Hello everyone, a warm welcome to each of you and thank you for joining us today for this virtual event on unleashing the power of AI for business transformation. I hope all of you are doing well and keeping safe. My name is Vijay Taripalli and I'm excited to be your host today. What is the future of AI look like? Why do enterprises need to think of AI as a key strategy pillar and a means of competitive differentiation? Is there a framework for a comprehensive approach to scale AI adoption and on how to drive innovation? Today, we are joined by three esteemed thought leaders on the topic of unleashing the power of AI for business transformation. Yugal Joshi, Partner, Everest Group, Nidhi Srivastava, vice president and Global Head, Google Business Unit, TCS Scott Penworthy, Office of the CTO, Google Cloud. Before we begin, let's take a quick look at this virtual event platform. At the bottom of your screen you will see a row of widget icons. These control the various windows which are resizable and movable by clicking on the arrows of the top right corner. Feel free to move them around to get the most of your desktop for better experience. During this panel discussion, you may want to maximize the video media window to full screen. We will be answering your questions as part of the Q&A session after the panel discussion. Do use the Q&A box below to submit your queries. Without further ado, let's welcome our panelists. Our first panelist today is Yugal Joshi. Yugal Joshi is Everest Groups business leader for technology services research and advisory memberships. Yugal has helped leading buyers and providers of technology solutions and services in technology adoption strategy, assessment of key market trends, go to market strategy, competitive insights, market positioning and expected future road map. He regularly interacts with industry leaders and drives thought leadership on core technology and technology services. Yugal's articles and commentary are regularly included in leading business news and technology media, and he regularly contributes to Everest Groups blog Hi Yugal. Hey, Vijay. Hi. Thanks much for inviting me to this very interesting and engaging panel. I think we are going to discuss a very important and pertinent topic and I'm looking forward to it. So true. A very warm welcome to you, Yugal. Our next panelist is Nidhi Srivastava. Nidhi is TCS Vice President and Global Head, Google Business, Google Cloud Business. She leads the company's Google Cloud Business unit, which is a full stack unit for cloud services. She guides companies to accelerate value from their cloud transformations and provides strategic guidance for enterprise cloud transformation, helping companies achieve agility, efficiency and scale. Hello, Nidhi. Hello, Vijay. It's wonderful to be here together this morning. I'm calling in from Chicago and I'm looking forward to a very vibrant discussion with Yugal and Scott. So it's wonderful to be here. Excited a very warm welcome community. Our next panelist is Scott Penworthy. Scott has demonstrated efficacy in reverse engineering mainframe systems and healthcare claims processing, searching for exoplanets and minerals on the moon with NASA, and cutting the costs in millions of customer chat and voice interactions. Scott has blessed with was blessed with the opportunity to build an amazing AI team without within the CTO office where the team instigated call center, AI document, AI, AI notebooks and many of Google's largest AI cloud deals. Scott holds a pH.D in AI with multiple degrees from MIT and the University of Washington. Hello. Good morning, Vijay and how are you And thanks everyone for spending a few minutes with us today. I'm experiencing RTO return to office. So I'm here in Chicago and if the lights go out and I do this because it's a sustainable office. And so anyway, I look forward to having a chat this morning and really excited about spending some time with you all today. Lovely. A warm welcome to you, Scott. A very warm welcome to you all. Thanks for joining us today. My first question is for Yugal from Everest. Yugal as a business leader for technology services, research and advisory. How do you see enterprises adopt AI and ML and how are they weaving it into their core businesses? Your thoughts please. So I think in our work with enterprises or speaking with. One thing is very clear that everybody wants to do something with AI. So we found out in one of our research that more than 72% of enterprises. Have embarked on their AI journey, so they are doing something with it and I'm not really surprised with this number because know the potential of AI is huge. And this tells you that people get it that yes, AI is extremely relevant for their business, for their future transformation. Having said that, the concerning thing for me in this research was that less than 10% of these enterprises said that they have meaningful AI running in production because what they are doing is given the promise of AI. They are running too many Pocs running in multiple direction. Just we're trying and testing. OK, this may work, this may not work. So a lot of Pocs are happening. Lot of ad hoc, ad hoc adoption is going on, but then a thoughtful implementation is still lacking. Which means very few enterprises, if at all. Anybody. I've really cracked the code of OK, how do you use AI in your business transformation? And eventually all of them realize that irrespective of the fact that today they may not be a mature adopter, but going forward, at some point of time they will need to industrialize their AI adoption. And we do see that cloud playing such an important role because a lot of the things that did not work for AI adoption earlier. Are now enabled by cloud platforms. So if you combine the power of cloud data and AI, this Trinity is really helping enterprises to drive business transformation and some of the capabilities that many cloud services provide almost, you know, out-of-the-box. AIML. Lot of APIs. Now even low code, no code. Capabilities. Enterprises can use them and then scale the scale, their AI journey eventually moving to an industrialization of AI adoption. Thank you. My next question is for Nidhi. Nidhi as the head of Business and IT consulting firm serving many large customers in a variety of industries, what is your perspective on scaling AI to shape the future of business?Yeah, So I think we have now moved to the eternal spring of digital. And. It is interesting to note that. COVID or the pandemic? Pushed us into a digital age. At a breathtaking speed we never imagined, I mean, I have to admit. And you can listen here on the panel. I think the analyst community, the consulting community, the thought leaders, the evangelists could not force the pace of digital the way. COVID did. So that's real. And it is. It is really. Eye opening to note the speed and the pace at which human behaviors can adjust. So what I did see during the course of the pandemic is that. Irrespective of age, demographics it irrespective of. The development of a country's economy. The response? At a human level was very much enabled by digital, so I've travelled to a few countries in recent times and everywhere as you go through various kinds of testing, the results are always delivered. Electronically now, so the coming back to your question, what do I see you know in terms of how AI will scale? I think the Holy Trinity that Yugal spoke about, which is data cloud and AI, I think this is going to continue to be at the center of the digital universe because you need good data to be able to train the algorithms. And the key is finding that good data because we have a plethora of data, but there is a lot of data noise as well. So how do you get to the data that matters? And then be able to find out the biases that are inherent in your data set today and make your AI explainable to me. The explainability of AI and the trust that machine learning will generate. Will drive in a big way the bold and the audacious use cases we see AI everywhere. It's there on my phone, facial recognition. Right, it is here, but the question is. How boldly will we use AI to solve societal problems? I think that's where really the challenge lies. In front of all of us. We have Scott from Google Cloud, and cloud is a huge enabler right now in this journey for AI adoption because cloud democratizes the use and the access of technologies, and Google has really been a leader on that front. Both from the standpoint of open source thinking and providing platforms to people so that you can consume AI and drive change. And last but not the least, I will say that. We need. A real sponsorship and commitment from the leaders of the businesses in embracing AI think they have embraced cloud in a big way. I have personally been witness. To how? COVID catapulted cloud adoption. And incidentally, both the words COVID and clouds start with the C and end with D. That's just uncanny, I would say. But I do see that trust, explainability, good data, and democratization of technology as some of the key things that will try massive AI adoption. Thank you. Thank you so much. My next question is for Scott. Scott, given that the teams you built instigated many of Google's largest AI cloud, first cloud, AI first cloud, what Google technologies do you see increasing in use as your customers transform their businesses? Your thoughts, Scott? I think if you go back to what Nitty was saying, Nitty, I spent the last five years trying to figure out we've got all this huge AI stack. You know, where is it useful. And if you look at COVID-19, where I've come to, the conclusion is that there's a reason Marie Kondo's in business, right. And there it goes. Look at that. It's so funny the light I do it. But the reason because in business is because humans are highly in structure. You know, I don't know what your closet looks like, but I could use some help, right. So And Marie kind of helps organize things and if you look at any business. A lot of humans deal with data that's not in the database. We talk to each other. We communicate via e-mail. That's not structured. We send each other PDF's. If you're in a hospital, you fax things still right, and a lot of the information we exchange in business is highly unstructured. And video now is becoming a first element of all business. You look at right now, we're all on video because we're some of us going back to the office, some are not. So we need technology that allows us to scale processes and deal with that kind of information. And it turns out these AI techniques are perfect for that. And so we're starting to see is where AI even like the search engine itself and say, well, how to use deep learning. I see this question from and on. If you do an image search, you're using deep learning. If you do, if you Google this morning, you're using deep learning, right. It's where these technologies are now coming in and where they're really helpful. Is they're allowing us to take processes? And things used to have to go through humans to deal with unstructured data. It's becoming the advisor to help you find an information you need at the moment you need it and process it. And you can take unstructured data, video, text, language, speech, music and then structure it and put in traditional IT systems. So that's sort of the new role that I see of AI and it's helping simple things of like how do you answer the phone when a customer calls right, and not put them on. Wait, how do you, how do you have a chat with a customer? How do you, how do you exchange my e-mail? Machines now can help with that. So that's where I really see a lot of traction coming. Wonderful. Thank you. Scott. Here's a question for you, Nidhi. In your experience of helping customers navigate their AI adoption journey, what are the strategic and operational challenges that enterprises generally face? Maybe. Good question, Vijay. There are three things that bubble up to the top. There are of course as in any massive. Technology change that has been driven over the decades. We see that you go through the constraints and the challenges and how you address or attack them drives how adoption picks up. So today I look at what are the key things. First and foremost is lack of good data. There is tremendous amount of data but. Finding. The right data to train. AI continues to be. A challenge? For organizations world over. And. At a certain level, the ability to move data on cloud. In fact, right now what I experience almost. Every single one of our clients, when I look at their cloud journey, the thing that they prioritize the most is moving data on the cloud. And I know why they're doing it. They're doing it because. In about 12 or 18 months. With the modernization, data modernization that they will do, they will have a good base to drive change in their businesses, change in their operating models using AI. For example, one of our grocery change launch retailer customer, they moved to the cloud three years back. They are the Pre COVID cloud adopters if you will and today they the way they are driving the product assortment makes in different stores the number of water simulations they are able to do. Has led to a very significant growth in their sales even in COVID times know where most of us were buying online at the know it's just like you have returned to office, there is a return to the store also taking shape now. It's good, it's always good to see more cars and the shopping malls finally you know beginning to see footfalls, it's a positive thing I would say. So the lack of data, data integration, that's one challenge. The second challenge that I I'm going to speak about is something we are all experiencing, which is lack of good talent. So good data scientists, good machine learning specialists, very, very hard. We are living and experience and experiencing what is being described as the wave of great resignation. You know, to. I mean, that's what everybody's talking about. The talent shortage is very real. So I do see that a lack of good talent also getting in the way and companies are doing different things to deal with it, whether it is. Looking at good partners, companies that can provide the services to developing talent in house as well. So lack of talent and the third thing that I will say is where we have to make more progress is in terms of. Explainability while we are testing out bold and audacious usage uses of AI, for example. People are trusting robotic assisted surgery. OK, people are going under the table. But when I reflect, what I notice is that in that scenario where it is a life and death scenario, it is AI being an extension of human intelligence because the surgeon is also there, you know you are not trusting your surgery to be done entirely by the robot. So I think that's why I have been asking myself if a person can go. On the Operation Theatre table. why would I as an individual not just putting my mother in a driverless car. I have asked myself this question and I think that I and I understood because in a driverless car probably I will trust myself because I can take control when driving, but my mother cannot drive and so I would hesitate putting her in a driverless car. So I think that the trust factor, but it'll get better today. We get into airplanes without batting an island. We don't even think. Because it has matured so much, I think we will get there. With AI and driverless cars too, so trust and explainability is the third challenge that I see. Wonderful. I recall the example of lifts when people were afraid of getting into a lift as to, you know, whether it'll take them to the higher floor, lower floor, and lift itself as a challenge at one time. It's difficult to imagine now, but getting into lifts was a challenge in terms of paradigm shift. Thank you Nidhi. Over to you Yugal. Ethics in AI is an issue that is much talked about these days. What is your perspective on the topics of explainability and bias as AI becomes mainstream yoga? I think you make a very valid statement that these are very talked about things and the reason I am saying because they are mostly talked about and the reason is. Currently, enterprises are so behind in the way they would want to adopt AI. That ethics become a good discussion point, but rarely something people really strategically think about. Which can hurt them going forward and we have had public cases in specific industries with specific clients who. Who struggled explaining to the industry, explaining to their clients with the regulators what did their AI system do? Now, ethics in general is a very complex topic, but from our vantage point, know you have a fairness factor to it. That AI system needs to be fair. And then the trustworthiness of it that other systems and humans need to trust. At the point, Nidhi also made that you may trust a robotic surgeon more than a driverless car. So those perceptions are there. But beyond this, there is one other element which Nidhi also touched upon, which is the entire explainability. And the reason I talked about fairness is if let's say, somebody does a crime, they can explain why they did it, but that explanation doesn't validate the crime. So explanatory explainability by itself isn't sufficient, but it's a very important aspect of AI, especially when we are talking about ethical adoption. Nidhi also spoke about, you know, issues around data, issues around biases. And the way we build our models, the way some of the algos are being built and more than that, the test data that is being used, a lot of it comes with their own inherent biases. Now it may not be representative of a sample set. And the reason is? The people who are driving adoption have to show some progress so many times they may end up cutting corners, which is a which is a dangerous strategy. At least we never recommend our clients to do that, but it happens. So the bias around it could be a representation bias, it could be a sampling bias, it could be another type of bias that is there plus. The way AI models have become so complex. And as you keep adding newer nodes, as you keep adding newer functionalities. And now people are also talking about a general purpose model where earlier AI models were normally built for a specific task, but now people say, OK, let's just try to build a general purpose model that can do not only NLP but something else as well. So things will become further complex and as you add more complexity, explaining that model and algorithm is very very. Difficult, almost next to impossible. And last but not the least is the risk associated with Nidhi would also talked about right autonomous cars or anything for that matter. So if an AI system. Takes a decision. Who owns it? I mean, who needs to be held accountable for it? Would the vendor be accountable? Who built that AI system or would the technology company where? Who are the building blocks of it, or somebody else or the client? Who is going to be held accountable eventually? I think those Gray areas are there. And all of this to us comes under an ethical adoption of AI that you use A where it needs to be used. Second, you use AI in a fair manner and then create systems that are trustworthy and to do that you will really need a good data set. I think there is no escape to it. There are a lot of research happening around how can you reduce the amount of data needed, but I think data has to be representative of what you want to do and then you will need to build some kind of call out the way we used to do in application development or software. Development where you have an traceability to something. So how do you build models that can trace? OK, I did this. Because of these many reasons, because AI systems are not like. Other system they can have for same input, different output or different input can produce the same output. So they are a lot more complex. And when you combine it with data challenge, when you combine with the security and the risk challenge, I think ethics become extremely important and something in enterprises need to think while they are trying to scale their AI journey, otherwise it is going to cost them in the future. Thank you. And as you rightly said, it's a very deep subject and is evolving and it will evolve through open debates, democratized debates. Thank you so much. Let me now extract. And by the way, we have come halfway to our almost reaching the mid time for our session. So from a time response perspective, I wanted to bring your attention to it. Let me now extract some more pulse of wisdom from the panelists. Scott, could you please share your perspective on a framework for AI led business transformation? And I have how many hours do I have to answer that question. You have plenty. You have plenty. Let me go back to maybe contextualize this might might help some people in the audience. I think, you know, maybe what you said earlier is spot on, right? There's how do you find the right data? We don't have the talents. Great resignation. And then how do you explain what's going on to your customer base or whoever else it is? So those are situations that are cute. We're all feeling it. You know, favorite restaurants are closing down, a lot of businesses going out because they can't find the talent. And so I think what we're going to see is a new dawn, a golden age of innovation in business because they want to stay in business. You got to pay for your kids going education. Johnny needs new shoes and you we have to prosper. And I think we're going to use to start to find areas where AI and these is a new tool for us. Let me give you an example of where I think what AI has to do is how do you take a job and make it 10 times easier. Right. And for example, going to a drive through right now we have people to help you go through a drive through. How do you make that such that AI can really take some of the mundane out of work so that we can do, we don't have to work ourselves to death. We just work as we are before. We have a nice work environment. We're doing 10 times as much work as we did before, but it doesn't feel any harder And how do you do that? So I'll give you, I'll give you one concrete example. Throughout the medical industry, this is state-of-the-art. They have these jobs called a abstractor. And if you remember what an abstractor does, but you get a degree and what you do is you, your job is to, you know, go to a, say, medical, medical document and read it and then try to extract stuff that you can put into a database for doing a sequel query. And there's hundreds of thousands of these people in medicine. And when I talk to them as well, how do you do it? And they, and they're trying to piece this stuff together. And I said, have you ever tried to search? He said, well, that's pretty hard, you know, 'cause it's not, it's non, it's not the public information, it's private information. And so there's an area where AI can really help because before it's on the public information, could we bring these technologies And if you look inside like to say that the Google search engine, we all use it all day long, right. And what's happening is you're searching for videos, it knows how to drop you right in the right part of the YouTube video. It can find the images for you. That's a that's a pretty sophisticated machinery. It's got a lot of AI baked in and we're looking at, well, where else could that be useful, right. And those and I think of if you think of that kind of mentality we do on the consumer side, imagine that in business. Where you starting to ask questions? And now these tools are basically it's like a it's like a wrench, it's like a typewriter, it's like a microwave. I think we're talking a lot about if you're building a microwave, what's the physics and what's the regulation of the physics and how do you do this safely and all? There's a lot of stuff to build the microwave. But everyone in business can go down the microwave, throw a thing of popcorn in there and get popcorn as a snack, right. And I think we're getting the point now where these microwaves will be like these new AI technologies in business, but it's going to mundane instead of having to be an abstractor and going through all these different documents by hand, can't you just go ahead and just type in a search right and things like that. And so that's what I'm seeing a lot more of is just how do you actually bring AI to the business place so that we can do 10 times as much work. It doesn't feel any different and we're much more productive. We have to figure this out. And so that's where I see these things coming in and a lot of things to think about ethics and applicability. Absolutely. Just like we have regular things in the microwave, there's a whole community around how do you build a safe microwave. We need to do that safely for an AI. But when most people are going to use it, it's going to feel like search, right? And it's going to be a natural tool because it's going to make a life a lot better. And another example is we type in Gmail and you get an auto complete sentence and it seems to know you well, That's a model for you. Well, why can't you just have and just extend it a little further and you want to draft an e-mail or you want to draft slides and here you go. Here's 50 emails to all the people at the party last night. Now you can review them. Used to have to type those things by hand. Same thing for slides, Same things for decks, same things for spreadsheets. Those are going to become question answers to an AI system. And then we're far more productive and your kids are going to say mom. You used to really have to put slides together for work. That seems so tedious. Who does that anymore, right? So I think that's where we're going. We're not there yet, but that's kind of thing. We need to get this 10X improvement. Wonderful. So it's from I want to know, so I want to know now do I want to buy have that experience right now. So what a, what, a, what an evolution that has come through us. Thank you Scott for your response on that. Over to you Yugal what approach do you advise your customers to take as they implement AI solutions of their own. So I guess given the fact that AI has been doubted as so transformational many times, what we have seen is enterprises may go the other end. I mean they may believe they have to do something so creative if they have never done. But some fundamentals rarely ever change, right. So you still have the technology part of it, you still have a people part of it and then of course you have a bigger process part of it. So these enablers continue to be important. How do you drive them may change. So for example, I think Nidhi and Scott talked about one is. What we call AI literacy. First of all, the leadership in the. In an enterprise need to be aware and convinced about the value of AI, otherwise the data earlier shared. That will continue. People will keep on doing a lot of POC, a lot of small projects here and there and be done with it, but that will. Undermining the potential of AI. So getting that literacy of AI is extremely important. I believe we all spoke about the need for talent of course, and then how AI can actually solve for that lack of time, right? The all the low code, no code development that are there can AI. And now we are also talking about AI generated code. So the point being the lack of skill of AI can actually to some extent be solved with AI IT. It's a recursive loop, but I think that is how it is. In addition, some of the fundamentals around processes now what does it mean beyond your AI strategy? So for example, business casing, why we are doing it, how we should do it, which use cases as an industry we need to prioritize. All of those things are important. But then what sort of off structure do we need and what sort of operating model do we need as an as an enterprise There will be a lot of change management. Who's going to drive it? Will our current, let's say our current organization model is going to sustain. If you remember when digital transformation really started many years ago, many enterprises created a post for CDO Chief Digital Officer. Now not many have them because they realize. This kind of a model does not work, So what is the right organization structure? Will we see a, you know, Chief AI Officer for that matter, CAIOVC, Chief Sustainability Officer, Will there be a Chief AI Officer as well? Maybe not. So every enterprise will need to take. Own bets around what works for them, what does not work for them. Last but not the least is the technology. And I think we have always assumed and always almost like Billy tells you, OK, technology is always there, we can use it. It's not that simple. AI is a complex domain. Using the right tools, right platforms, right partners is extremely important. So which type of cloud platforms are you going to bet on? Which type of data platforms are you going to build, How you are going to remove biases from your data? How you are going to build a foundation? Operationalize that foundation and then scale it so all of those elements will need to come together. For a customer AI journey to succeed and I think it is going to be the usual. Crawl, walk and run. Kind of a model, but then every enterprise will need to find. Its own way. What works for them? What doesn't work for them? Having said that, focus on internal organization structure, focus on the right technology bed and last but not the least, have your leadership buy in your business case sorted AI literacy into the organization and then be willing to drive a lot of change management. That's how at least we have seen some of the successful enterprises adopting AI at a reasonable scale. Wonderful, Yugal, thank you so much. Over to you, Scott. How will you rate the industry's adoption of AI on a scale of 1 to 10 with one being just starting to 10, where AI is fully embedded in the businesses operating model and processes? Where do you see it going next, Scott? I think it's like me throwing darts at a bullseye in a bar. Look, I think of like hitting all from zero to 10 I think depends on who you talk to Vijay mean what I'm finding more and more is that as a fascinating technology think of even a on quantum, my gosh, we actually get that running right. But. Where it's really going to be pulled into business and fast is because a lot of these jobs that we used to have with minimum wage, right, there's a lot of minimum wage jobs people just aren't showing up for anymore. You're reminded of many, many years ago, a couple, 100 years ago, we used to all have to wake up in the morning and go slop the pigs and dig dent, dig, dig dirt and you'll feed our food and go eat the food and then and there's a lot of farming jobs that just don't exist anymore for most families. I think now we're seeing it and it's a step function, change in a lot of retail. A lot of restaurants, a lot of other businesses where they say I was going to hire someone, we'll pay them a minimum wage. They'll do that job. They need that job. And people aren't showing up for it. And so now you got to figure out, I got to run a restaurant with one waiter. How am I doing that? Right? They said, well, maybe someone's going to figure out how to take a Roomba and put a dinner plate on. It's in the Roomba out there, right? That's a robot. Now it's just a Roomba with a dinner plate, right. So we're going to start to see innovation. Same thing for people that, you know, you do a drive up right now, we have minimum wage people there and you're seeing drive ups, you know, closing down. Why isn't that like a call center that you're talking to? Right. And I think we're going to see a lot of innovation pulling in saying. Here's our tasks that humans just don't want to do anymore. How can we automate that task and change our process so that we can accommodate much fewer people and actually allow do the same process we had before. So I think we're seeing a lot of innovation in the retail. There's a lot of innovation in the food industry, in the service industry where they're just people aren't showing for those jobs and we've got to figure out how to stay in business. Well, I tell you there's nothing like that kind of pressure to figure out how do I bring an AI, how do I take a lot of this paperwork out and use AI for that technique. And that's going to bring, I think a rush of these technologies in. As tools to make a sort of like superhuman, the sense of we now have these tools, let us do things before we had to use labor for that labor doesn't exist. People want that job. So I think it's going to force us to rethink that. And we're seeing it now in droves. You know, just in your local hometown, right? We like to go to the corner restaurant. They can't find people. Same thing for a yoga retreat, like how do you, how do you do this? So I think we're going to see tremendous innovation here in just the next few years now because people won't stay in business. And there'll be an entrepreneurs who are going to figure out a new way to do things. AI, where you're reading, you're writing, you're seeing, you're perceiving. And those are technologies now that we didn't have just five years ago, right. And that's where I see a lot of this coming in. And I think then that'll help us think through as you bring it in, what are the ethics of this tool as it's been approved and all those issues that you guys bringing up now become as part of that tool. This is a safe tool, right? And I think that's where I'm going to see an awful lot of innovation. I'm already seeing it personally in these enterprises, right, as they go through these transformations. So that's my two cents. Thank you. Thank you, Scott. I'm going to pick up some questions from the audience and I would leave it to the panel to decide on how you would like to respond and who would like to respond. How do you see AI playing a role? This is from Senthil. How do you see AI playing a role in hyper personalizing the customer experiences based on intelligent insights? Hyper personalizing the customer experience is based on intelligent insights. Right. I think I'll just put two cents in here. You want to go? Why don't you go first? OK, I just wanted to bring in a moment of just over here. I wanted to say that. AI And especially if you look at AI in a retail experience, the way it is able to read your mind and predict your interest is sometimes getting even better than your spouse's ability to know what you want for your anniversary. It is, you know, it's maturing to that level. So it's scary also at times, I will say, but I do think the. If you look at the way how we have learned to shop in the last two years, we've really embraced online shopping and it's going to be now an irreversible change. You to go back to my old behaviors and the ability of the algorithm to remember what I was searching for, when I was searching for and then to give me recommendations is something because you know that's where it is rivaling. The human mind's ability to retain information and leverage it. But I will say one thing because I do want to hear from Scott as well that we must remember, and this is my firm belief, AI will always be an extension to human intelligence. It will never be able to replace it. So I think if we look at it from that standpoint, I think. We are only experiencing hyper personalization. And we will begin to experience it more and more because just randomly, in the middle of my work day, sometimes Uber Eats will tell me that depending on my general food preferences, a restaurant has a deal out there. Would I like to order something? So it is uncanny, interesting, but something that we all will embrace a lot. So let's hear from Scott. Yeah. A couple things are Nidhi. One is we do have superhuman technologies today that we buy. You can buy a superhuman thumb today at Home Depot. The superhuman thumb is called a hammer, right. It does far more than we ever could in a finger. But we've gone from a rock to a hammer and it's pretty and we can buy for a couple of bucks. It's superhuman. I think we're going to see digital intelligence as a superhuman capability. But it's a tool just like a hammer, right. And we're and you're and you just pointed to 1 area with it is shopping right now still. We still browse, we look, we shop. But you could probably see indications of the trend where it's starting to know really what you want. So once you flip it. So at some point, why don't you just say just ship me what I need, I'll return what I don't want. And so now we're having robots that are being programmed to go, it looks like a beer cooler with wheels and an antenna, and it goes there. And his job is to go pick up stuff that you don't want. And so at some point does e-commerce flip and now you have a provider that knows you well and that's where you're seeing with clothing, we're seeing very early versions with clothing. If you buy some want to subscribe, get a 10% discount because they have a, they have annuity net streams from you now because they know that's what you want. And so it's in their commercials best interest because they can't hire the sales people. No one wants to stand up for 10 hours a day in retail anymore, right? They don't want to answer the phone. So why not just ship from the warehouse to you? And if you're an annuity stream, there's a there's a financial incentive to get it right and that'll be a superhuman ability to understand what you need. But it flips e-commerce with a lot less labor. That's just one example of this, where the idea of personalization, understanding who you are. A lot of times we did this at Google, we stepped back and saying do we really want to deploy this? And we go through the whole thing. We look at the ethics of it, the privacy of everything else. A lot of times we don't even do that metric because it's just it's a creepy factor. We don't put it in. And So what you want to do is figure out what's right for the customer. But I think the ability to understand patterns and what your behavior is showing versus what you think it is showing, no wonder it's it may know what's better for, you know, recommending for your anniversary. Like you want to give a birthday present for your daughter and like, what do I get for a 12 year old? How about this? Their friends are doing it. Oh, that's pretty cool. What does it shipped you a week ahead of time with the wrapping you've liked and pre wrap like that'd be amazing, right? And so I think that's what we're starting to see and that be 1 innovation driven by we can't find people to stand up 15 hours and wait around in a shoe store anymore. What do you do? Brilliant, you got your thoughts? No, I think most of these things are covered, hyper Personalization and Nidhi also said. As an end user, it sometimes scares me, Yeah, OK, what the heck is going on. But at the same time. Businesses know that this is needed, but the interesting aspect of AI will be it will allow you to do it in a scale manner. So what is called mass personalization or hyper personalization at scale otherwise. As a business you cannot keep chasing individual as your target segment. So but you will create your strategies and then use use AI to target individuals around. OK, this is where personalization is needed. And this is where personalization is not needed, because many times human beings want to be part of a cohort. So the AI system need to understand when they, when this person will be a part, will prefer to be a part of the cohort versus an individual who needs to be served individually, if I may say so. But of course AI has a has a very interesting and an important role to play, otherwise we won't be able to do it. I think it's as simple as that. Brilliant. So in the process of I've seen plethora of choice architects coming up to influence the decision making process of consumers. So brilliantly said, thank you for sharing your opinion. I have a question from. Avimanu. It's on biases. Since AI learned from our past knowledge facts which might have positive and negative issue biases, how do we formulate a system to detect and minimize the biases, specifically in the criminal justice system? Who would like to go first. Should I repeat the question? OK, that's a question. Yes, I agree. And we will only take Scott's view for this one. There are lots of questions. So we'll try and get through as many as we can. Yeah, lots of questions. Yeah. Thank you. Yes. So I think it's here going back to me, it's, it comes back to data and understanding really the origins of data and just how careful we have to be with it. So there's a very famous. Article in the last decade about recidivism. What's the chances of this potential person, they're going to come back to jail, right. And they used to give recidivism scores and what was what we're finding is that these things were highly biased. And to talk to the math nerds behind this and what do you mean? I I'm not putting anything in there about your background, about your racial background. But you look at the questions and the questions in the data they were collecting was collected as a whole. Like what zip code were you from? Do anyone who's been shot in these other areas? And if you look at that statistically, that's a tell for a particular demographic, right. And they were using that demographic. It's just like another very famous study in AI where it was an amazing like a 99% fit of detecting breast cancer. And they said, why is this so good? We've been doing this for years. It's not so great between two hospitals. And he did an analysis of where the model was thinking, and he realized it was zooming in the lower part of the picture, like what's going on in the X-ray. And it turns out it was looking at the serial number of the X-ray machine, if it came from the inner city versus another city or a different location. It was saying if it, if it's close, just call it cancer. And they're like, that's not what we want. So because AI in some sense is lazy, it finds the pattern. Just be careful what you ask for. And So what that means is that you're going through. That's why a lot of times we have. New AI products. AI just amplifies our own biases. That's something we have to function with bias. That's how we don't go crazy. It's called like, what do we focus on? Because there's all these sounds going around that you know, how your socks feel on your feet, how your shoes feel on your feet. Probably think about that now. How's the chair against your left arm? Thinking about that now you weren't before. Your brain has to bias against all information. Focus on something so you don't go crazy. So what happens is, like Nidhi, what she's focusing on and Vijay, what you're focusing on, you go focusing on. We're all different. And so we come together, You go, well, says, I know what the answer is. That's, well, what's the answer? And he goes, it's as big as a wall and you can't move it in a mobile thing. Like, really, if he's just like, no, it's not, It's bushy, it's bushy and it's easy to move. And then he's like, what are you talking about? It's big and flappy. It's like, AI don't know. It's like a big piece of leather. We're all talking about an elephant, right? Because we've all basically looked at our own biases and as a technology where if you're going to ship something, you've got to have people, different backgrounds, with different biases at the table. And then collectively you can look at it and then say, I see what you mean. You shouldn't ask for zip code. Have you been shot and do you know anyone who's been in jail? Because that's just that's a tell for a particular demographic. Get those questions off right. But you that will only come out if you're with people don't have the same biases as you do. It's almost guaranteed if you look at an AI team and all the same background that probably is going to be biased and they're not going to see it because we all have hidden biases. And so I think that's where a lot of this comes out is that you know it's a really good really good question. It's sort of the, you know, the trolley issue in self driving cars. You know a trolley's coming across how do you stop, you know, someone's going to have to die. These may be artificial scenarios, but it comes down to if we're shipping a product, how do you then make sure that you have. We try to reveal these biases we have as human beings, right? And the algorithm itself is also biased that we talked about for the breast cancer or civilism. It'll find the pattern right? Just be careful what pattern it finds, right? And as you do, how do we evaluate the pattern so that. Is it bushy? Is it a wall? Is it a flappy leather thing? No, it's an elephant. Very interesting. Very interesting. Ashish has a question. In healthcare, we are seeing a significant use of AI. However, in US we are still seeing it limited in limited use in the area of medical diagnosis. Do you agree and when you think it will be mainstream? Yeah. I mean, it's been the last couple years with the AI in healthcare. I don't want it anywhere near a diagnosis, right. So in the sense that AI is a tool just like just like an X-ray machine, just like anything else. It's a very powerful tool. And it's really good for helping caregivers get to the right information at the right time. Right. So think if you're going in, you're going for help and you're saying, how do I, what's going on with me that how to? And they only see you for 5 minutes, 7 minutes. How do you make that really productive for that doctor? And how do you get the information she needs at her fingertips so she can make a better decision? Right. So can you use AI as a technique to say there's all these latest reports? All these latest new treatment that are posted, all these new drug trials and new information coming out, and she's just a real doctor in Missouri. How do you get her the information she needs at that time so you can have a better diagnosis and treatment? Like was the inner city, Philadelphia? You can't afford to go to the doctor. You know the parking is $35. You don't have $35.00. So how do you, how do you do telemedicine and how do you use that, which is like again, video and information where A is now assisting the doctor, right, and where the doctor makes a decision. But it's sort of like having, you know, the Physician's Desk Reference and be able to search with that instantaneously. To help the doctor in her decision making, right. That's where I see a lot. And AI has made that mistake very early on, the very early AI. When I went to school at MIT, we were studying symbolic AI long before deep learning. And it's very classic systems where they try to use as a direct diagnosis. And because there's so many subtle things in one of those complex systems in the world, Our bodies, our bodies are amazing machines. They repair themselves. Wouldn't they be great if your car did that? But it does, right? So in order to understand that, we need to have the human element. But how do we use as a tool to help her? Make those decisions better. By getting the right information at the right time, I think that's where we're going to see a lot of AI. And AI may have an indication of these areas interesting on a skin to look at. Let's take a look at those as opposed to that. You definitely have a polyp in your column in your colon. These areas might want to take a look at and have the doctor then decide, right. That's what we're seeing, a real effective use of AI in healthcare. And a lot of it comes down to things like just scheduling nurses, you know, how do you, how do you know what nurse and how do you and can you use like some basic math to help nurses, you know, find a job. And makes it make the job less stressful. So that's where I see AI and medicine an awful lot. I think it gets a lot of the press are the diagnostics. There's something we're doing at Google, which is not only can AI read what's in your blood, it's basically coming like Grail or Oxford Nanopore using AI to basically understand what's in your blood. We can now predict the structure of those things using AI. And we have a company called Isomorphic Labs, but that's assisting in drug design, but it's not designing the drug, it's helping the drug designer do a better job. And I think that we're finding is really lot getting a lot of traction not in what we first read about is like, wow, it can diagnose a breast cancer, it can diagnose these things. That's a really interesting pattern recognition. But in practice, let the doctor make the decision, right put this in her hands and use it as a tell for information. Yeah, very, very good insight, Scott and Ashish. The point I will make is that this is the industry where we will see a lot of AI being used in the next few years. And some of it you would have seen that telemedicine became a real thing in the last two years and prior to that also telemedicine was available. But I would just feel more confident having met the physician in person. At the GPS office, so but the trust was built. The other thing that I will say where I think AI will have a huge role to play in Healthcare is that as the technologies as cloud and. The ability to do slicing and dicing of data becomes more and more affordable. You will see a big uptake in. The technologies being consumed by healthcare, whether it's by doctors or. You know whether you're doing scans of X-rays, looking for cancers and even in personalized drug. Design for patients, I think it is a few years away, maybe anywhere between 3:00 to 5:00, but it is coming. And I think this is an industry which is poised very well. Retail has used it, banking has used it. They will continue to use, but Healthcare is poised for the growth. Is my view. Wonderful. I think this is an amazing audience. The kind of questions that are pouring in are difficult to be handled in the short span of time. So I would request the audience to be patient. We will respond offline from the panel side. So I would have one question that I had in mind and I would like to ask. And I would like to ask all the three panel members, beginning with Yugal, then possibly Scott and then Nidhi. So the question is what lies ahead for AI? Maybe a minute each, please, starting with Yugal. So I think. Yeah, go ahead. So I guess adoption may be cautious adoption, but definitely an adoption and using AI as it permeates different part of our lives, our work and social fabric. So that will continue. I think it's a one way St. much like cloud wars and. To be fair, AI has been around like forever. And but has not got the needed push from the required building blocks which are probably now and they are further evolving. So I'm quite confident and hopeful that going forward. Enterprises and consumers like us are going to adopt AI lot more strategically and be more willing to trust AI more than probably what we do now. Thank you. One minute, Scott. What lies ahead? What lies ahead? For AI know I want to see it to be as easy as search where you know, there's a, there's a 10,000 people who are really good at it and they're designized algorithms and they think very thoughtfully through it. But it's kind of thing where, you know, I would love a world where you're an enterprise and you say, well, I got all this data, it's OK. Well, let's get that data ready. We'll put into a data site, like a website and maybe TCS helps you build that. Right. You got a data site. OK, We'll pick a surface. Well, you want to talk to it. You want an API, Put in your process. Do you want to, you want to chat with it? Do you want to just search and have a panel come up? Well, we pick a surface. And then ask away. Right. And then the idea is that can you look at a business process and say where do I need information at my fingertips because I don't have the labor to do it anymore, right. They can make my decision. And that to me as is a wonderful future. And you know, I think our CEO talked about the future of Google and AI search. I'm like what I think we're talking about. How do you apply that technology? It's rich in AI, but to basically help solve IR problems, information people, problems in business and they're everywhere. And I think that we're going to start to use AI and if you can Google. You can use AI. That's where I see it going. Nidhi, over to you. What lies ahead for AI? My perspective is very bullish. I think we will see a lot of AI. We will also see. New businesses, new offerings and new jobs. And we will see some jobs being sunset as well. For example, I'll give you Uber Eats as an example. So we have what was originally. A transportation offering from Uber. And we had the restaurant business and how the two came together to create a new offering. So I expect to see a lot more blurry of lines learning cross pollination between industries because of the digital front the business will have. So your ability to share data, your ability to share insights and then build something. As trust and collaboration builds, it won't happen that the forces of business and economics and consumer forces will be there, but I see that. And I see AI alleviating human beings from doing a lot of repetitive jobs. Those are the jobs that the machine will take over, just like there were lots of jobs taken over by the machine by the industrialization that we have seen during our history. So you machines will take our jobs away. So and yes, so will machine learning, but we have to be prepared for it but and but at the end of the day human mind. Human intelligent intelligence. Trump's artificial intelligence. So I still see that it's I have an optimistic view of the future with AI. And I hope to get into a driver. Wonderful. This has been a great discussion. Unfortunately, that is all the time we have today. Our thought leaders have presented A framework for a comprehensive approach to scale AI adoption along with how to drive innovation. You will share his insights on how enterprises adopt AI and ML and are giving into their core businesses about ethics in AI. You will share his perspective on the topics of explainability. And bias as AI becomes mainstream on challenges that enterprises face while navigating through their AI adoption journey, Nidhi gave her insights on ways to overcome some of the most common ones did. He also succinctly outlined along with Scott and Yugal on what lies ahead for AI. This has been a wonderful audience that has shared a lot of questions and we're going to get back offline. And Scott, of course, talked about Google technologies increased in use as customers transform their businesses. Share this perspective on a framework for AI led business transformation and it's assessment on the industry's adoption of AI and on where it is heading. Thank you. Yugal Nidhi Scott for the wonderful discussion and for sharing your rich insights with us. Thank you so much dear audience to learn how TCS and Google Cloud can help you embrace cloud for purpose led sustainable growth contactgpu.marketing@tcs.com. With that we will bring today's session to a close.