used in-house, in service area. Everything that we are building with AI, we are building in a franchise model to apply AI and increase productivity in the implementation. AI, what we are doing via AI, a proposal of AI, cobrindo todo o AI space, documentação do white space feito via AI, benchmarking of services and TCO reductions.
We have things that work. We have some initiatives that are functioning. We have some initiatives that are functioning and working. And some that we gave up. We are using AI because we will have a disruption in terms of sales and services team. This has a scale gain in the mid and long term.
[Foreign language]
We are getting to the end from the beginning. We believe that people matter more than ever because artificial intelligence and all the opportunities, disruption that it brings to us will demand much more collaboration. Our peers have mentioned some examples of initiatives that we had in TOTVS, but with collaboration among our business units, but among our teams as well.
We believe that in TOTVS, its AI is much less than robot and much more than about humans. People have to be in the core of it because we've had lots of information, data, and it brings challenges to us, which is crucial, and we need to manage a paradox of focus and distraction because the main value here is our most valuable resource, which is our attention, people's attention. That's why we believe that as the human area, we need to train our people to be prepared to create those likely futures. As the largest technological company, to train our customers and make them ready in order to capture the value. Here to address this challenge, we have four focuses. First, to increase our people's proficiency in data and AI. We have lots of laboratories, tracks, assessments.
We have a big investment in people's qualification with this gaze that people should share it in order to scale and prepare our leadership because when we do it this time, it will be very different. We need now to encourage everyone like the experiment of it. We believe that our leadership should encourage the curiosity as well as the efficiency of our people and invite them to automatize everything that is possible to be automatized, to have more time, to be more strategic. Three, we need to be prepared for this future. We've had an evolution in our values and propositions and to be more connected. The AI and the human intelligence will work together.
Last but not the least, we believe that we need to be more, actually, as we invest in people's qualification, we need to be a desirable company that has a relevant brand and people want to be there. It will demand new success KPIs, and we should have other measures. We will be able to measure all the increase of it. Here, we believe that innovative performance is based on data, collaboration, creativity, but very important empathy. We have to train people to have critical thinking, our leadership, to have ethics, and the responsible behavior that will take to sustainable outcomes are the core of our capabilities. We have a collaborative methodology to work with those four pillars, a methodology that is much less about, actually, it is not about humanoids, it is about humans.
We have to make them concentrate, but as well as to have innovation. We have not started now. We have made lots of efforts. For example, in data, AI data, we had a number of qualifications that has more than 10,000 hours, lots of content. We also have what we call G days that are qualified people, trained people in order to train other people regarding these topics. We have mentorships for that. We also have used many platforms to increase the impact on our workers. When we think of a leadership that is for AI, we have worked a lot with content in a program that we call Next, which belongs to TOTVS and our workers share their knowledge. We also bring some market reference.
All the summits are there, shared with our people in our web summit or even the travel that some people had made to China. We need to blend this from inside to outside, and then we can make this feasible. There were more than 5,000 hours of leadership qualification, more than 10 contents created for them, and we were able to train more than 2,000 people. We also worked with our value and skills, and we have a metric that is very important to share with you that comes from our performance assessment. 40% of our people have in their PDIs, their individual development plan that is focused on the increase of their qualification as well with the job. They wanted to deploy what they've been learning, and 82% of our people at least meet our knowledge technological target that makes our business feasible.
Today, in our internal team, with Gustavo's team, we have created more than 18 agents in-house, but we still have many others ongoing. We also worked with the increase of our brand relevance. The most important indicator that I can share here with you is our internship program. This year, we had more than 91% of enrollments. If we really believe in the education as well as the knowledge application, we will have much more engagement. We also believe that to do this, we need to be contemporary. That is why we had this evolution of our mission. That is why we believe that we have to evolve people and business, and we need to have people as everything. Our customer is our core, and we really need to innovate together.
We believe in human intelligence, much more than AI, but it will be done with collaboration so that we can be sustainable. It is good if it is good for everyone. Here, going to the end, I will use those final minutes to talk about an investment that goes beyond the financial return, which is the IOS. That is a social opportunity that turns people's life into opportunities. We have been training people for more than 27 years in administration management with the excessive use of technology. Then we do the production inclusion. We generate opportunities, future, increase the competition in our country. The focus here is on training young people on AI, all the contents that have been demanded by the companies themselves. More than 47,000 people have been with us in those last 27 years with the increase of income.
We are speaking of families that live in a very poor condition, and they receive in our institution opportunity of qualification. We can offer this job opportunity, which really changes the game. We have several international recognitions, the Doty Good, which shows the seriousness of our institution. At last, I would like to invite here to the stage to talk about the scalability with who is Isabella.
[Foreign language]
T hank you, Vivian.
[Foreign language]
Okay, so IOS goes beyond the qualification. We have the role of finishing with this poverty cycle. How do we do that?
By bringing those young ones to the job market, doing this inclusion and doing this social mobility through those 27 years of IOS, we put more than 14,000 young people between 14 and 29 years old, and we are giving opportunities to them because many of the times they are the first ones to have the formal job. It has this impact on the income of the family, 59%. When they are qualified, we have like 65% of those people who are graduated, we already give them a job, and it happens because of the partnership that we have with the private sector. We can say that most of our partners are TOTVS, Dell, Localiza, Renner, Pernambucanas. We are in every sector, and it is just possible because we work with this partnership with the public and private sectors.
I would like to invite you to learn more about this institution. When we started this employment work with those youngsters, they generate more than BRL 22 million a year. I would like you to get to learn more about it. There is a specific family I will invite, Katia Melo, a former student. She also works at TOTVS, and I will also invite Gabriel and Fernanda, and they are IOS students, and you will see the reality very close to you. Good afternoon. I am very excited. I was from the second group of IOS 26 years ago. It did not seem to be feasible, and IOS gave me this opportunity. I took the course, and TOTVS gave me the opportunity to work 25 years ago. I am very proud to share this. IOS did not change only my life. It changed the life of my whole family.
Now I can give a future to everyone in my family. Today, I'm a leader. I'm a marketing manager. I have more than 40 people in my team, and the IOS and TOTVS gave me the opportunity to be where I am today. I'd like to leave a message here to really touch you, to give opportunities to those young ones as I had this opportunity, the chances. You should also give this different future to all those people. That's all. You can generate opportunities, and to close this up, for us, as human relations is about business, but as well as about innovation.
Let's play a video that will show what we did with the product team, as well with the idea team, and it will be used in-house and for the IOS students to help us to make our PDIs be more assertive for people and for really, so that people can understand what this is, this part is, all the history that we have. Like 360 degrees assessment, and then it will create a summary of the journey and will bring recommendations for the development of people. And even check players to give other recommendations based on the plan that was created. Human resources is also about business, innovation, and investment in people, which for us is something very, very good. That's how I finished my speech, and I will start here, Cesar. I will invite Cesar.
Okay, so you can stay here and then invite all the executives to be together with us here on the stage, and then we'll start our Q&A session.
[Foreign language]
In the meantime, I will just reinforce something nice about this event. The box that you are receiving today as lunch, there are two nice things. First, the box is from TOTVS customers. So inside the box, all the products are from TOTVS customers. And secondly, there is a voucher inside the box that is to go to our gift shop and buy something, and then you can help with everything that Vivian has already presented here. Let's start our Q&A session. We have some virtual questions, and we have some raised hands here. I think Leo was the first. Let's start with Leo, and then Marcelo and Maria Clara.
Hello everyone. Good afternoon. Thank you. Thank you for this event, for the invitation. It's very, very good, as always. There are many nice things to discuss here. I think there are many good things, but I will focus on one thing. I was surprised by the tech thing and mainly something that we talked about last year. About the possibility of distributing the supplier products to the customers of Itaú BBA. First, if you can tell us if it's a novelty and how it would work and give a general overview about the governance and the integration with Itaú BBA. Thank you. Thank you, Leo. Okay, let's start from the end. Governance is very good. I've said that every time someone asks me. Yes, of course, that we would think of something very good. Otherwise, we wouldn't have this connection. So it's even better than what we have planned.
Everything is working very well for governance, the alignment, all the visions. Everything. Very good. Any kind of concern that we could have before the creation of it, like about the competition, no, there is no competition. It is totally self-independent. It is free to choose the paths. All the commitments for funding are being filled fully and bringing advantages, knowledge, attention, know-how. Everything is going pretty well. About the distribution of the supplier products from Itaú, it is a possibility. We had already talked about it in other opportunities, and it is still advancing. The supplier's product is a unique one in the market. There is not any direct competition. Itaú does not have this offer in its portfolio, and it has shown an interest, and it is something that we have been working with them. It has not been concluded, but it is a possibility, of course.
It was not apart from the original scope, but it came up as a demand and opportunity as Itaú does not have this product in its portfolio. They see a demand, a need for that. Marcelo. Boa tarde a todos, Marcelo Santos de P. Morgan. Obrigado pelo evento. Thank you for the event. I would like to talk about those circles of performance of the AI agents. What would be the main ones that you would say that they are more important? What is the easiest part of it? I will give you the frame, but not functionality itself, because there are hundreds of tasks. The frame here is the one that more complex techs and more critical ones are the last ones that will have a higher level of autonomy, and they will be the last ones to generate the highest value to the customer.
Because on one side, this complexity brings a challenge of accuracy or even the processing cost for that task, but mainly because of the criticity of this assignment for error, even though the accuracy is 99%, depending on the topic, 99% is not good enough. Even when you do not accept the error, for example, but the cost of it is not feasible for TCO that you will have for such a long time. The opposite is the same. Any kind of test that is simpler and less critical will give you some ability to test this autonomy faster. Gustavo Bastos's human resource product uses this concrete example of a payroll. That is where 99% of right things. If 1% of this payroll was wrong, then we will have a very bad problem here.
We showed concrete examples here when I'm interacting with my schedule to release a payment that is in the flow here. You will have a low-ranking fruit that will be very clear, and it will be totally autonomous. Maria Clara from Itaú. Congratulations for the event. My question is for you. I'd like to talk about the tax rate that you showed us. Can you give us more details about the profile of the customer, if there is a standard for percentile, or if you see a curve through the years, and then doing a follow-up regarding the AI agents? How can we think of this deployment for the curve of customers, for example? There isn't one standard. There are some features that will make customers go further in this take rate process. For example, the level of sophistication will make the difference.
A customer who has a level of higher sophistication within his management, like more robust processes, a company that has been formalized beforehand and has a higher level of organization, it helps, for sure. On the other hand, when you have companies that are closer to what we call consumer-facing, we will probably have this disruption level because of technology, and it will make them invest more in technology. Another important element is that we should have done a deployment in the customer that is something that would be very, very good. TOTVS is not perfect. The deployment process can be a little bit stressed, can bring some stress. This curve will increase this cross and upsell. The better, the more organized the process was, then it will prepare much more the customer.
There are many elements that will help with the definition about the customer. If the customer will be in that level of 1% or something similar to that, or not in the 0.20. And Apendino, where is Apendino here? Apendino brought some concrete elements here. A customer that has already migrated to the cloud advances very fast in the take rate, and he enables, activates many things. The cloud has been a very important product, but also an enabler. When we speak of recurring services, IMS, the Prime, all those products are somehow big enablers in our customer. They open the doors so we can offer other solutions, as human resources, for every specific segment. There is a set of features that make the customers have a higher take rate. For the AI, I think it's the same step.
Again, as Apêndino said, if the customer doesn't have the last version of the system, the chances of having agents operating and bringing benefits will be zero. We will not allow this to happen because we want to have the control of the entire environment. It doesn't mean that it will be only us who will provide agents to the client. No, our vision is different, but the client needs to be in the environment we know. Otherwise, we cannot ensure security and governance of what is happening with the client. Okay. Your question now? Yeah? Thank you. Related to this ecosystem and in this discussion of being a platform of final user, you are developing many AI tools using the ERP data that are unique because this environment will be very accurate for this.
When you have to develop a tool versus to open the ERP dataset to other companies, how do you see this? How do you trace this line? Thank you. First of all, not necessarily we have all the answers to the questions, and even those that we believe we have already, maybe in the near future, we decide that we do not have. If you see what was our understanding of AI last year in the universe of last year, it seems that it was five years ago, but it was not. It evolved a lot and changed a lot. Again, this disclaimer is valid. Since it is a new thing, we do not know everything necessarily. TOTVS has a history. The TOTVS DNA is open. It is an ecosystem DNA. We have the distribution, the franchises. We have 150 business partners.
We have this tradition, and I do not think AI will change anything. What changes is what I said. We want to have more visibility, telemetrics, and control of everything that is happening. We want to know what is accessing what, and we want to measure what is being accessed because TAS that will come together with SaaS in many situations, the TAS will be linked to consumption elements. For you to charge this, you need to measure. You need to know what is happening. I think that the border does not change. We will continue with this open position, attitude, this attitude that we are not able to develop everything alone. Every task is subjected to become a specific agent. You have a very long agent, so to say, and you have a very specific agent. We are talking about thousands of agents over years.
We alone cannot develop this. Again, we will have this attitude, but what we need to do is to advance in this pathway of visibility and control. Control in the sense of knowing what is happening to be able to measure. Control is integrity, integrity of the data journey, and to avoid putting several agents. If you do not register correctly your invoicing, you will not have integrity. It is not to have a closed system. No. We have several partners, and we will add more partners. Orchestration is very important, not only because of security, but capacity. If you do not know what you are doing, you do not have capacity to meet a demand. Now, Luca, next question. Thank you, and congratulations. I have two questions. Can you give more details of AI?
You mentioned that it can be a market 16 times greater than managed, but how much can you acquire of this market? What is the timeline? How long will it take to reach the maturity you expect? Second, you mentioned the TOTVS differentials in AI, data access, investments, and several other points. Okay for the locals, but when you have the big international competitors, what are the differentials in relation to these AI players? Wonderful. Okay. First, no, we do not have any type of concrete roadmap for numbers or when we will be able or how much of this addressable market we will be able to gain. We just gave a context about the size of it. We have never imagined before AI.
I myself, I have never imagined that I will look to the BPO market and say, "Wow, maybe this is a market for TOTVS." Just the opposite, I said. This market has a totally different dynamic. It does not have the same scalability logic of a software. Today, with AI, with the ERP, I, yes, see a possibility of some time, maybe six months or six years ahead. I see the possibility of having many of the BPO or HR tasks being done by the HR system, for instance, or even applying it to any type of ERP module. Again, I do not have any time frame in my mind. Now, concerning the competitive differentials concerning the international players, they are the same. The level of knowledge of what is an SMB, and an SMB does not follow the same dynamics and practices as a large enterprise.
This knowledge and within the Brazilian business environment is something that SAP or Oracle do not have. They have never had. If so, they could compete with us at the same level today, regardless of AI. They do not have these differentials. Likewise, the go-to-market, having a platform with the sales machine, with the capillarity we have, nothing of these companies, these competitors have. This will be mandatory for decades. In this case, AI does not change anything in terms of competitive differentials that TOTVS has relating to any global player. As Cloud did not bring it, AI does not bring any change in the competitive scenario concerning international players. Next question.
Thank you very much, and congratulations for your presentation. My question follows Luca's question.
To understand the charging AI model, how do you imagine the monetization of this AI? Will it be a direct charging, or how will the monetization of it? The second question relates to CapEx. If the deployment of this technology requires a big investment, this can bring some pressure on CapEx margin in the next years. Some impact it will have, but this is happening as we speak. We are investing in AI intensively for almost two years now. It is not that we are starting to invest now. We showed you a great amount of things that are already a reality. If you go in this space we have here, you will see things that are already available and being used by the client. Of course, as things advance, some additional investment will be made.
Anyway, we'll have, on the other side, a concrete return on these additional markets, at least today. Maybe I will change my answer in the future, but today I don't see any change in the margin profile of the company because of AI. This dynamic or technological market in a B2B market, as TOTVS operates, and in particular the management market, they are more complex. This pace is different from the AI adoption pace in consumer-facing markets. The pacing is slower for us, for good and bad. With the ability of testing more things and adjusting the course in a different way, I always make this matrix comparison. In our market, many times you see the bullet coming from your enemy gun; you see it in a different speed.
You do have time to adjust, but in other markets, you see the natural speed of the bullet, and you cannot react in many cases, but we can react. On the other hand, the speed of capturing all of this is small. So when you have 16 times the business, I don't know, it will take long. We don't have this inflection point. We didn't have that you go from 0 to 100 almost instantaneously. You have a curve that will increase and sustain this growth level for a longer period. And we are used to this, and we have an investment process, a decision-making process that takes all of this into consideration. But it does not mean that we notice any important change in this scenario. We are ready to adjust our course of action. Now, going to the final of our Q&A, we have one more question here.
Good afternoon. I'm Luiz. Eu sou o Daniel de Corrizones, CEO. Xispei, XP. The first is 50% of clients using AI, but very few in value. Why do you think what is hindering this value capture is the maturity of the client, the culture? The second question relates to Protheus and Apcel. You said 75% of sales are in this model. What is the driver of this in terms of a main product or solutions that are carrying this Protheus? I will start with the first question. When we look and you see this in your day-to-day, especially when we are talking about generative AI, people are capturing value for them and not for the organization. Not necessarily the time the person is experiencing or trying the tool they are having any productivity.
What we noticed, it was clear in our survey, was this lack of orchestration, thinking, metrics, practices. Many times people try the technology, they insert this, implement this, but they cannot extract value. What does it mean to transform something in a game? If I say I personalize a task with an agent to say it was fancy and cool, this does not bring any gain. You generate a new cost, but you did not transform this into a performance gain or process gain. This is what we saw. People trying testing in a less instructed area. People extracting value for their protein task, but not for the company, for the business. In fact, we organized this in the way we show to make this feasible, to provide an orientation, a context. Things can be used not for the sake of technology, but with a specific return. Okay?
What Gustavo said is very important. If you lo youok at the nature of all the content we bring to the universe, and this year more than any other, it is exactly this attempt to translate, to decipher, to materialize things to our client. Our client, they expected this from us. The client sees things happen, but alone, this leadership is not able to go to reach and organize the answer. Part of our role is to help the client to find this more organized pathway to extract the entire value. Now, concerning your second question, we have a big diversity, a big range of products that make this 75% of Protheus or Apcel. Cloud is a great driver.
It is still a very important offering that we have, but within this entire segment, every segment that Marcelo's structure deals with, every economy segment has a different degree of use in applications, and this is still a very important sales item. The human capital management and experience gives all the collaboration tools, and to me, one of the most important is everything that Apêndino showed as recurrent services, services that help the client to better manage the structure, the differentiated level of support. It is a wide range, and it is important for us, meaning that we do not depend on one or two items in our portfolio. Every client has a different journey, and that is why in my answer to Maria Clara, I said we have some profiles of one single profile. Depending on the segment and size of the client, the journey will be different in this Protheus.
Unfortunately, we do not have any more time for the Q&A, so we close it, and I thank you for your attendance, the attendance of investors, reinforcing the invitation. Enjoy the event. We have many, many nice things happening. Dennis, please, before reinforcing the invitation for those who are following us remotely, to be with us next year to interact with the executives and to feel what is TOTVS world. Thank you. Dennis, your last comments or final comments? Oh, just one thing. The material used here is available at the IR site. It is available with the investors' relation website. I would like to thank you for your attendance and congratulate the marketing and IR team, and reinforce this point for us. To have the Investor's Day within the universe is an opportunity to give access to TOTVS Kitchen, as I say.
You have total freedom to go wherever you want, to ask whatever you want, to whomever you want. Enjoy this opportunity. Thank you very much.