DigitalTransformationTalk: Scaling cloud native - fuelling agility, resilience and cost control

Speaker A: Foreign.

Speaker B: Hello everyone. Good morning, good afternoon, good evening wherever you are, whatever your time zone. This is Digital Transformation Talk and I am, um, your host, Kevin Crane. Welcome to the show. Today we will look at scaling cloud native strategies to deliver agility, resilience and control. Cloud native adoption is accelerating with the 2025 Cloud Native Computing foundation survey showing that 90% of organizations report faster software delivery with cloud native approaches. Yet many struggle with the orchestration, the multi cloud complexity and scaling securely and effectively. So in today's session, we'll explore three critical dimensions to preparing for AI driven orchestration and intelligent automation. How can organizations implement automation effectively? We'll look at cross, uh, multiple cloud environments. How can companies maintain security and compliance while leveraging multiple cloud providers? And we'll look at overcoming security and cost management and some of the complexity issues with Kubernetes scaling and what strategies can help manage costs and ensure security and really make operations scale with more efficiency and effectiveness. We're going to dive in in just a moment, but first I want to say thank you to everyone attending today, and that includes the folks that are joining us live on LinkedIn. Hello, everyone, and hello everyone joining us today on Zoom. Thank you. During today's discussion. Look, we want to hear from you too, so we'd like to encourage everyone today to participate. Join in with your comments in the chat section and if you have a question along the way, just jump on in. I will attempt to get some of your questions and comments into the flow of the show in our discussion as we go. All right, let's get to our great guest today, starting with Shadi Khattab, Enterprise Container Platform Sales Manager at suse. Shadi, are you with us?

Speaker C: Yes, Kevin. Hi.

Speaker B: Hello. Hello, Shadi, where are you calling in from today?

Speaker C: Right now I'm calling actually from, uh, Dubai.

Speaker B: Dubai, Wonderful. Uh, well, it must be, what, in the middle of the night there for you, at least here. I'm on the west coast of the United States, so all the way across the world. It's great to connect with you, Shadi. Thank you so much for being with us. Also today with us is Johan Von Dine, Chief Information Security Officer at Do and Co. Johan, are you with us?

Speaker A: I think so. I think so indeed.

Speaker B: There you are, sir. Thank you for joining us. Where are you calling in from, Johan?

Speaker A: I'm calling in from, uh, cold St. Albans, of all places.

Speaker B: Oh, my goodness. All right, well, very good. Well, thank you. Thank you both for joining us today. Look, um, we have a great discussion planned, but I'd like To start our discussion today by drawing our attention to an article published recently by InfoWorld called I Built an Agentic AI System Across Multiple public cloud providers and it discusses the challenges and lessons learned from developing a decentralized AI system that operates across various cloud platforms. Public and the author highlights issues such as complexity and costs and the need for careful orchestration in a multi cloud environment. Despite the challenges, experiment demonstrates the feasibility of building scalable and resilient systems while leveraging multiple cloud providers. Uh, it's an interesting article. I'm wondering if I could get your take on it today. Shadi, what are your thoughts on the article?

Speaker C: Uh, it is really an interesting one specifically with the growth of the whole agenti use cases and uh, how the AgentIQ is transforming the whole processes. What is more interesting in this is that they are trying to overcome the multi cloud dilemma where basically um, the use case was to automate the workload and to have it deployed on the right cloud, depending on the right cloud. And that is also one of the growing uh concerns now on how you can have control on the financial aspects on the public cloud and at the same time utilizing the right technology across the multiple cloud provider. Um, throughout the topic um, it dives deeply as well into. It is easy said than being done uh, due to the different um cloud structure in each of them and selecting the right workload might not automate in the different cloud providers. However um, I think it is at the beginning of a great solution that would come at a later stage that can overcome uh, one of the problems that goes now across the organization of who is the right cloud provider that is right to my workload and where uh, I should place my workload. How can I automate my workload, how can I control the financials of my workload? So um, it is a very interesting architect, very interesting use case and indeed it will be one of the growing uh growth in the area of agentic to be smart enough to deploy directly across all the clouds uh, based on the current organization needs.

Speaker B: Shadi, you're hinting at seeing perhaps a solution in the future quickly. What do you mean by that?

Speaker C: Well, so um, the, the idea of building this agentic AI is uh, one of the discussions that we will always have with our customers uh, on which cloud provider is right for their workload. Of course after checking the sovereignty and the presence of the cloud provider it comes to a discussion of which one would be better as well as a revenue generating from the cost perspective and which one can provide them with the right tool. And the right performance. Imagine having all of that being done through AI. The AI uh would check your workload and then select the right provider accordingly. Deploy the workload on the right cloud provider, manage the growth as well. So for example sometimes you will have a spike in your workload. The AI will decide to continue deploying on the same cloud provider or a different cloud provider and then having that in control and controlling all the financial aspects of it. That's as per the article and as per the current uh, market status is going to be the major problem. But as soon as this is developed and evolved enough I believe that that solution will be in need for every organization to avoid the manual selection and the manual movement across the multi cloud provider and leaving the AI to select the right one based on the performance and the cost.

Speaker B: Wow. Okay. That will be an interesting future. Um, Johan, what are your uh Johan, what are your impressions of the article? The author stresses the need for careful orchestration across multi cloud environments. What from your perspective, what are some of the uh, important factors to consider?

Speaker A: Um Office. First off I think it's a wonderful use case for agentic AI and as Shadi mentions there's a product in there somewhere in the future and m maybe a whole new cottage industry um, going forward. Um, with my security hat on. Um, obviously I thought I read the, the phrase um at its core an agentic AI system is a self governing decision making system. I was scared to the very core of my being um, because what happens when that goes wrong? We've all seen it. A user or an administrator who's supposedly trained in this kind of stuff, um, misconfigures um, something somewhere in a checkbox or in a drop down box, um, in a configuration um menu and suddenly your cloud costs skyrocket um for a day or two and um, then somebody has to back up and somebody has to plead with accounts at Microsoft or Amazon or whoever.com um so that highlighted to me, gosh, you have to know what the agentic AI is allowed to do. You have to know what it's what you wanted to do and, and, and you have to be absolutely confident in the parameters that you feed it. And then obviously you need to, you need to test it to the point where you can trust it. Um, the article uh, author openly stated states was a small test and he picked on quite a few problems, quite a few areas where the interactions weren't quite what he expected and he had to keep on tweaking, tweaking, tweaking. Um, in a production environment you are going to need a human in the loop who actually understands what's going on. And that is so often, um, something that we quote when we talk about safety with AI keeping uh, a human in the loop. But in cases like this where there can be quite, quite exciting cost impacts, uh, if it goes wrong, um,

Speaker B: great advantages and great risks at the same time.

Speaker A: Yeah, yeah. So obviously it is some, but with so many other technologies, um, there is that side where what if it goes wrong? But if we think it through, if we know what we're doing, if we architect it properly, if we make sure that we are paying very close attention to detail, we can get it right. And I believe with enough practice we will. It's just how many bruises do we pick up on the way? Um, but very, very exciting to see that this is being done. Very exciting to see that you know, that this is being done in multi cloud environments. And, and of course also exciting to consider, you know, what, what happens. Add a bit of kubernetes to the mix. Um, add a bit of, you know, start being, just being able to shift at, at this, at a speed that no human can, can match.

Speaker B: Uh, well, all right, well let's explore, explore some of the uh, things to consider as we design our strategies and address some of these issues. But uh, folks, the article is from InfoWorld. I built an agentic AI system across multiple public cloud providers. It is here in the chat section. Take a look at the article yourselves and let us know what you think. Does it ring bells for you? Uh, did it miss anything important that uh, we should consider? Uh, it's right there in the chat feature. So. All right, well gentlemen, I'd like to get us to our first discussion point today and that is preparing for AI driven orchestration and intelligent automation. Look, research shows that organizations using AI based organization can reduce operational incidents by up to 50% and unlock things like faster innovation and downtime. Getting automation right ensures faster delivery with fewer incidents. But let's discuss how we do that and how can companies prepare for automation to gain these benefits? Um, Shadi, how do you help enterprises prepare for their environments to leverage AI orchestration effectively?

Speaker C: So um, I'm going to refer to what Johan has mentioned on the whole security aspect of the AI at the moment. And the race of the AI is still going right now. We're still defining the use cases that the AI will be used for, but we're still not being exposed to the amount of the security breaches that the AI will come with. And every day we come across A new uh, type of an attack that never occurred to us in the typical world. And that's where it's becoming very important that when we build a platform we are building a platform that is secured enough and observed enough in the way that will allow us to control it specifically as well on the guardrails that we will put to avoid the hallucination of these AIs. So, so um, right now what we look at in the organization um and as you've mentioned it will enhance the process massively and it can cut the organization process by 50%. But there is a lot of effort that is being put internally to build the on prem AI platform, build their own LLMs to have it more secured so that they can control it as much as possible. For example in suse we build their own um SUSE AI platform where basically we look at most of the applications that comes on the AI as driven by kubernetes. So they are mainly going to be containerized at this stage. And then we need to build a platform that will cater to these LLMs and at the same time will give you all the security that will uh. I cannot say all because we're still have not seen the different type of attacks but as much as possible secure the environment that can somehow bring this on prem, control every aspect of the access to that LLM M and build the right guardrails on it so that you can avoid this um, ah AI from hallucinating from the actual use case that you are building.

Speaker B: So what you're suggesting Shadi is sort uh of a hybrid environment of part on prem. Part cloud. Doesn't that add a lot of complexity? And isn't that sort of against the theory of using cloud resources to begin with? And I'll ask you this then to follow up on what you were talking about before and Shadi mentioned it again. I mean here we have. How do you, how do you assess the risk and ensure compliance when applying these tools, these automations across critical systems and across different cloud environments? What is your suggestion based upon your perspective?

Speaker A: Yeah, um, it's. I think it's going to be a little while before anyone will be able to say that. Assessing that we are very mature at assessing the risks inherent in um implementing and deploying AI in any shape or form. Um, especially when it comes to giving agentic AI so much, dare I say agency uh when it comes to um, to uh orchestrating the systems upon which we as an organization rely on in order to um, make revenue. Right. So we're a good While from being particularly mature at that. However, um, we do need to understand what are we, what are we placing within the purview of this system? Which processes are we automating? Um, do we even understand the process? If the answer to that is no, step far away. Go do your homework first. Never try to automate a process that you don't fully understand. Um, I think that should be a key lesson. Unfortunately, it's bad news for a lot

Speaker B: of people, but, well, people have been doing that for years, right? Automating processes they don't understand. Or. So what do we do about that? I'd like to open this up to both of you gentlemen. Um, so, so what do we suggest that we do about this In America, we might say, isn't that horse already out of the barn? In other words, agentic AI. The horse is sort of out of the barn. Uh, are we at real danger here? Uh, what are some of the suggestions that we have for our viewers that might be addressing this very thing in their organizations, um, moving forward with ways that we, they can ensure they are both secure and effective in tools that are already out there in their environment?

Speaker A: Yeah, I think for a lot, you know, the horse has already, for many organizations, the horse has already bolted. Right. Um, and it's, it's, it's out there and it's difficult because the agents are not operating point as they were expected to. Um, and organizations are realizing that some of their processes are more complex than they thought they were, uh, and that the need for an experienced human is greater than they thought it was. Now I don't believe that in all cases, um, that cannot be, um, worked around with enough experience. But we do need to go back to process architecture and process engineering and then we know what we are asking the agents to do. We need to know very clearly what we are allowing the agents to touch. Um, and all of that. The risk assessment there is very clear. Uh, what's the impact if they touch the wrong thing and they input an astronomically stupid parameter into a textbook somewhere? Sometimes nothing. Sometimes that can cost us quite a bit. Uh, in other cases it cost lives. So we need to understand that. We need to deeply understand the process and what is at stake. We also then need to be able to, um, understand how well we understand the agentic AI that we are deploying, because that's the only way that we can, that we can make any kind of judgment as to the, as to the likelihood of something going wrong and the likelihood of a certain impact, um, condition being met. Um, and Then we can start determining the risk situation. I um, think we definitely have a situation right now where there are too many people claiming to be AI experts whereas not enough of them actually understand the math and the um, and the algorithms um, that businesses are being, are deploying left, right and center. Uh, but uh, there is good material out there for those who want to learn. Um, and a lot of the top architects out there are busy, are very openly attempting to uh, get anyone who will listen to hear them when it comes to how to safely deploy um, AI, especially genetic AI.

Speaker B: All right, very good. This is Digital Transformation Talk. We are here with Shadi Khattab from SUSE and Johan Von Dine from Doe and Co. We're talking about uh, scaling across multiple cloud environments using AI across multiple cloud environments. And Shadi, I'd like to move us on to our discussion point number two today and ask your opinion about operating across multiple cloud environments. Considerations like resilience and vendor diversification and regional compliance, these are all issues that face organizations with multi cloud adoption. Over 75% of organizations use uh, multiple public cloud providers according to Gartner. How do you folks at SUSE uh, uh guide organizations in designing multiple cloud architectures that can remain flexible and resilient?

Speaker C: Uh, thank you Kevin for raising this. It's actually uh, you still hear me, right? Okay, perfect.

Speaker B: Yes sir.

Speaker C: It's one of the topic that is on the race specifically from the lessons that has been uh learned in the recent period in the market of um, different vendors that we have seen locking into customers, their data, increasing their uh prices. So customers now understand that they are always going to be looking for flexibility in the solution. And to have this flexibility or building what we call a robust uh, future robust roof um design will always be based on an open source solution. A solution that will give you a choice. Because we all know that technology keep changing uh all the time and we make decisions today that can change next year, the year after better technology would come on. What we in SUSE um do our best in maintaining is to always give them the choice choice that will allow them to orchestrate the workload across the multi vendors, the multi cloud providers and the multi UM technology providers. So in suse for example we have Rancher prime and Rancher prime is a multi um container orchestrator. It is an orchestrator that will orchestrate all the Kubernetes flavors so that we will not be mandating to our customers to use a specific Kubernetes flavor. For example our one is called Rancher

Speaker B: Kubernetes Uh, Johan, let me turn to you. Um, so we're talking about multiple cloud environments. How do you maintain visibility and control when running critical apps across multiple cloud providers?

Speaker A: Uh, you mean other than uh, implementing agentic uh, AI?

Speaker B: Yes, yes indeed.

Speaker A: Yeah. Um, generally speaking, uh, uh, it's, it's, it's, it's always a question, isn't that. No, you, you want to be run multi cloud in order to increase agility and um, and resilience. But most of your engineers, most of the people managing uh, your cloud systems are going to prefer one or the other. Uh, so yeah, you're going to need to train up people to deeply understand those cloud providers that you wish to uh, use uh, so that they themselves also understand which ones are best for which use case and also you know, bring in your legal teams to, to, to deal with uh, um, you know, with, with sovereignty issues. Uh, because some of, some of the cloud cloud providers know their sovereign clouds aren't really worth the um, the um, paper that the sovereignty claims written on. Um, so, so there's a, there's a lot there to unpick but most certainly um, you need to plan for that. If you're planning to, if your planning is fully based on a single cloud um situation then running multi cloud is going to feel frightfully complex. Um, and either way, uh, it's going to be difficult to um, negotiate because all the cloud providers are going to have a slightly different way of doing everything they do. Um, there aren't any great cloud standards uh, that are being imposed on the world yet. Um, and that may be for better or for worse or for both. But so at the moment we are left to unpick that if we do make the bold choice to be actively multi cloud, um, and to want to distribute our workloads across multiple cloud service providers. And we will need the software, we will need the management systems, uh, to support that. Um, and we probably need to rather than cook it ourselves unless we have the expertise in house, we will probably need to look for vendors who can assist us with the management of multi cloud, um, scalability, security operations and of course resilience.

Speaker B: Now from a security standpoint, how do you ensure compliance with different regional regulations while maintaining operational efficiency across multiple cloud deployments?

Speaker A: That is good fun. Um, yeah, because in general maintaining multinational, just maintaining compliance with multinational um requirements, uh, just with, I don't know, the current state of NIST2 for example, it feels like you're being pulled in multiple different directions. Right, all at the same time. Um, and yeah, it's difficult. But fortunately you don't necessarily have to carve the world into all the various countries um, that you need or that you operate in, but you can do it on a more regional basis. Um, so European regional requirements may be met by a sovereign European cloud or it may be met at some point if, um, if there are better protections for, um, for uh, cloud service providers who uh, belong to a certain other country, uh, when they are operating sovereign clouds within other countries, um, and within other regions. So the, the protections afforded to the likes of um, Azure, aws, et cetera, when it comes to operating sovereign clouds elsewhere, um, at the moment those are not very strong. Um, and therefore we can't really uh, trust them to be fully sovereign yet. Um, it would make our lives a whole lot easier if we could. Yes. Uh, so I'm hoping that at some point we get a diplomatic solution to that, uh, that then allows that to be much easier. In the meantime, I, uh, think we're kind of between a rock and a hard place because even the European regulators are not really going to tell you, don't uh, use American CSP yet.

Speaker B: We are here today with Johann Von Dine, Chief Information Security Officer at Doe and Co. And joining us again, welcome is Shadi Khattab, Enterprise Container Platform Sales Manager at suse. If you have a question or comment for our guest today, please feel free to pop them into the chat feature. We will try to get some of your comments and questions into the flow of the show here in the time that we have remaining. And Shadi, we do have a question or a comment coming in. I'm hoping you can help me with, uh, Rebecca is back. Rebecca, thank you so much for your contribution today. Um, her question is this. As organizations scale cloud native environments, how do you keep agility and cost control in balance without compromising resilience? Shadi, do you have a comment for Rebecca?

Speaker D: Yeah, that's actually a great question and um, I'm talking again so I hope I don't get disconnected. Um, um, so basically, um, uh, they are divided into two aspects. The first is the observability and how you can have the full, uh, observability monitoring across the multiple cloud provider and the local provider because the cost control will only take place to as far as you are observing the amount of the workload that you are deploying across the cloud. Um, and then on the other aspect of it is the agility. The agility will happen was how your product is able to scale as quickly as possible. That will Be as a scaling up where the container size expand or scaling out where another container is deployed on the cloud or even on a different cloud provider. So for all of this to happen you would need the orchestrator, the topic that I was referring to before, um uh, I lose the connection um M and this is where for example in SUSE we have the the orchestrator we call Transfer prime. That is the layer of control that can control across the multiple cloud providers. So it can provide the single pane of glass that will allow you manage the different Kubernetes, the ones that you deploy on cloud and the one that you deploy on prem. And then with all of them being deployed you have the observability integrated to it. So you will be able to see all the workloads that you have deployed on prem and that ah, you have deployed on the cloud. So with that you will have the bird view to the amount the containers that you have deployed that and scale it. And of course all of this has to come with the security um deployed across. For example when we look at suse our full solution comes with what we call SUSE security. So basically when you deploy the workload on Prem and you expand it on the cloud, the security will follow you. And this is very important for organizations that deploy compliance. So basically they want to make sure that when they deploy the workload and they are using one of the cloud tool on the public cloud, the compliance policies will follow them. So they are deploying the workload on the cloud or on the PREM with the same compliance policy and there will never be a breach across. So um.

Speaker B: Well Shay, that's interesting. Yeah. I wanted to ask you in our time remaining about Kubernetes and overcoming some of the security and cost management and complexity issues. I'm not an expert at Kubernetes. Kubernetes is an open source platform that automates the deployment, scaling and management of containerized applications which of course you are an expert in. Um, but it comes with some challenges. Scaling um inefficient clusters can increase your cloud cost. Misconfigurations can create security vulnerabilities. So how do you folks ensure reliability and performance when scaling Kubernetes and managing hundreds and thousands of clusters?

Speaker D: So uh, it is very true that Kubernetes is a great technology but comes with a lot of challenges. The first the Kubernetes proved itself to be very good for the business and for the application. So the moment that the businesses start using it, it is scale rapidly and then that's when we lose the control. The other problem that we are seeing with the Kubernetes is that in the market of the Kubernetes there is a lot of different flavors. And by flavor I mean a lot of different solutions from a lot of different companies, which even add a lot of burden to the team that is managing this environment. And due to the growth and um, the different flavors, we wanted to make it easy for our customers by giving them back to what I said as a single pane of glass. So we want them to worry less on the type of the Kubernetes that they are using and to worry more on how they will manage it and the scale it and also to look into the security of it. Uh, because one of the things that we see now with a lot of our customers is due to the growth that happened quickly, they overlooked the Kubernetes security. When we talk about Kubernetes we're talking about application. And application is what is defined in the OSI model or the layers of the network is layer seven. So we need to look at the tool that will be able to monitor what is happening inside the application, inside the microservices m that these applications are having and not just looking at the physical hardware itself or the IP communication. Um, so in suse, for example we have the SUSE security, there is a layer 7 product that will implement the whole compliance policy and then it will, it has machine learning ops embedded in it. So it will learn the typical behavior of the container and based on that it will secure it automatically to make sure that if we oversee it, it will secure it itself. And as back to the previous question, if it was deployed on prem, then it is controlled inside. But if it goes to the cloud, the policy will follow it to the cloud. So it will make sure that no matter where you are deploying it, whatever the policy that you have agreed on it, it will follow you in that perspective. So these are the two major challenges that goes into how you scale it and how you secure it at the same time.

Speaker B: And Johan, what uh, about the tools that you suggest? What tools or frameworks that you suggest that can help us maintain cost efficiency security as Kubernetes uh, grows?

Speaker A: I think to a large degree, uh, in addition to management frameworks is also the uh, choice of, of the uh, containers we choose. Because Kubernetes grade gives us an incredible opportunity to minimize our uh, attack surfaces by choosing absolutely minimal configurations within the containers, um, focused solely on what they are intended to do. Um, and that a is more often Than not, it is great for security. Um, more often than not it is also great for uh, the efficiency of the container itself. And it's great to have more efficient containers and hopefully uh, overall uh, a better uh, performing container for uh, the uh, processing resources, uh, that it requires. Um, so I think that opportunity should not be missed when we're looking at um, containerization in general or kubernetes in particular.

Speaker C: Um,

Speaker A: I think planning for explosive growth is one of those things that absolutely needs to happen when you, when you even think of going down that route. Because it's, it's the kind of technology that once people understand what it is, um, they, everybody wants a piece of it. Um, and, and so building. Look at, looking at the micro components, the individual containers, what's inside them, where can we optimize them, optimize them for security, optimize them for performance, optimize them for scalability, uh, and then being able um, to then. At every level. At every level.

Speaker B: Well, so now my question becomes one of observability, Shadi. How can we maintain governance and observability across rapidly scaling kubernetes environments?

Speaker D: Yep. So um, uh, I recall this sentence and every time we speak to our customer they say that we start a container as a project, but then we realized that we have another project to secure it, and then we secure it and then we realize we have another project to observe it. It is very important to have that in mind. Where, sorry, um, you hear me right?

Speaker B: We just about lost you again. But in the moments that we have uh, left. Uh, Shari, your final thoughts?

Speaker D: Yeah. So, uh, on the observability, I was referring that observability is very important to understand that it relies on correlation and correlation is seeing how the components are related to each other. The container, uh, complexity comes from the fact that there is a lot of components in it, Container, namespaces, pods and then the servers. So having the right observability tool that will show you how every item is connected to each other is extremely important because when there is any change in performance or in the application, you need to spot where is the problem and you need to spot it quickly. For example, in suse we have SUSE observability that is able to do this correlation and at the same time has the machine learning ops to tell you that hello, this is the problem, that's why the problem occurred and that's the resolution of the problem. So it is very important to anyone discovering the container to be also discovering the tools that they will use. To observe how the workload, uh, can be monitored.

Speaker B: This is Digital Transformation Talk. I am here today with Shadi Khattab from SUSE and Johan Von Dine from Doe and company. Gentlemen, we have only just a moment or two left in the time remaining. I'm wondering if each of you could please provide us with one quick action item that our viewers can use to take advantage of, of the ideas and advice today. Johan, do you have an action item for us today?

Speaker A: Indeed. I just like to remind everybody that you cannot secure or manage what you cannot observe. So build that into the very foundations of what you do. I mean, building the capability should almost come after you've considered how you're going to observe this lot. Because you cannot optimize what you can't observe. Um, you cannot, um, you cannot control what you can't observe. You cannot monitor what you can't observe. You cannot pick up that something that your agentic AI is hallucinating or that your cloud costs are suddenly, ah, increasing, um, out, um, of control if you're not actively observing. So observability has to be your number one, um, capability. After that you build processes on that and, and then once you've got observability sorted, go math.

Speaker B: Wonderful. That is Johan Van Dyne from Doe and Co. Johan, thank you so much for being with us today. Shadi, do you have an action item for us?

Speaker D: Well, um, uh, two actually. Uh, the one, the first one is when it comes to selecting technique, technology, I would say always look for how you will be in control of having a choice and avoid the lock in. Look for open source. That will always drive innovation. The second one is on the agentic AI. Um, um, if, if you haven't discovered that yet, then start discovering and start seeing how even it can automate part of your current or day to day, uh, job. So yeah, these are the two things.

Speaker B: Wonderful. That is Shadi Khattab from Sousa. Shadi, Johan, thank you so much for being with us today. Your comments and perspectives were spot on. I hope that we'll have a chance to talk again soon. So thank you so much and to everyone joining us today, thank you too. Um, join us next time on Digital transformation talk on November 25th with another great panel of guests. We'll be discussing the topic of turning digital sovereignty policy into practice. That should be another great discussion. In the meantime, if you'd like to connect with me, you can do so on LinkedIn. You can find me there and check me m out. I'm Kevin Crane and check out my weekly audio podcast, the Digital Transformation Podcast. But for now, that'll do it for this episode of Digital Transformation Talk. And until next time, I am Kevin Crane saying thanks for watching.

Speaker A: Thank you.