Hey channel insiders, thanks for tuning in for today's episode. Real quick, how did you get here today? YouTube, Apple Podcasts. What if instead of coming to us, you could get this episode sent to your inbox every week along with the top stories and insights from across the channel ecosystem? Well, now you can. Head to channelinsider.com and sign up for our weekly email newsletter right on our homepage. It's never been easier to stay in the know, and now you can, too.
Thanks and enjoy the episode. Hey Channel insiders, welcome back to Channel Insider Partner POV. I'm your host Katie Pavoso and my guest today is Mang Kong Tong, CEO of Sodate USA, a global IT services and consulting company offering everything from product engineering to various technology services. In the age of AI, workloads are more intensive, timelines are shorter, and infrastructure is often overlooked or underprioritized before moving forward with AI initiatives.
MK and I dig into the importance of infrastructure as a service in the AI era and what it means to get it right. Man, thank you so much for joining me today. >> Thank you, Katie. Thanks. I'm happy happy to uh join the podcast today. >> I'm very happy to have you. As we were talking earlier before we jumped on, I know you're in Dallas today, so you're experiencing some warmer weather than I am here in Massachusetts. And I'm very happy for you for that.
I hope you have sunshine as well because I don't know what it looks like at this point. >> Wonderful weather today. So good. Much better than >> we are experiencing some winter storm. >> Yeah, you did have some some freezing I heard in Texas, which must have been wild to experience that. What is that like when you go through that in a state that's probably not really ready for it? I mean everyone just uh we just stuck at home right so we have already that we have enough supplies or food so for the at three days you know either you are working if not you are just you know uh spend times on doing some development work and you know uh spending time with the families >> got to pass the time well thanks again for making it and I'm glad you weren't you weren't frozen out from from the interview today so I'd love to dig into sodate because looking at everything you do you do a lot and it is very very massive as far as not only global full reach but everything that you specialize in.
So you're a global IT services and consulting company offering everything from product engineering to services for cloud data and AI. So when you're explaining the company to a customer or even a partner for the first time, how do you describe what sodate actually does best today? >> What we do best actually is uh product engineering. uh we always uh call oursel as the product engineering, software development as a service and uh data analytics and AI solutions development company right nowadays um development is still the main uh bread and butter for many of the service organization including ourself.
So um this will generate a lot of our uh re uh revenue right in fact uh from this particular service offering itself and all the blockchain uh development uh AI development services they all parks under the product engineering side. So we call ourselves as product engineering uh software development as a service company. >> So what kinds of customers are getting the most value from your company right now. Can you tell me about in terms of their size, their vertical and their maturity level?
Who are they and what problems are they typically coming to you to help them with? >> We typically serve uh three different verticals uh which includes the uh banking and financial services and uh insurance companies. We call that BFSI in short. We also have healthcare and life science. So that is the second vertical that we usually served and the third is going to be manufacturing and high-tech. So these are the three uh key uh verticals that we serve and we support from startup companies all the way to uh tier 2 organizations to fortune 500 clients.
So uh um of course different types of organization will have different needs. Startup companies will look will likely come to us for a quick turnaround to develop something short and quick for a P or demo unit or MVP uh minimal viable products to show to their investors right that is mostly for these startup companies and uh tier 2 clients they're mostly looking for leadership and how they can scale their operations to the next level right and then how they can focusing on their innovations or leveraging the innovation that we have uh so that they can increase their revenue right and uh uh uh or support their uh scalability of the operations.
Uh whereas the enterprise usually is looking for some sort of a cost cutting or offshoring because of our global presence in different um cities and countries. We have about 1,400 500 engineers right now uh located across the globe. Majority of them is actually in Vietnam, India and other Southeast Asia countries like Malaysia, right? Indonesia, Philippines, so on so forth. So they look for us for offshoring or IT outsourcing services or in in current case blockchain AI solutions development because we do have a lot of AI uh accelerators that we have developed inhouse uh that we are offering to different enterprise client as well. >> Amazing.
So like I said very wide reach there. You just named a lot of countries in which you have thousands of of those developers and engineers living in and working in. I'm curious to know what brought you over to Sodotech because you joined about a year ago now as CEO for the US division. What drew you over here and what makes you stay? What makes you get up every day and want to be a part of this organization? >> First of all, right, the uh uh consulting IT services is my passion, right?
So, um even though my first part of my career is not in this field, it's more like engineering and software development R&D type of uh kind of work. uh but my second half of my career are focusing mainly on servicing and uh consulting basically uh leveraging my knowledge or leveraging my team on in helping customers to solve their problems. That is how uh I have been very motivated to do for the past like 15 years right. So um why I joined sort of tech is mainly because I think uh Vietnam is one of the very good alternatives uh to some of the very matured outsourcing countries right like for example India which is traditionally a very popular uh location for outsourcing but of course everyone knows that uh postcoid right for the past three four years um there's a tremendous change to the whole paradigm in the outsourcing landscape so India's uh particularly that probably dominate 80% of the market um is also experiencing a big shift uh we You know a lot of enterprises clients are coming to us because they are having uh problems with management overhead right the cost has increased probably doublefold right if not threefold right some of the cost is uh getting very ridiculous now so it's you know it's become a very difficult and so I selected sort as one of the uh companies to uh to spend my entire my my remaining of my career uh for the next 10 over years 10 20 years at least is really because we have the same uh management style we the the management team has the same vision.
Uh we have a strong and supportive board member and uh most importantly I think we can tap into 200,000 engineers right uh that is graduating from uh Vietnam every single year >> right um so um I can see a group of very very hungry energetic uh young workforce that is that I can tap into uh that we can scale our operations right to support some of the you know uh uh emerging challenges and you know instead of focusing on all the older technologies uh they jump straight to the AI world.
That's why you know we are one of the uh biggest player in the blockchain side right um you know blockchain uh is not easy to find and um I'm so surprised when I see that the companies are founded because they have blockchain development company uh you probably can't even find that kind of a scale in Malaysia or Singapore in a more advanced space so but they have like hundreds of engineers that is certified in blockchain so so it is an eye openener for me uh as someone that has been in the industry for so many years so I thought you know I want to uh offer that to the clients or enterprise that is looking to solve their problems using cutting edge technology as well.
The name itself, Sod Tech stand for state-of-the-art technology, right? So, you know, we are using technology to solve business challenges. >> You want to live up to that name and really showcase why you're called that. I agree. I love to hear about hearing about your passion and hearing about what drew you over there, especially on such a a global stage. You're going we're we're placing our bets in a new market now and seeing the talent and the the expertise coming out of Vietnam.
So, I really enjoyed hearing about that. I'd like to start digging into uh a bit more of the topic that we're here to talk about today, which is infrastructure as a service. It's been around for years, but from your perspective, what's fundamentally different about infrastructure as a service today now that AI workloads are becoming mainstream? >> Okay, this is a yeah, it's a it's a very big topic because uh as from my angle, right, infrastructure has grown tremendously, right?
Uh for the past probably 15 to 20 years, right? uh when I just get to know uh we are working with network you know databases and all that which is part of the whole infrastructures we all manual everything is manual right we are dealing with something onrem before uh that was like 20 plus years ago where things are onrem uh cloud is not popular at all right AWS is still trying very hard to convince people to convert hey you know use my cloud infrastructure right uh it has all the scalability and you know uh cost savings and all that all the perks so uh but of course I will classify that you know into probably three generation, right?
Uh first generation is probably uh you know about 15 years back, right? where the cloud is really starting. people are trying to deploy uh you know on-prem uh resources and infrastructures to the cloud focusing on uh having virtual machines in the cloud having on demand services or having you know how you can scale your operations right through reducing your capeex by using more more of a cloud infrastructures right so that I consider that as a first generation where uh people are really looking into uh reducing cost we're really into that is a key driver reducing cost and scalability right that is the the the the time uh in fact uh the reasons why I uh came to us about 10 over years ago is before because of that reason right client is looking into migrating on-prem data centers to the cloud and they're looking for a company uh to help them to manage that uh this is a mega five fortune 10 companies right and because of that uh I moved from Singapore to US to help them out with this uh uh initiatives and of course that is the first generation and of course shortly after that I think um um probably 10 years back even until today people are still looking into how they can optimize uh the infrastructures using code.
We call that programmable infrastructure, right? That means that instead of managing the infrastructure itself, people are looking into how to manage it more effectively, how to manage it more conveniently, right? So people are looking into using code as a way to manage the infrastructure. That's why you have a term called infrastructure as a code, right? So you use technologies like you know containerization you know kubernetes you use for cloud formation you use um different different technologies right uh to help to do all these things right uh in fact terraform is one of them as well which is a very popular uh today >> so uh using that you are able to manage the cloud more effectively right um you are able to uh more conveniently as well because all these are programmable right become programmable that I will classify as the second generation right you have one step uh better and of course uh u third generation which is current right which is an AI AI world things change tremendously again right I I I think that you know because I used to offer something called phops right uh cloud phenop services where my team of members going to the client side to help them to manage uh their infrastructure right uh how to optimize their cost how to effectively you know manage their infrastructures better so that at the end of the day to have a more streamlined services more streamlined processes right to to approve improve certain infrastructures deployment.
Um that is like the second generation time. Now with AI you can do that in in your screen right with natural language. You just don't you don't even need to um deploy yourself. You just have to describe what you are looking for and AI will do it for you. You don't have to worry about hey you know my engineers don't need to crack their head about how they can deploy how what kind of auh resources they should do based on what kind of memory size CPU size is needed and all that all this can be done automatically with AI now so nowadays all the uh uh SAS platform that is used to manage cloud infrastructures right um is all equipped with AI capability of course all with a varieties of different maturity right some really do what they say some no right so But um uh that is the the current thing what that is happening right now right uh infrastructures become invisible.
So people are no longer seeing that you know because uh last time you used to log into AWS and you know Azure GCP to start to manage your infrastructures and deploy right all your policies and all that now just through a chatbot you you communicate with AI it will help you to do all these by uh automatically so uh it's a much more advanced stage now. So I call that as gen 3. Of course with all this you know evolution um the skill set of people also changes right I I used to hire uh database administrator network engineers and all that right about 15 years 10 12 years back when I first came to US then after the migration I convert them or reskill my uh resources to be cloud engineers right they do cloud engineers they do devops you know development uh devops pipeline development so on so forth now it's all you know hidden under the uh the phrase of AI engineers where they do the same thing but again they trained to do better prompting.
They they're trained to do better tuning to make sure that the AI really help them to deploy what is needed uh correctly uh correctly and effectively. >> Very well put. And let's keep tugging on that AI thread that we were both getting into there because AI it sounds exciting. It is exciting but it's also very resource inensive and expensive. So what mistakes do you see organizations making when they rush AI initiatives without rethinking their underlying infrastructure?
This is actually quite um dangerous in fact right if enterprises are not careful about um uh what they are doing right eventually based on what I'm seeing the trend is that eventually right now if you want to deploy some services you lock into AWS you can check what do you want and then you have a price right uh based on your infrastructures that you want to deploy or you know allocate you will have a list of pricing eventually it will be all token based right in the next probably two three years, right?
Maybe sooner. No one can predict when will happen, right? It will be token based, right? That means that you just buy token from AWS. It's only one price a token, right? So, if let's say you say I want to do this, right? Um the AI will convert the token into whatever spending, right? um uh that is needed for the uh allocation of resources automatically and scalable and then after they come back and say okay based on what you need you know or this will be deployed these are the resources and then you know you'll be using about you know I don't know uh five token or three token or whatever right this is uh giving a very a lot of convenience to the to to the enterprise but you you will not know what is behind the scene right so uh the conversions and everything is still yet to be defined right um you may just burn away all your money uh very quickly Just like from first generation to second generation just I mentioned right so uh when customers come to us for cloud migration they are looking into cost optimization objective but after they migrate then they realize oh actually they spend 40% more than more more than their infrastructure cost in the past then they come back to us and hey MK right we thought you know migrating to the cloud is going to save cost the and the real the the reality is that it is not always the case right it depending many factors right you didn't if you didn't optimize your architecture optimize the deployment properly and you don't have a right governance and processor oriented you know team it may go up very drastically so 40% of increase of cost right after migration so they are shock of course then of course we are being deployed again to employ you know to kind of like study the whole landscape to see how we can reduce the cost the cost eventually does drop after about 6 to 9 months it is not dropped immediately right so it's a long cycle same thing will happen from gen two to gen 3 right u kind of a scenario where people just jump into AI bandwagon and thinking that they can uh solve some business challenges and reducing cost that will not the case always because the focus is not about cost right uh and if you ask most of the big players nowadays enterprises AI the powerful thing is not about cost savings it's about how you can turn compute right into innovation with the compute with the money that you spend you can speed up the kind of innovation that you may take you about 5 years, 10 years, 15 years.
So that is the key power of AI. So people that jump into the wagon shouldn't be thinking about cost. They should think about how they can maximize the output of AI to evaluate you know innovation that can be evaluated. >> I see. I see. So one one concern that we hear often is that infrastructure as it becomes more automated especially with AI organizations fear of losing control. They fear of if we automate it too much, we won't be able to manually do that.
And you talked about the manual tasks bogging people down, but now we're seeing the overcorrection of worry. So, how does Sodk help customers modernize with infrastructure as a service without sacrificing visibility, government or security? So uh that comes with uh the experience that you have right because we have worked with um multiple enterprise before especially uh I mentioned to you just now the three verticals that they are focusing two of them are highly uh regulated right so uh banking financial services and insurance is highly regulated health so do so does uh healthcare and life science right so uh with our experience working in uh such uh enterprises and industries we know and we have deployed and managed cloud infrastructures for the past 10 over is we know how to uh deploy the right governance right of course AI is always um evolving right no one can come and tell hey and say hey I have the whole set of rules this is a rule that you follow right no one right now can says that but we do know that what is needed in different industries right what is needed to make sure that we have the HIPPA compliance what is needed to make sure that you have all the different uh compliance required right uh for different vertical and industries for that we will code that and make sure that you know whatever AI deploy we'll to have a double check right to make sure that this is the you know running accordingly.
That's why often a time we have a team of people uh that is trained on that and to do to run audits right to make sure that deployments are being done correctly and effectively and following the guidelines and governance that that is defined. Of course the governance in a playbook in a form of a playbook is always aligned with the clients uh before we we uh use that as a a handbook right for deployment. I would love to know a little bit more about a product that you spoke to me about in our our call leading up to this.
Can you talk to me about ACE? >> ACE is actually um in a form of a rack where you can actually deploy um infrastructures automatically by giving using natural language, right? Um that means that instead of hiring a bunch of cloud engineers or you know DevOps engineers and all that helping you to deploy your resources, it is exactly in Gen 3 where you can deploy that automatically, right? So you talk to a chatbot and tell them that hey you know I want to develop this particular applications that is uh to be served in the what industry what is the purpose of the applications and uh what kind of a user base that you're talking about in which country then ACE will automatically calculate what is needed to be done what kind of resources is required what kind of memory size and is needed uh if there's AI features in it right what kind of GPU power is required right u and then after that we'll ask this is the summary of what is going to going to to be deployed and You have two options for option A, option B, right?
What is the pros and cons of option A? What is the pros and cons of option B? Then clients can choose uh what they want and when you choose option A, we'll deploy accordingly. And then after that, we'll evaluate and you know uh all the governance, all the playbook because often the time in our industry uh is a highly customized industry, right? Uh customized uh service offering, right? So we we have to make sure that whatever that we advising, the AI is advising uh it has to be applicable to the customer's environment, right?
And their processes. So, so this will uh be part of the uh conversation with AI. AI will communicate with the customers. Hey, you know, uh this is the options and this is our current playbook. This is our current uh kind of policies. Uh do you agree or do you not agree? If you don't agree, of course, you give me what you have and then after again, I will populate it again and then after that, you know, once we made an alignment, then we'll deploy automatically.
Uh this this is considered a setup process, right? Because I'm talking about the initial setup phase, right? Once everything is set up, the AI is very familiar with um you know your environment, very familiar with your uh your your infrastructures, your architectures, your apps requirements, uh your industry model, then you know everything will be done very quickly, right? >> You work closely with SAP environments in the cloud. What makes SAP workloads a good or challenging fit for infrastructure as a service and what should organizations get right before they migrate? in general right uh any SAP environmental migration requires a lot of effort right even within SAP you are trying to upgrade from version to versions uh it is also taking a lot of effort I think the main thing is uh the the preparations needed is really the uh the processes the workflow that is associated and embedded in SAP itself because SAP is ERP system right uh which is uh most of the time interlink integrated with uh varieties of ME system right or CRM system or your own databases, your or AI intelligence systems, right?
So all these are actually their scattera system. So depending on the industry, so all these are inter integrated with your SAP system. So when you want to migrate something uh SAP versions to either to within SAP itself or to another software, it requires you to understand the database, the schema, right, the processes that that is embedded into the system that is programmed into SAP, all the automation that you have already created, right? uh can you be uh directly uh migrated to another platform SAP sometimes offer direct uh uh uh migration but many many a time it doesn't it is known to be notoriously known to be difficult to do that that's the reason why companies like us exist because we have done so many of those kind of migration so and we have done a lot of preparations framework best practices that we can just you know leverage and we use right in fact we have in-house accelerators tool that can help to migrate uh in fact instead of you know maybe you spend a lot of time uh helping to prepare some of these.
We can use the tool to scan through automatically all the link and everything linkages and processes and integration points right API uh scanning and all that instead of doing it manually we can scan through and then give you a list of things that is being uh you know associated with your current SAP tool and then after that we can uh deploy it right you with a tool that we can do deployment to uh different tools different platforms that customers requesting of course all this depending on really there's so many ERP systems nowadays right uh big small customized different countries use different thing right uh there's there's too many so we we have done a lot of migration activities but you know uh those those playbook or those framework are still evolving but uh definitely uh it's much easier than someone doing it manually. >> I'd like to know from your experience what separates organizations that manage infrastructure as a service well from those that struggle.
I I could imagine you answer this question saying well the ones who don't come to us are probably having an easier job but the ones who do come to us are struggling. But what is there is there a key difference or a pattern that you see for companies that are doing it well compared to those who are not? a few things right whether they come to us doesn't matter I think main main thing is that the um first they have to keep up to the trend right of what is needed what is available what what kind of tools is available in the market that they can use some are open source some are not right understand the technologies that is available and uh probably don't jump into the wagon immediately right um study that do a P try it out you know migrate a small little piece of it I have he too many times that customers you know just gum straight they didn't do thorough preparations and investigation And they just you know based on some research you know online they do some studies point A point B right you know hey this is the pros and cons you know let's do it a management buy off and then after that you know it's become a total disaster migration can be done probably spend two couple of million dollars to migrate after that again they don't achieve the objective they do what that happens all the time right of course we we come in to add value because we have done it and we can tell clients hey you know customers hey I did something similar you better avoid the mistake as well right so things that that is our value but of course putting sod aside um you know I think technologies is important right understanding that doing proper preparations and uh investigation and due diligence is important and u making sure that the team many of times is um this kind of migration work or optimization work are actually people oriented very very heavily reliant on the people that is doing that right um people has to be willingly understand that this is a need and then migrate that you know and and then making sure that they are actually wanting to do that because many customers because with advanced of tool many a time it means you know less work right because automation is in place people are worried about their job you know nowadays so uh but of course from on the other hand right uh I'm seeing that customers is u this is part of the whole preparation as well many of times they can uh better prepare uh the team by rescaling them as well we're giving them a plan that hey this is going to be the new platform you to be trained to use this new platform right and with this new platform right you become more effective and efficient then you and use whatever the remaining of the time to uh implement something more innovatively right.
So people spend more time doing more uh value added you know uh activities uh instead of worrying about hey you know I my job will be replaced of course as I mentioned right uh in summary right technology due diligence you know making sure that the people are well skilled at the right area um and then process and making sure that you have the right process in uh before and after right you have to map it out um the right processes all these are part of the whole migration you know activities that is needed you know >> I'm curious what your take might be on this next question uh because As infrastructure becomes more intelligent and automated, some managed service providers worry about being disintermediated, essentially cut out of the deal, cut out of the relationship.
Where do you see the biggest opportunity for MSPs in an AIdriven infrastructure as a service landscape? >> I have an interesting discussion yesterday with a group of CEO, right? So, um there's a two groups, right? One group is still talking about hey AI is going to replace the works that being being done. The second group has a totally different ideas, right? I think what you ask is kind of a relevant because as an M MSP in the past you provide bodies you provide resources to help customers to solve their scalability problem right you help to scale their problems you have to manage their uh IT operations at a cheaper location you know those are the many things that one of the things that MSP are doing >> um and with AI things are getting automated right people are worrying that hey last time I used to propose 300 people to work on this particular thing now every customers is coming to see how much AI is inside every deal that I'm coming across nowadays right from software development how much element is done by AI are they worry about whether AI is there of course they may be worry about whether some codes is AI whether it's you know effective whether is secured or not and all that right but the main question is that how much money they can save right so instead of uh 20 people 30 people in the team you probably need 10 people right if you don't tell me the same right it will be different no matter what so so from that kind of angle right people are already expecting some AI is like regard right regardless now in this uh current uh time.
So with that I think the the way that MSP or companies like us IT service provider or consulting company right is really to make sure that the the human interaction is still uh enforced. Um the domain right the domain is something that not something that is be can be replaceable by AI instead of worry about hey you know I have tons of resources that can help you to scale which is most of the value proposition of MSP. Now it should be how much value I can bring right what kind of security that I can help you with your AI that you're trying to deploy what kind of uh cross country border you know governance and security and observability that I can use to help you with that and how can I help you to clean up your data better so that your AI can have a better output right so that is the the the focus is different you have to change according to the focus as well you cannot use cost of course cost is already an expected element right in all engagement whether is before or after cost is already you know a standard and expectation but on top of that right um instead of focusing so much on manual resources and stuff you know focusing on things that how you can add value to the client to drive AI implementation better how to use you know AI in a better effective way in terms of managing the service operations we have a tool that can manage using a agent AI to do customer service right but is it replacing people no right the tool can help in a certain extent But at the end of the day, you still need a layer of resources to be to provide the human interaction.
Nowadays, AI is very often used for layer 1. We call it layer 1, L1, L2, L3, right? L4 there layers. So, but at least you know L1 is can be done pretty effective effectively uh AI but not L2, L3, L4, right? So, those are still very human intensive, right? And so, uh we have to the MSP has to shift their focus a little bit, right? when they sell their value proposition and of course with that they also have to shift their whole operation model right we are shifting our whole operation model to make sure that all the we call the SDLC right the development life cycle right it's all has AI element in it and then with AI element in it we can either make it more efficient at the same time we can also reduce the cost and transfer the value to the client right at the same time we have proper governance in place at the end of it to make sure that you know whatever that we're trying to use from AI has to be properly vetted through and then scrutinized by uh experienced engineers as well.
AI can also use to kind of like scan through for example test cases, right? Software development unit testing, right? So software development test cases generated by AI. You also have to be vetted by another AI. It it is possible nowadays, right? With multiple different LLM modeling, you can have something generated by AI and then vetted by another AI, right? And then vetted by a human. So it's it's possible. Yeah. With that in mind, then you've talked to me in the past about encouraging customers to focus on smaller high impact AI use cases first.
What do those wins look like and how should organizations think about aligning those early AI wins with their infrastructure strategy? >> First of all, understanding their whole infrastructure strategy is important, right? Before even think about the smaller use case that can align with the bigger picture of the infrastructure strategy because the the basic question right now is hybrid. Whether are you going to do u um um AI yourself, right? Whether you want to deploy data centers yourself, right?
There's a group of uh companies that's wanting to do uh data centers, AI data centers themselves without relying on the cloud, right? So um uh they are often the national the government agencies, right? There's something called sovereign AI, right? That is they want to deploy this inside the premises, right? So there's also companies, enterprises that wants to do in the cloud in a hybrid way. There's no another strategies about vertical uh infrastructure as well uh to deploy an AI because a lot of AI nowadays um is general AI right you have a general generic LLM that is deployed most people are using that but the trend is also moving that you know are you deploying this specifically trained uh AI in a special mode like for example for healthcare you have a healthcare LLM for banking you have a banking LLM so that is going to be what happened right in the future you have individual verticalized LLM that is trained for particular industry and not only that now he's talking about training alm in future you have cloud AI accelerators that is also designed that mean from the silicons from the chip perspective it designed that way that is optimized to run in your environment in your vertical right so those are all happening I mean 6 months ago you ask me I will tell you probably 3 four years down the road now I I don't have an answer it can happen like now or maybe even like you know few weeks or few months down the road it's AI just changes everything right it's so fast to the kind of deployment and innovation so as a CIO of the company for example you really have to think beyond that right think about five years down the road what is really happen and then even if you don't know what it will happen but at least you need to know what kind of trend is going to happen right for the few options and how you are going to mold your organizations infrastructures towards that direction right that means your goal at the same time making sure that your current operations your current projects your uh resources everything is still in line right to intact right without going haywire.
I think that is the number one thing that u every uh CIO has to think about it right um looking at the future looking into what we have right now and how we can map towards that uh direction without losing track. After that, after you have that direction, then you think about, okay, what do I need now, right? In order for them to quickly demonstrate some use case, right? Quickly to deploy uh some some of this use case to uh their upper management or to the board, right?
For example, to have a better budget, bigger budget allocation or you know or investment for startup companies, you really have to demonstrate in the short term, right? Because infrastructure project are usually long term, right? Take take about 9 months, one year kind of down the road, right? Even with AI is probably six months still right it's going to be long term right uh but people are looking into hey can I see something three months I want to see the ROI in the two three months time three four months time right so um that is the the reasons why I I'm telling you that uh people are moving into um very specific use case right so this use case can be in the case of a sales agent right it can be maintenance agent customer support agent it can be onboarding agent so so these are all in general customer support in the past but now with this whole process of customer support, you can dissect into so many mini, you know, uh uh role that is needed, right?
And you can use AI to test out each of these area, right? Uh from how you can scan through the whole uh communications, you know, and sniff out some business opportunity, right? and convert that into potential opportunities uh in a CRM that using the CRM to uh AI train to understand and analyze each of these opportunity whether this is worthwhile to pursue or know what is a risk level what kind of a uh uh angle that we should penetrate this particular account to how you can run okay once the customers is interested to know more right how you can pursue them after you pursue them close the deal how you can bring them on board to let's say a bank How to bring them on board to be a merchant for example, right?
How to bring them on board as a merchant with the process of bringing on board itself is highly manual in the past, right? I need to sit look at your face. I need to check through your documents, right? I need to make sure that uh you have the right credential, your registration companies, everything is correct. You have the your your P&L, your statement, bank statement is all correct. In the past, all these other manually, right? Now with a agentic AI you can do all these pretty much um you know accurate uh I would say 90% accuracy 85% to 90% accuracy right and then auto highly automated right you have EKYC you have uh OCR technology they can enter documents um the OCI is being trained right the AI is being trained on a particular domain in this case is banking domain so they'll pick up terms they'll pick up the regulations and in the future you need to train the AI to be localized to particular country that you're operating Right?
Malaysia, Singapore, China, India, US, they all have different regulation, banking regulatory. Can you imagine how complex is the system? Right? Your your AI has to be specially trained for that particular location, for the particular country, for the particular vertical, for the particular process. So once you have all these small things defined, okay, then you can see um you know, you see where I'm going, right? So the you can deploy things, you know, at a at a smaller scale effectively.
The reasons why I'm saying why smaller scale because once they get too big you just lose track you unknowingly pump in more and more money more and more money the you know there are a lot of unknown right is in the in the world that you are trying to bring in the moment that you are trying to figure out all the unknown by the time it's going to be like 6 months down the road right so it's better to make something more manageable something small size chunk chunk it up into something small and then focusing on targeting that and on generate ROI immediately quickly clearly define the goal clearly define the ROI and quickly drive service company like us to meet a goal or drive your own team to meet a goal right and of course as as you can see but where I'm going with all these smaller things that you are working on depending on the projects it has to align with the bigger picture right so the bigger picture is the top down and bottoms up approach right once you have the trend you understand where you're going the CIO uh you know defining the infrastructures landscape then you look at some of this project whether this some of this project is fitting nicely to the goal that you are the infrastructures that you are trying to that that you're trying to bring the uh organization to.
Right? So that is the the top down and bottoms up approach both where you have to consider >> something important to consider very much. As we start to look towards the end of this interview today, I always like to ask my guests when they come on for a customer success story. So in the time that you've been with Sodate Tech, has there been any sort of customer success story that you really love to tell as a use case or or a scenario where when you go to a conference and you say, "Hey, this is an example of why we work so well with what we do." What customer success story comes to mind for you? >> Infrastructures um as an immediate one, right? that is the success stories that I helped uh sort to to to perform uh is really the uh cloud migration uh example right so um uh as I mentioned to you after the migration right cost exploded tremendously right so um uh customers doesn't expect that to happen right we do a thorough analysis uh to pres present to them uh what is needed to be done and then of course we also uh immediately uh quickly spend some time because customers doesn't expect that to do that but uh we actually created a phenov platform.
That is the foundation of why I said we have a f phenov platform that can help customers to optimize their cloud spending. Right? So we create a phenov platform so that we give back the control to the client. This is a phenop platform. This is the bearing. This is a dashboard. Right? You can see how much spending uh how much cost that you are incurring in each of these service offering each of these infrastructures and it drill down to individual.
This particular person is spending you how much X dollars this person is spending you how much X dollars. So you have a big pictures of all the cloud spending that you have right now right so customers can say okay I want to reduce 40% let's say okay you can tweak this button tweak this button tweak this button right so it's easy for them to have one look to do that and we spend about 2 months at our cost to develop this uh to the client and customers is extremely impressive because they never expect company like us to do that right company like you know big players like IBM and all that they have those things already right but uh for us we want to be something very customized to client environment.
So we developed something customized for them. Right? So that is the I I think a very good success case study because we have been with the customers for many years and ever since that we have been supporting them with a managed service model uh leveraging the cloud platform at the same time helping them to implement security as a service you know we implemented uh infrastructure as a service infrastructures many things uh with the manage service platform right to support them.
So that is uh one uh pretty good not non AI relevant uh to be uh frank with you uh on that use case but of course right nowadays our phops platforms already integrated with AI insight right that means you can ask AI instead of customers uh coming back to tell us okay you should tweak this this this AI will recommend you okay you should do these five things in order for you to reduce 50 40% of the cost without risk right you can hit 60% of the cost savings but with mitigate the risk right So this is the things that you need to do and these are the security policies that you have and these are the things that you can do right things are getting much easier now with with a more advanced AI integrated uh phops platform that is we deploy them to many customers after that right total almost 300 million of asset uh infrastructures is being managed by the cloud phops platform that we have so uh that is one case study right so the second one is actually uh I would like to introduce about a bank as well right one of the largest bank in Southeast Asia so uh they not in US uh but the world best bank uh for seven consecutive years in Southeast Asia.
So uh traditionally they have been using uh big players like Asentro and IBM specifically IBM right they spend about 350 million uh on consulting company like IBM right so um there are projects that they couldn't move through right despite uh all the effort that they have done so uh when the CIO that we one of my friend uh come to us right uh asking for help they say hey you know I've been using them for for for a year spending a lot of money but they couldn't optimize a simple thing right uh which is a difficult part right uh is a cold storage migration we call it co- storage optimization they are trying to improve the performance right uh of the particular um uh data analytics platform that they have developed for co storage and it's in terms of millisecond right the performance improvement that they're looking for is in terms of millisecond CPU certain percentage of reduction you know memory size certain percentage of induction cost certain percentage reduction so and that is happening during covid time the bank that particular bank has very strict policies no one can work from home so uh it's very difficult back back then right and we are the only team that is going there on prem in the in the in their office right helping there and there's no other employee there we are the only one so uh we went there right um and the whole plane we are the only team there right there's no one else in the plane itself so we went over there and then we helped them out to to to solve their problem we spent about 6 months right three months in the office in in the country in that location and then three months you know offshore uh helping them with that and the customers is extremely impressive after 6 months right so um that is only one small project.
The project is not big size but the the next years she they gave us about 15 million projects within one year right after that project because customer is so impressive to work that um um you know with a small company like us right comparing to you know mega big giant player in the market they're able to resolve the problems right receive customer service and they can see the passion people are you know risking their life to go all the way to the office helping them to you know solve the problem and then uh they are very very impressive basically right and um the next years I have seven projects from them 15 million revenue right so all related to um um um data analytics right we become the uh data analytics agency of record right uh for that particular client and then um we do data observability we do uh a whole legacy modernization migration work we do cloud cloud migration work you know all these works for them so so it's a lot of work for the client uh after that so it's a it's a pretty impressive uh case study right from my angle I will definitely agree with that.
Very impressive and great to know that they're a happy returning customer. For those who want to learn more about you, about Sodate and to potentially become a customer or even partner with you, where can those interested go to learn more about you? >> Uh first thing of course they can learn more things from our website, right? Our website has a tons of information. In fact, we have um recently also listed all the AI tools that we have in the market uh in in in serving different clients to the website.
We have about 20 plus agent AI solutions in our website now or maybe reach out to me if I'm of any assistant uh to anyone. So uh feel free to do that as well. I can be reachable in my email and uh cell phone. >> Oh, amazing. Well, Mang Kong, thank you so much for your time today. It's been a pleasure getting to learn about Sodac and about your journey and I look forward to following along in 2026. >> Thank you so much, Katie. Thanks for the interview today.
Thank you. >> Thanks so much to Mang Kong for joining me today and thank you for watching or listening. You can check out every episode of Channel Insider Partner POV on channelinsider.com or watch us on YouTube at youtube.com/ channelinssider_news and trends. You can also listen to us as a podcast wherever you get your podcast from. Don't forget to like, subscribe, and follow wherever possible so you never miss an episode. Come connect with me, Katie Bavoso, or Channel Insider on LinkedIn and X. Once again, I'm your host, Katie Bavoso, and I'll see you next time.
This transcript was generated automatically from the
video's captions and may contain errors.
