Modernizing applications and eliminating 200M+ lines of legacy code

Hello, everyone. Thanks a lot for, for coming. My name is Fabio Napoli. I'm a CTO at Itaú Unibanco. I'm responsible for, for Cloud I chair, Architecture and Operations and uh join us today. Bruno. Bruno is the CO and the founder of Zoi. Zoi is a leading technology company that was founded in 2011.

And uh today in our session, we are going to divide in two parts. The first part, we are going to talk about our digital transformation with Itaú. Our digital transformation uh including our platform modernization. And in the second part, we are going to talk about Tech Spot. Tech Spot is a cloud native platform developed by Zoi and it a u that is supporting a lot to modernize our legacy applications and move to the cloud.

So I hope you enjoy. Let's get it started.

So just just before we start talking about our transformation, I'd like to introduce Itaú. Itaú is a universal bank with a full offering of products uh and financial services for different segments of the market. Um we are the largest financial institution in Latin America and one of the word by market cap, we have the most valuable uh brand in Latin America and we are over 97 k employees.

And what is really important to emphasize is that Itaú is undergoing a huge transformation, a huge digital transformation. And the reason that we decided to move this, the reason that we decided to change is to become much more connected with our customer than we used to be.

And for us, uh what is the transformation, the digital transformation that we are undergoing? It's a mix of two aspects. The first one is future that is 100% related to the way that we work today. Uh one example is that the technology and the business are 100% integrated uh creating the products according our customer needs our customer's behavior. And the second part is the technology that the technology uh is the core of everything that we are performing right now.

Uh and how, how do we operate, how do you work? Everything that we start when you think about a product or a service for a customer, we start from the design. Uh the design is really important because we understand the customer behavior. And once we understand the the customer's behavior, we can create the journey and we can enhance the experience for our customers.

And how is possible to do that. We use the data uh with data, we we look for to understand every single interaction that the customers are performing with the bank. And based on that we create our data points, then we understand that, that our customers behavior and we use AI to support, to support that process.

And once we understand, once we have the data, we create our products and our products are according to our customer needs, our products are just solving real problems that our customers are facing. And one thing that's really important, we work, we feed the back loops all the time to really enhance the products that we are delivering for our customers and the technology, we are upgrading our platform, we are modernizing our platform. Otherwise it would be possible to work in this way to unders to really understand the customers and to create the products and the services and enhance the experience according the customer needs.

And here in terms of technology, uh how, how is, how is our journey so far? We started in 2017, modernizing some workloads and we built our private cloud and it was really to learn how to work with this new technology, how to modernize and how to prove the value that one system modernized and running in private cloud could bring for for Itaú.

In 2018, we were allowed by the central bank to use public cloud. So our decision was to get rid of our private cloud and modernize and move to the public cloud. So it was our first contract with AWS when we were really understanding how to work in this environment until 2020 when we decided to move our clothes to the cloud to modernize and move our clothes to the cloud.

At that time, we changed the way that we were operating our our environment, our infrastructure, we called cloud operating model where our focus uh it was velocity. We really wanted to provide the best tech products for the developers to create the business products for the users very fast.

And now 2023 and beyond, uh we are focused on the platform because we understand that with a platform that is Tech Spot, we can optimize our environment and Tech Spot can consolidate our tech products and our stack to simplify and optimize our environment.

And uh we are going to give more details about the platform some results that we, we reached since we started, we have more than 5.9 k AWS certifications is the largest certification program in the world since we started, we reduced it in terms of incidents, we reduced it 98% the incidents running in production.

Uh in terms of velocity, we increase it in around 13% 13 times. Uh the number of deployments that we used to, to, to do and we have around 50% of our platform already modernized. And uh what does it mean? It means 70% of competitiveness that we understand that we have to, to get so we still have too much to do to modernize and to move our clothes to the cloud.

Talking about Tech Spot. Uh we are gonna cover more, more in details, but we really understand that now that we we have maturity that we learn it since 2017, how to work, how to create our tech products, how to create our stacks. We understand that we can have a platform to consolidate everything.

Tech Spot is the platform that we are using in Tech Spot is providing much more quality, velocity and efficient, efficient in terms of cost reduction. And we are going to cover in a few slides in Tech Spot.

Here is one example how we are using in the bank. Uh and it's, it's not just a matter of a platform but that Tech Spot enables the way that we are operating. So when you see on the left of this light on the top left, we have the, our CIOs, our CIOs are fully integrated with the business, creating the products for the customers. They are 100% customer facing.

When we see in the bottom left, we have the CTOs that are creating the tech products, the products that are consumed by the CIOs to create the products to the customers. And when you go to the to the right of the slide, we have a Tech Spot. And here is an example with two different layers, one layer uh is where we have our plug in stack.

So one example here is this layer is where the CTOs are working. So when we have a product from the AWS, for example, is it true? Uh we bring we internalize to the bank, we put the graduates in terms of security in terms of reliability, in terms of uh fines. And once we internalize it, we create a plug in and this plug in is connected to the platform.

And this plug-in connected to the platform is available for the developers that they are customer facing where they work is the Tech Spot EDP. So once they need a tech product or a stack, they consume that stack that plug in. That is uh that's done for the for the engineers.

So we reduce a lot the cognitive load from the engineers that they are working uh with business solutions with business features that must be delivered. So we reduce a lot the cognitive load and this Tech Spot enables this way that we work.

And now I'm gonna call Bruno. Bruno will explain in detail how how his Tech Support works, right? It's the green. Ok.

Hi Bruno. Let's let's now take a look at uh closer looker to this tax pot platform. We have two main products here. The first one is our EDP, the EDP. It's our enterprise developer platform here is the place where the architecture teams is able to transform their technology vision into a Zachar architecture product to be consumed through a whole company.

And the developers can access the platform to, to find and use in an easy way these stacks. After this is very important too to the architecture teams set some configurations here that will be applied in the whole process. When the code getting production, this is the way that you can ensure some policies and and it's very important for security here.

The second product, it's our tax spot called services, the cloud services. It's a managed service for ideal for companies that don't want to is in transition to the cloud but has a lack of in depth knowledge how, how to do this. We work together with Itaú to create this product and we patched more than seven years helping customers to go to the cloud and not include only our AWS resources here. But the best practice and configuration the finns to escape with cost saving resilience about the application monitoring and, and to policy enforcement. The security is very important when you set up a cloud configuration for the whole company, everything here is packaged in a red shoes product and we are able to create abstraction layer to provide to the company a very fast way to the developers in focus on the customers to create business features and we take care about the cloud operations for them.

Let's delve deeper to the EDP. We have four foundation layers in our developer platform the first one is the tax pot studios. The studios is the place where the architecture team creating their products, transforming the technology vision, the the architecture can nor normally are documented. We transform into stacks here that is a execrable product can be called it by the developers.

Here has an important hold to create a strong adoption plan. It's very difficult in the companies, the architecture team talk and with the developer team and the platform is the way that we connect both. We have a good opportunity here to create a strong inner source culture in the company and bring all developers around the company to help their protect team, to create it and improve their products. This is a very, very good approach to become a more unified company about technology.

And after here, it's able to, the architecture team can are able to enter and set some policies here. When you give power to the developers created infrastructure and applications, you must control some some some configurations. For example, we don't, anyone wants that new database will be created without backup, right? And here you can set up this execrable products like a plug in that implemented RDS for example, that when the developers will consume the backup will be set. This is an important role to guarantee policies, security and quality about the how, how patterns and and guarantee that that these patterns will follow the code into the production to production

The second layer here is the workspace, the workspace, it's how we can organize different specific needs in different areas in the company. Normally the tack tackle of the team or tribe will enter here to create their, their space. Uh the workspace can be divided, for example, for in a bank uh could be an investment team. Another workspace could be account balance team or credit card team. You can organize different needs and different configurations for each area of the company here in the workspace. The the technology said each stack they want to use.

This is very important because this this will be specified when creating new applications, what stack, what technology will be used there? And the developers just need just need to enter in the in the stack spot here and create a new application using this stack available in in their in their work space here.

The workspace, the technology lead is able to add new configurations based on the conf confirmation was made by the architecture team. Uh putting additional features or suggest configuration to to make easy the creation a new application infrastructure to the developers.

The next layer, it's about the run times, run times. It's very important to getting the code to the pro to the production here. It's the layer that we can create it abstraction about the infrastructure code. For example, if you want to create a new dynamo table and this code is packaged inside the studio in a plugging. And the developer can reach can can change this code. They just use this plug in through the stack and we can able to change because here is when we protect the code to apply and enforce the security policy in production.

If they want to change the scope, they must improve the product for the whole company through the inner source process. Together with the architecture team, improve the ar the total ar the the whole architecture in the company, improve this plugging for everybody in the company, not only for my project in that case and in the wrong times, uh we are able to configure here. Uh what kind of infrastructure we need to connect?

We support three different kinds of infrastructure. The first one is AWS called accounts. The second is the the on premise environment. A lot of companies is, is is in transition to the cloud. They have both environments. It's very important, the platform support both and manage the cloud. If manage the cloud is not your preference. The best option here is this tax pot code services. We manage it for you like i said before and now the run times guarantee that the policies are configured in the stack. We will be applying production every component that we created or f pattern that will be applying this zro architecture will be applied here.

The final layer is our stack spot catalog, the API it's the best way to interact with the system to connect with other systems. And our catalog can understand when a new API will be published in the company and automatically store this information in your centralized catalog. A lot of companies has more than one API gateway and it's difficult to the developers find what is the best API in different places.

We centralize every API documentation here to be, make easy to find and access these APIs ok.

About these studios to make clear the studio can be created based on specific technologies. For example, Java studio Python studio or can be created by types of solution could be front end um mobile applications uh could be a studio for mission critical critical applications, et cetera or as a single enterprise wide studio.

The architecture team define their strategies about these studios to understand the better all the layers I will explain three explore three journeys here from the technology vision to production.

The first one is about the architecture team, the character team entering the st spot to create accounts for the whole company. And after they define the technology vision, this technology vision, we need to create a studio that will implement that technology vision.

After this studio was created, we start to split this technology vision in different pieces called it pluggings. These plugins implement the whole technology vision about that studio. For example, if we have um a microservice studio, we need a database, we create an infrastructure plugging call it um r ds. I need a table. uh I created uh uh dynamo plugging to implement a dino table for example. And we have another infrastructure plugging uh called kubernetes.

And all these three pluggings running this application after we need to go through through the application layer, and we create a new plug in called it a scaffold of microservices. And we implement it using java and spring boot for example. And after we need another plug in to implement the observable using called watch.

When you package all these pluggings, we can implement the whole uh technology vision about that studio. After this, you get these plugins to create a stack, they stack, it's a composable by different pluggings. And the stack is the zac architecture product that will be delivered for the whole company for all developers.

And each stack we can define different music cases called starter. The starter is the best way when the developer creating a new application, they fit what what they need with the best starter because they start is based on the use case that the stack can solve in this process.

The architecture team are able to enter in is able to enter in the global configuration to set up mandates here or suggest configuration. The best practice here and the developers can change after but the mandates the developer can't change because we will be guaranteed that we will be applying production when they finish this this configuration.

The this stack is ready to be consumed, consumed by the company.

The next journey here is the workspace. The tech leads of the team or tribe enter here in the st pot account to create a new workspace in this workspace. They need to add the cloud configurations about how we will connect to the aws accounts, the budget clouds or the on premise environment that will be using that workspace in that project.

After they need to choose which stack they want to use. This is important uh important point because all applications that will be created using this workspace will follow that stack that they choose here. And after he can add some additional configurations to suggest configuration for the developers to make it more easy, create new applications or mandates here too.

And after when he finished or he or she finished this process um is available for any developer in this workspace, creating the application so fast, very fast.

The last journey is about the engineer. The engineer just need to access the this tax patch account and choose their workspace that we is working now. And after they create a new application, new microservice, new mobile application, et cetera.

And after they just need to choose what is the starter from the stack that they are, they are using that workspace. And after the application will be created, the new repose store will be created with the whole code. The zacho architecture will be coded for, for, for him with all pieces implemented by each pluggings inside the stack.

After they just need to start to code the business feature, implement the business rule and when they finish, they just need just need to follow the the the pipeline, the the the pipeline, the process of the pipeline and the codes we will get in production using our runtime engine that is integrated with the pipeline to do this, applying the how our standards and policies here.

All this journey journeys can be configured very fast. The first journey uh about the architecture team, we can, can take almost one week to do this. And if, if the company doesn't know how transform the technology vision into plugins, we have a professional team to help with. It's easy. But if you don't know, we have a professional team to help you with, with this and we, we are able to access our catalog.

The customer are uh are able to access our catalog in our catalog. We have a lot of stacks implemented. We have more than, than, than, than, than than than eight years, helping customers to do mo modernization and we have a lot of stacks ready and you just need customized. We can reduce 80% of the effort to do this in this step.

The second journey takes less than two hours. It's very simple to configure a new area inside the company and you just need to configure a new workspace, a new kind of project or a line of business. It's very simple to set up with output patterns here.

And the last journey less uh it takes less than 30 minutes to create a new application. With all this architecture is is here with this. We truly believe the developers even the first day can achieve productivity because every day they need, it's ready here to bring the best developer experience for them.

And now let's take a video to show a p in action. Let's go

Well, when we launched the edp to our customers, we help a lot to improve the developer experience. All this configuration that you saw here, it's very easy to use for the developers with just one comment, all this, all this configuration will be used for them. But during the creation of the business rule, then the developer creates a new application and start to code the business rules. We don't help so much the developer in this phase.

And this is i am very excited here to to introduce uh at aws to invent the new stack spot a i that we are looking to focus and help our developers that use our platform to accelerate the business feature creation. This tax spot a i is a code suggest an agent harnesses the power of generative a i to make it possible that developers achieve the hyper productivity.

Thanks for the hyper contextualization capability of the platform. How it works. We are able to create a different kind of minds for different kind of companies. Each company has a different technology landscape and our agents are able to adapt for each kind of context.

For example, we training our agent based on a lot of technology information within the company, for example, enterprise software architecture, cloud configurations and cloud architecture. And we can training with api and even catalog software documentation source codes. Hypo is security policies.

And if eu use our edp every information defined in our stack, we will be training automatically here with this. We truly believe that we can provide not regular good suggestions but a most the very, very close to the real code that is going to the production.

We very very align our suggestion with the strategy technology strategy of the company. And it's very easy to find how you can connect with your systems and documentation and epis and bring to the to the to the to our agent.

Why st bought a i is ahead of ahead of the market. We have three main aspects here. The first one, it's everything that the developer need is inside the id. The id is the place where developers are building and evolved applications within the company. And the best place to develop a focus on their job is there to achieve more productive.

Now, the developer just need to install our agent there and hold this knowledge, all this knowledge that was trained, it's ready to be used inside the id. For example, if the developer wants to create a new, a new code that access a customer information, they just need to interact with our agent and our agent will suggest what is the best api in the company to do this.

But not only this, they help the developer to generate the code to access this information based on the stack on the whole technology vision of the company. This uh is the best way to bring every information to the developer interface here.

But complex companies has a lot of process and it is difficult to to to implement our process here. For example, some kind of uh products we need to approve budget to put in production to create a new infrastructure, how we can solve inside the id. This kind of problem we can solve using our custom quick comments.

The custo commons, we combine our specialized mind of our agent with different systems to create workflows and complex workflows if you need and we can interact with the systems and compact every this process, our process, the whole process in a single command that they can use our agent to execute this call this command.

And we can patch the whole developer experience in the company within the. And we have some be quick comments too to help the developer to create it. For example, uh documentation, the task of the code based on the stack to factor the code to keep align with the the technology vision that is is evolving to.

But the custom quick comments makes the developer more powerful. They can create custom quick comments. For example, to gather our daily tasks in the backlog systems and help you to create this the code suggestion how to fix the daily tasks.

And this avoid that the developer experience in a company, they must access a lot of systems to solve their problems, to perform their their tasks. We centralize everything here in the id.

The last point here that aspect here it's our knowledge and experience to to do the modernization, the legacy modernization. We have a lot of ears in the market help customers to do this transition to the cloud and modernize the applications and we we partner with it to improve this mind of our agent.

Our agent is ready with a deep knowledge how to do this. I am not talking to translate code from legacy to a new code. I am talking to reimagine the code because when you translate, you don't create a tru truly new api or a cloud native application and to create a cloud native application line with the stacks and technology vision, you must imagine how you organize the application and it's very easy to the developer when needs to create a new ser micro services that modernize one lega application, just talk with our agent through the id and our agent will be will find what is the best piece of the code, implement that kind of business rule and transformed to a no uh to a modern code.

And after his factory, he imagined that code to fit with the stack and technology vision of the company with this kind of approach, we can reduce a lot. We can the time the effort that we need to do the modernization.

Now you can do faster, cheaper than ever with all these aspects and the features that we have here, we believe the developer can achieve the hyper productivity and generate more value to the customer to to their cus their customers and the company can be reactive faster to the customer needs much more faster than ever.

Now, i will play a video about this tech spot but the tax spot a i but before i'm finished here, i want to talk about the com the how we can combine the ad p with st port a i. When you combine this both solution, you, you you deliver amazing developer experience for the developers, an incr an incredible uh pro productivity and you can create a company with legacy free because when you modernize an application today in the next after five years, that application will be a new legacy because the architecture, the technology vision of the company is evolving during the time.

And when you combine the ed p that we know about the evolution about the technology vision of the company. And you have our agent working together with the developer we can do is small factories and our agent can suggest the small factors in the code to keep a line with the evolution in the edp.

And when you combine this, you can create a legacy company and don't need modern modernized application after five years, after five years. Ok. Let's play.

Thank you so much. Let's play the video about the stack spot a i.

So thanks a lot, bruno, so many new features functionalities and uh uh we are working uh with zooy uh developing together this tech spot for the last two years. And uh we do believe that once the solution, the platform is working for for it a u uh it's going to work, it's gonna work for every company that you are modernizing legacy applications and moving to the cloud.

Uh and in the end of the day, every cto s, every cio s every companies are looking for efficient quality speed with security and with stacks spot in terms of efficiency. Uh here are some examples. Uh we stacks, spott ensures reusable products and the final agents with stacks spot a i is one example instead of having uh engineers looking for cost reduction, uh it was developed at agents that they are looking for how, how to reduce uh how to optimize the infrastructure all the time in terms of quality.

One example here is the stack with observable. So once the engineer uh wants to, to run a co production, the stack that's using comes with observable in terms of speed. We, we really believe that st pot is much more uh than a copilot because uh here when, when someone is developing a software, it's developed by, by, by context in terms of security, uh st spot a i was, was viewed, was designed according to ws we have accepted security framework.

So we are in the end of our presentation, we have a, we have a booth, so we invite everyone to visit uh this text spot booth to understand more and again. Thanks very much for, for your time.

Thank you.