Fibish Madathil | IBM | Cloud Expo Asia | Singapore
Andrew: I am now delighted to welcome our friends from IBM, Fibish Madathil the cloud adoption leader for IBM in this area. So, let’s start from the beginning. You are the cloud adoption leader, all things cloud. Do you want to give us a little bit of insight about what you do and within IBM?
Fibish: Sure, so I’m responsible for helping our enterprise lines on the adoption of cloud platforms. So, helping on the strategy roadmap, architecture and the use cases also my team, work to deploy that platform. So, you know in the past is all about infrastructure as a service. So now it’s all about applications and how do you develop new, innovative applications on the containers and microservices architecture. So this is the new innovative cloud adoption happening with our clients and so we have platforms, we have private clouds, we also have public clouds. Right, so this part does on frame and off frame and all the eyes and other blockchain capabilities on the cloud. We can help our client to adopt as the adoption matures and evolves. So my role is, if you look at it from a more business perspective, I am a technologist as you know, so it’s really on the technical side also understanding the business value and the use cases around that because most of the time if you look talk about AI, it’s jargon. You know artificial intelligence, then how do I apply that to my organization? How can I create a business value? So my role is to help them with that. You know, what is the right architecture needs for some clouds so that it can scale the need for data for you to analyze and give food into the AI platform and then you put an application to the consumers and then start using it.
Andrew: That’s interesting. So, we’re in Singapore obviously and you cover this area and the surrounding areas. Is cloud adoption here? How have you found it? Is there a specific cloud that people are more prone to enterprises? I supposed you deal with more public but can private, hybrid friends be found?
Fibish: I think if I look at this region specifically in the RC and AP region where I’m responsible for IBM, my role is all about hybrid, right? So the majority is on premises private cloud. I would say 60 to 80 personal enterprises are on private. But there are enterprises that already have applications on the public cloud but it’s a combination right. So sometimes if it’s really in the business needs to appear like a new innovative application, which is not possible, or they don’t have the necessary skill set or infrastructure when they go to now on the public cloud. So, I think if I look at the adoption barrier in this region, specifically on countries like Indonesia, Singapore, Philippines, they have regulations where the GDPR and similar to that, it’s either PDP and all those things. So, the question is what load I can move to the cloud? when I say cloud I mean public cloud and what I should be building or architecting on-prem. So that’s kind of situation in the past, but I think now things are changing. So, there are solutions available to analyze those data patterns and say what is the right workload? Can I move to on-prem or off-prem? those kinds of things. So, in summary, in this part of the world, I would say 60 to 80 percentage is on-prem and remaining 20 is off-prem, but with evolution, it is going to be using hybrid anyway, so there are AI solution where IBMare helping.
Andrew: So how are IBM working in the eye in this region? Tell me a little bit about what you’re finding
Fibish: Yeah, so. Like I mentioned in the beginning, AI is just everywhere right. But I think the challenge we are facing is really the business value, the use cases, the trust and transparency and the are the skill right. All of this constitutes the adoption challenge for AI and IBM. So like I mentioned we are Watson. So, we can provide Watson capability on private, on-prem. So, I think the concern about trust and transparency is kind of you know, we can take that to on-prem where their data and everything is on prem and we have a data science platform where we can analyze that data. It can model some of the data patterns into the AI engine. So, we have been working with lot of clients, especially in the financial sector, starting with a virtual agent. virtual agent is just an app right, then you have a lot of collective intelligence from human and computers it all comes together right? So not everything can be processed by the AI engine, but you still need to have a process of approval required. We have been working with some of the insurance companies to build a virtual agent for smart claim processing and things like that so that the adoption is happening. The advantage of IBM is really the capability. I think when it comes to trust and transparency, we can even talk about okay, we start with the minimum viable product on on our platform. We can put that back on, in order to clear the same capability on your own right? So that’s the time spent in development. The second is really the skill set. So, we have the engine already available. You don’t build a new invention, so we can use. You know engines ride the natural language processing part WCS and you have visual organisations, you know, all of it depends on use cases. So, we start from the use cases, we help them to understand what business value to get. It’s not just about cost saving or eliminating job is about enabling the humans to do better, you know, we all know that is the key.
Andrew: We’ve used Watson and we use the translation tools within Watson text to speech from these types of interviews and it’s translated in many languages. It’s been really good. You are a cloud adoption leader. You’ve written this acronym here and I thought I had heard this acronym before I don’t know if it’s yours or if it’s an industry term. The ABCD weekly Eye blocked in cloud and data, which is very interesting. But data tell me about data, what were you doing with?
Fibish: So, you know every client is in the transformation stage. They want to be on the digital age of the business, right? So, the digital transformation is the big umbrella and there are multiple initiative within that. Cloud is the foundation which enable all the transformation on the technology space. So, it provides scalability, it provides compute and memories and all the resources that they need to build an application, right? So that is something that you build on an application development platform whether you build on private on-premises or on the public cloud like AWS or IBM or Google, right? So that’s the foundation now the second is really the data. Becuase Data is a golden asset, which is hidden in most of the enterprises. So, there is no clarity on how do I get my customers in most of the places? Some customers are so advanced, they have all the analytics and data engines available, right? So, the data platform is so important. So, we have a platform for IBM Clouds on private. So, we have DSX data science experience, we built on that where you can run a PR with various data sources, you know, which pulls the data we can use. And now most importantly, you see the foundation in the cloud architecture, you know hybrid cloud architecture. Then you have a data platform sitting on it. And we analyze the data based on your modeling and what’s inside and your needs and then you feedback that back onto your AI and blockchain. That’s why I said ABCD. The data architecture, that engineering of the data science or analytics is so important in this individual transformation because otherwise you don’t have any way of helping to build a new capability in every digital solution because AI is all about end of the day data, how do I kind of connect that and so we have been working with many clients, you know in this region. To command and analyze and you know discover, explore and create the modeling and the pre-staged area and then also creating the DSX. We have a platform a simple one that we can use in our various capabilities of this blockchain or AI induction. This is something we have been working with many many organisations. As we think this is where the supply chain is. So that’s why I say ABCD to you.
Andrew: One last question because I asked someone earlier and I’m just curious. We speak about data and I’ve got this term I don’t know what the actual term is, but I call it dirty data. What I mean by dirty data is these companies have just accumulated all this data for years. How do you actually get the good stuff away from the bad stuff? You know what I mean by that? How do you go about doing that? What do you tell them? Do they just start from scratch or how would it usually work? And I know it’s a tough question and I put you on the spot.
Fibish: Yeah. It’s interesting because to me, you know, peoples data is a golden asset in any organization. It’s within their enterprise not external, right? So, if it is structured or semi-structured or unstructured that’s key and some of the data is just there, they don’t know what it is and they have no clue why they put all this attention there. So, I think the first step is really to understand right? So, do they have a real governance in place for the data? What that means is whether is it classified as critical or business critical or less critical there are different categories when it comes to the compliance and governance, right? So, we need to look at that and then think how do I understand? How do I help them understand whether is kind of searchable and some of them are not searchable so we can classify that from a business perspective for my information data perspective and then put that back on the technology data sources, you know, it could be running on open source technology like MongoDB or to the Oracle or SQL or things like that. But then can I use an application with an API to search those data sources, right? So, I think the challenge, the complexity of the organisation is spirital and regional. So, we need to bring them together and maybe we are putting on a workshop and get those details and data, so, what datasets are needed if it’s a 360-degree for affirming this customer. There are multiple details, right? Whatever course you do, how do I get those details from them? And then we put that back on to structured modelling. We do a discovery of the data and which sort of methodology to use. I can talk for hours on that. You discover, then you analyze then you remodel it and then you take that down to the inside of what you want and you put that back onto API for AI applications.