Our Own Devices with Nandagopal Rajan

How AI has evolved over the past decade

In this episode of The Indian Express tech podcast – Our Own Devices, host Nandagopal Rajan, COO, The Indian Express Online is joined by Sandip Patel, Managing Director, IBM India & South Asia. IBM is one of those organizations that has been at the forefront of AI development. In fact, they brought out their own AI platform - Watsonx. Sandip Patel shares how watsonx is making strides in the world of AI. He also talks about how India is definitely an early adopter of AI and how companies are benefiting from it. Interestingly, a major part of employee - HR interaction at IBM happens via a chat interface without any human intervention. But does that mean artificial intelligence will render us humans redundant? To find out tune into today's episode of Our Own Devices with Nandagopal Rajan.

Introduction 00:00
IBM and AI 0:55
IBM's AI platform - Watsonx 2:24
Working with Indian companies 7:27
Challenges 9:20
Skill requirement for AI 12:60
AI and India 18:26

Edited and mixed by Suresh Pawar

Duration:: 23m
Broadcast on:: 24 Jun 2024
Audio Format:: mp3

At your job, do you ever have to deal with a nose roller? How about a snub pulley? Well, if you're installing a new conveyor belt system, dealing with the different components can sound like you're speaking a foreign language. Luckily, you've got a team ready to help. Granger's technical product specialists are fluent in maintenance, repair, and operations. So whenever you wanna talk shop, just reach out. Call clickgranger.com or just stop by. Granger, for the ones who get it done. (upbeat music) - Hello and welcome to another episode of our own devices. This time, to talk about, again, our favorite topic these days, which is AI, is Sandeep Patel, who is managing director of IBM India and South Asia. And if you understand, IBM is one of the companies which has been in this way before, maybe even the word AI became something people knew, right? Because they have been doing so many things in this space for so many years. And again, we keep emphasizing that, even though AI is barely burst out of there, everybody screams literally in the past couple of years, it's something which has been there for maybe decades. And to give you that kind of a context, you need a company like IBM. And when a company like IBM talks about what's happening right now, you understand the kind of knowledge that it's going to give you. So Mr. Patel, welcome to the show. - Great to be here. Thank you for having me. - So if I can start by asking you, if you could set the IBM context in this space, which again is many years of working on this thing, and you suddenly come to a point where you have a real side view of what's happening. - Yeah, look, as I think you rightly said, we have been working with artificial intelligence for a long time, whether it was the early days of Watson to the work that we were doing in the health space to even intelligent systems and automation that we were driving. I think they were all flavors of AI in some way, shape or form. And I think in today's world, what we are seeing is that a lot of people ask me the question whether, you know, is AI here to stay? Is Genii just a fan? I think we are beyond that now. Genii is definitely here to stay. It's a real game changer for Indian businesses or businesses anywhere. And the real game changer is going to be the ability to build and scale Genii that users can trust. So I think there is an element of trust that is important. I think an element of ensuring that as we scale this technology, we do it through open innovation and we do it in a way that becomes relevant for business. And, you know, one of the things I'm hoping, you know, to share with you is some real examples and how this technology is actually evolving and becoming more and more practical for business applications. If you could also tell us again, Watson is something you've had for a long time. But I guess Watson has also evolved for the Genii era, right? If you tell us a bit about where is Watson now, I guess it's Watson X. And how is it different from what it used to be? Yeah. So, you know, maybe that the way I can help put it in context. It's always good technologies and context, right? Yeah. So why don't I put it in context in terms of what are we seeing as AI and AI drivers for adoption? And then link that to how Watson X is evolving as a technology. And I think, you know, one of the things I'd just like to set as a context upfront is last week at our annual flagship event, Think 2024, our CEO, Arvind Krishna spoke about our open innovation strategy for AI. At IBM, you know, we've always been committed to evolving technologies in a very responsible and practical manner. And we are strongly committed to building and collaborating on an ecosystem for AI. If you tie this back to the robust investments and energy that surrounding artificial intelligence in India, the opportunity is clearly, you know, phenomenal. Now, let me give you to a couple of interesting data points because it's always good to put technology in the context of how it's being adopted and how it's being used. So we did a survey last year in 2023 called the IBM Global AI Adoption Index. And it had some very interesting findings. First, it showed that 59% of enterprise scale organization, which is over 1,000 employees that were surveyed in India have AI actively in use in their businesses. And interestingly enough, India was one of the highest in terms of the adoption index. You know, the 59% was one of the highest across all the other markets that were surveyed. That was my number one. The second interesting data point was that early adopters are clearly leading the way with 74% of those Indian enterprises already working with AI and having accelerated their investments in AI in the last 24 months in areas like R&D, workforce reskilling, et cetera. So as you look across the Indian business landscape, we are seeing three main drivers for AI adoption in our region. First, the advances in AI tools are making them a lot more accessible. Second, businesses are relying on AI to reduce costs and automate key business processes. And third, the increasing amount of AI embedded into standard off-the-shelf business application, it really enables businesses to easily commence their AI journeys and integrate AI into their operations. And the other thing that we are also seeing is the maturity of classic AI models creating a solid foundation for Gen AI, where we are starting to see innovations that we've never seen before. And I think all of this collectively is what we have used in developing our AI platform called Watson X, which essentially is comprises of three different elements. First, which is Watson X dot data. There is a term that we use quite often saying there is no AI without AI. So if you don't have appropriate information architecture and the right data architecture, you will not have AI that you can rely on effectively. So the Watson X dot data component of this Watson X platform basically enables you to establish the right data architecture, have the right data sets, and others that you can pull together to basically make AI work and be effective. The second component that we have is Watson X dot AI, which is more of a developer studio and which leverages large language models, both ones which the IBM has developed, but also third party LLMs, which they can leverage and use to create applications and business applications for AI. And the third component is Watson X dot governance, which I think is a fairly unique thing that we have built into our platform, because as this technology starts to proliferate and be used and starts to scale, the need to govern the LLMs that you're using, the need to govern data that you're using, the need to govern the applications you're building within an enterprise is going to become extremely important. And this is a element of the platform that enables you to drive and do that. So that is what Watson X is all about. It is a platform based approach to generative AI, which enables you not only to use LLMs and others that IBM has developed as proprietary models, but also open source models that can be leveraged from other third party providers. - So it's quite interesting what you said, because I was about to ask you about adoption and you were saying adoption is already very high. So somebody, I'm gonna flip that question a little bit. Like how do you work with Indian companies, for instance? So of course there is adoption and in many things we know India is considered to be an early adopter because as you said, there's a value for money angle that comes in, which Indian companies are quick to lap up. But do you also have to work with them on opening up the opportunities? Because you really don't know, maybe what all can be done, right? Because there is so much possible, but and those possibilities you haven't explored because it's completely new. So how do you work with them? And also this architecture that you mentioned, like you have legacy companies, which have been there for a hundred years, maybe. And the data is not in the structure that you wanted in. So how do you retrofit into something like a Watson X into all kinds of data? Because I've heard from multiple people that getting the data right is one of the biggest challenges. - Yeah, so let me answer this in two parts, right? Let me talk to you a little bit about, so we have actually been very active in working with companies both in India as well as around the world, in driving pilots, because as you very rightly said, Gen AI is a new technology. I think people are still grappling to say, they want to do things with it. There are some very interesting innovative use cases that people are thinking off and trying to experiment with. But getting that pilot right and ensuring that they can then ultimately scale this for enterprise use and for benefit of the enterprise, that becomes very important. So one of the approaches that we've taken is we work with clients through our client engineering teams, which are deep technologies that can work with clients to do pilots that can then ultimately be tested for their use and applicability within the enterprise before they are scaled effectively. So what I want to share with you first is some of the common challenges and what we have learned as insights from over 700 pilots that we've done with IBM Watson X globally, right? So the first challenge that we see, as you very rightly pointed out, is that businesses don't have a clear data architecture in place, and this is very crucial because you can't have an effective AI model without having your data architecture in place, especially when your data is spread across hybrid multi-cloud environments. So as a result, several Gen AI pilots have not made it to production due to challenges with data quality, access and security. So one of the things that we do with clients is to help them establish the right data architecture and actually do that effectively through Watson X dot data so that, you know, the right data can be organized. It can be curated over a period of time. Second, customers are concerned if generative AI models can actually be trusted. So the adoption index talked to you about about 80% of the leaders who were surveyed for that. They expressed ethical concerns as a major challenge in scaling AI across their businesses. Today, most AI models that are offered are trained on data sets that are of unknown quality and provenance. And this actually leads to legal regulatory ethical inaccuracy, those nightmares. And we've seen a lot of this out in the press in terms of issues that have emerged from there. So data provenance and quality really matters. And along those lines, there are two components within what's next that we try to work with the clients to make sure that these challenges are addressed. So besides the clients being able to bring their own curated data sets to Watson X dot data, we are very carefully curating domain specific and internet data sets as a first step to train trustworthy models. And I'll talk about language models in just a second. And before training these models, we cleanse those data sets and filter them for hate, profanity, biased language, licensing restrictions, et cetera. And this has actually allowed us to indemnify our clients when they use our large language models, because we are able to stand behind them based on the data accuracy and provenance. A second part of this issue around trust is that Watson X dot governance actually helps clients govern the training data that they use and the AI they deploy so that they can operate, scale, and succeed with trust, almost as a level of assurance within their enterprise. I told you there were three major learnings. First is the data architecture in place. Second, having AI models that can be trusted. And then the third one, which is actually becoming pretty critical, is skill gaps that remain a barrier to gen AI adoption. Businesses need to provide employees with the relevant skills to work with gen AI. And it's not just technical skills to build models and so on, but it's also training existing employees, re-skilling and retraining the workforce, to work with gen AI as an enabler so that they can improve their productivity and actually do more enhanced work for the enterprise. And we are now also seeing clients higher for gen AI roles that did not exist previously. So that's the whole structure we're trying to sort of build out. And these are the challenges that we're trying to work towards. >> So again, on the skilling bit, and I remember conversations from five years ago when people were talking about AI becoming, it's almost as powerful as electricity. It's going to be there in everything we do. We do this and there was a skilling conversation happening then also. But are you now seeing that the skilling conversation, like you mentioned about gen AI roles, but if you're looking at data architecture, is the skilling requirement also percolating down into other things, which are in a way going to enable the impact of AI? >> Yeah, so I think there are probably three elements to the skilling dialogue. Number one is the technical skills of people who can understand this technology, who can build models who understand how to use these LLMs and so on and so forth. So there is this whole, the technical community that actually is evolving and growing, so there are data scientists and all the rest of it that comes through. The second piece is there's this ongoing debate and questions that you hear from clients. And it's not just clients, I think there's the industry that sort of asks you the question. Is AI going to make humans redundant? And that is a moot question. If you look at time immemorial, right? Every time there's been a technology evolution or a technology that has evolved, it has brought productivity. It has brought new ways of working. It has brought new ways to automate things and therefore innovate other things that you can do both for the enterprise, for business, for customers and so on and so forth. And so the ability for employees to start to use AI effectively, improve their productivity so that they can become more efficient for the enterprise, that becomes really, really important. And what I'm gonna do is I'm gonna share one example, actually I'll share with you two examples, which will be really interesting. One is our own example. At IBM, we have deployed a conversational AI-powered platform called AskHR for IBMers globally. And through that, as much as I think 94% of all employee interactions now happen without human intervention. So it's all through an AI chat interface and this has transformed how IBM HR team works helping them to focus on more value-creating tasks, right? So it's clearly improved productivity of the entire HR team, but they are now able to focus on more value-creating tasks as we continue to sort of grow and scale, right? So that's one. The second area which I think is going to become very, very relevant is something that we are doing called the IBM Watson X-code assistant. It leverages Genii to accelerate code generation and it increases developer productivity while maintaining the principles of trust, security, compliance at its core. So I would encourage you to sort of look at some of our, you know, announcements that we just had at IBM thing 2024, but we previewed a new generative AI-powered tool called IBM Concert. And essentially IBM Concert will provide Genii-driven insights across client's portfolios of applications to identify, predict and suggest fixes for problems, particularly as clients go in for ID and app modernization. Now, why is this very relevant to Indian businesses? Is that in the current context where there is a huge push for application modernization and digitization that's happening across all enterprises and I think it's across all sectors, particularly in banking financial services and so on, where businesses are continuously modernizing their core banking applications, et cetera. A tool like IBM Concert, right, which is Genii-driven, it gives businesses the power to understand and remediate risks across their business applications which are usually fairly distributed across different infrastructures and so on. And it ranks common vulnerabilities and exposures according to that impact on business applications, including operational disruptions that could happen and cascading effects of expired certificates, et cetera, that can enable focused and efficient risk mitigation. So just think about it, right, in a world or in a market where you're driving a lot of IT modernization, where there is a depth of skilled resources that you can apply to these projects that you're doing, right? As if you can use some of these tools to improve productivity across some of these coding tasks and risk mitigation solutions that can be automated, it just enables you to drive efficiency and productivity in a very significant manner. So I don't know if I answered your question, but what I was trying to get to is what are the challenges and then how are these solutions starting to help and then what are the skills that will be needed? So it's not to say that a coder needs to rethink how they code, but they need to start thinking about can they use some of these tools effectively to become more productive and efficient so that they can focus on some of the other, you know, more complex tasks that have to be done as a part of the overall IT modernization. So the other thing I wanted to ask is that since we have an India angle to this and you already mentioned that the adoption is high, but the two other things that I want to ask you is that, you know, often you hear with a lot of technologies, the way India uses it, you know, opens up people's mind, hey, there is a way, there's another way to use this or you create new use cases which are very different, which works for the rest of the world, hey, you know, why didn't we think of something like this? And also again, given the kind of beast that AI is, you know, having the kind of scale Indian companies will have, you know, is that helping AI itself accelerating its change, its growth, you know, its evolution? So if you could just talk about that. - Look, given the scale that India operates in, I think the data points that I shared with you already is suggesting that Indian businesses are at the forefront of, yeah, miming the use of this technology, you know, as you said, we are also finding that as we, you know, our whole thinking around what's in X was that we build it as a platform that can grow a very vibrant and a very impactful ecosystem that can actually scale this technology effectively. So short answer to your question is, I do believe that what happens in India will, you know, scale in different parts of the world as we've already seen with, you know, the likes of the payments platform that we've seen in the life, you know, the work that we've done with SBI, you know, for example, that is, you know, that's now being looked at as a way to create banking platforms around the world and so on and so forth. So I do believe that we will continue to see this notion where I've always said that IBM is very proud to operate in India where we make in India for India and then for the world because whatever we do here actually scales to different parts of the world. So let me give you a few examples of what we are doing, actually, and what we are seeing in conjunction with business partners. So for example, we launched a new innovation lab in Bangalore to expand our relationship with AWS and test prototypes of joint solutions including those related to Jenny and so that is actually working extremely well and, you know, looking at new innovations and others that are coming out of there. Similarly, in April, 2024, so it's fairly recent, we launched the IBM Microsoft Experience Zone also in Bangalore, which is a dedicated space for customers from around the world and across industries to work with our consulting team in various technology zones to co-ideate and co-create gen AI powered solutions that leverage a sure open AI service, co-pilot and other Microsoft technologies in addition to what's in X. Another very interesting example, and I can't name the client because, you know, we'll be announcing it fairly soon, but I can give you a flavor for the kind of innovation that I think is happening in India that will scale to other parts of the world. A leading partner organization in cybersecurity has recently embedded what's in X into their AI ops platform. So leveraging Jennyi, they've built an engine that sits on top of incidents and errors that offers businesses a 360 degree view of their entire enterprise. Further, they are also using large streams of synthetic data to continuously ping the enterprise instead of waiting for incidents to occur. This is in the whole cybersecurity space. So this combined capability allows them to continuously monitor for latency and other functionalities of enterprise security. And what's in X is offering, you know, the much needed transparency into how the LLM's work and ensures the privacy of their data. So these are the kinds of, I think, solutions that we are seeing in India. And I think we will continue to see how this, you know, works for India and the world. So, Mr. Buddy, you know, absolutely fascinating stuff. And I'm sure it's good to hear that India is at the forefront of this and on adoption, on thinking about new ideas. So I want to thank you again for being on the show. Thank you. Thank you so much. So you were listening to Sandhi Patel, who's managing director of IBM India and South Asia. And we'll be back again next week with another guest and a lot more insight. Till then, we're available everywhere. You'll listen to your partner. You will listen to our own devices with Nand Gopal Rajan by the Indian Express. This week's episode was edited and mixed by Suresh Pawar and produced by Meen Harakananda. If you like the show, then do share it on your social media handles and do not forget to tag us. We go by express audio. You can find us on X and Instagram. And if you have any feedback, do write to us at podcasts@inexpress.com. That's podcast within S. And do not forget to tune in next Monday at 4 p.m. for another episode of our own devices. (upbeat music)