Generative AI – Implications on Enterprises, IT Services & Beyond

Sairam:
Good morning, everyone. my name is Sairam Vedam. I welcome you all to this exciting series of the Cigniti Podcast editions. these podcasts and our digital dialogue flagship editions are going to cover some intricate aspects of modern-day technologies that are disrupting as we all, go through these interesting times. Well, generative AI that’s taken us all by storm. It could be chatGPT , it could be BARD, and there are a myriad of technologies that are flavors and variants of that as, a leading IT services organization. Obviously, we are looking at generative AI and its impact, how it could enhance productivity for us. What does it mean to our customers and more than all of us as an ecosystem Today, on that note, I’m very excited, thrilled, to be honest, to be having Srini Athreya, Chief Data Scientist. joining me today, a practitioner, who does significant amount of work, not just on generative ai, but a whole bunch of ai ml and data engineering technologies. And he’s someone who’s been at the forefront of, implementing, and I would also love to use the word, experimenting a lot of these technologies in our labs and putting to them for effective business use cases that have transformed some of our customers’ digital initiatives. And he also, someone who’s very passionate about teaching, sharing his insights more from a pragmatic standpoint and not just at a theoretical touch level. That said, let me welcome, Srini. Srini, How are you doing today?

Srini
I’m doing good, and thanks for inviting, Sai. Good morning everyone, and thanks for tuning in.

Sairam
Well, before we, we go into whatever we wanna discuss today, Srini, I just read the stat, which was staggering to say the least. The Goldman Sachs report tells about 18% of the world’s workforce could get computerized. Impact of generated wear is as sweeping as that. I’m sure there are tons of things that have been already written, but let’s, let’s get to this today. Simplify this whole thing and probably share your insight. So you, are you looking forward to this?

Srini
Oh, Absolutely. Absolutely. Sure.

Sairam
It’s Gonna be fun. Well, what’s the future of generative AI and it’s ethical implications? Maybe that’s something that we could start off with.

Srini
Yeah, thanks for the, the most difficult question right at the beginning, I guess, but yeah, so predictions are always dangerous, and so I’ll be careful with whatever I say here. future of generative AI. see, generative AI is here for sure, and it’s going to change, the way all of us work, you know, that, that much is, is, is, is now pretty clear. now how exactly it’ll unfold. I guess we will have to wait a little bit. We still, you know, we still don’t have all the answers, but, it’s ethical implications. Now that is a, now that is an important one, right? Because fundamentally generative AI has the capability to multiply, the, the human impact many, many fold. If it’s good, it’s gonna be very, very good. If it’s gonna be bad, it’s gonna be very, very bad. And I think that’s where the, the biggest challenge for us is mm-hmm. in terms of, in terms of how do we want to use it and things like that. And the biggest thing here si, is it’s not really about the models, right? It’s about the data. And we really have to be careful about the type of data that we use to train these models, okay? And right now, you know, all of we are doing is we’re just scraping data off internet and kind of, giving it to these models. And that’s something that I am not personally comfortable with. It’s, imagine you’re, you’re teaching kids, you gotta curate textbooks, right? You can’t just give them, you, I can’t just, scrape all the calculus stuff off the internet and ask somebody to study calculus. That’s not possible. That’s what we’re trying to do with these models now.

Srini
So that to me, is an ethical concern. And, hopefully we will get around to solving it for the next.

Sairam
I mean, I agree. It’s all about the data. Well, there are all these leading companies, meta, Microsoft, Google, yes. Big, big, big names. How do you think they put Generative AI to work responsibly?

Srini
Yeah. At least the one thing that’s happening right now is they’re being kept honest, frankly, because of so much open source community. And there’s all of this spotlight at at least these guys are, these guys are being kept honest to a certain level in terms of what they’re trying to do now. one big question, of course, is how much of it is truly open sourced? How much of it is known and how much of it is unknown? And things like that. With open ai, there have been a few concerns raised, especially with Microsoft and Open AI, because they have not really published too many technical papers detailing how, how they have trained open AI and things like that. The last paper they truly published was in Jan 22, which was a paper called In Start GPT Post that they haven’t done it. But, we are hoping that as time goes along, they will publish a lot more details. I think Google has been a lot more forthcoming, honestly, in, in, in this particular aspect, Facebook too. So they are probably a little, more transparent at this point. But having said that, the hope is that Open AI will do it too, and, and we’ll all get to know a little bit more.

Sairam
Sure. Looking forward to that. Well, you know, you, you’ve been working with enterprise customers, for a long time. Yeah. As far as chatGPT is concerned, do you think, is it ready for enterprise segment?

Srini
Truly, The short answer is it is and it is not. Okay. And I’ll try and, and, and, and, and kind of, explain that a little bit. It is, because like all revolutionary technologies, it’s not 100% ready, but it’s never going to be 100% ready. Right? You gotta take it and make it ready for yourself. So the owners at this point, lies on a lot of enterprises to do it. Okay? Because if they don’t do it, they’ll be left behind, which is not something that you really want to do in today’s day and age. so, so in that respect, I think there’s, there is still some work that needs to be done by the enterprises to really start using them, but I think it is ready from a use case perspective, cuz there are definitely efficiency use cases, automation use cases, that can be today used, by, by enterprises, right? And when I say it is not probably the biggest concern is around security. Mm. Okay. what exactly happens? And, I think all of us read just yesterday or something, Samsung had this big problem where some Samsung engineers uploaded some very confidential R&D data to, to chatGPT. And, it’s now creating issues for Samsung, right? Because all of that data is now with open ai, because they, they’ve kind of uploaded that. So, so there are security concerns for sure, and organizations need to have clear policies around what needs to go in, what does not need to go in, and things like that. So at that level, there’s still work to be done. But definitely from an efficiency and automation perspective, it is ready. It need needs to, and like I keep saying to a lot of people, the first step would be to just let your people experiment with it. Right. You know, don’t try it, ban it, don’t try it high, that’s not gonna work. Okay. You kind of first let your people play with it for, for, for, for a few weeks or months and use cases will bubble out. Got it. Themselves. Okay. So that’s the way I’ll have to Say it.

Sairam
So, as a continuation to that, why do you think some companies are still scared, if I could take the liberty to use that word, scared of this whole generative ai? Is it lack of understanding or is it, you know, an inhibition or the ability to experiment? What could be that?

Srini
It’s all of those, right? It’s all of those, and this is not true necessarily of just chatGPT, right? Any new technology that comes in, the first reaction is, is that of fear. Mm-hmm. Because we fear the unknown. When we don’t know something, we fear it, and it’s a natural sort of, reaction to anything unknown. But, but again, history is a kind of proof that people who have gone beyond that fear and embraced those technologies have performed a lot better than organizations that haven’t. So I’m not saying that the fear is irrational. The fear is absolutely rational. The only point is you gotta go beyond it. You gotta see what you can do by minimizing your risk, but still taking advantage. So I think that’s really the point. And why are they scared? They’re basically scared of the same reasons. They’re scared of security reasons. They’re, they’re, they’re scared of ethical reasons. They’re scared of legal reasons, and most fundamentally, they’re scared because they don’t understand it. Got it. Okay. So, so, but the effort needs to be, to try and understand. And, and I think that’s where companies like ourselves can play a big role Right. In, in, in kind of taking this.

Sairam
Interesting, well, you know, we all know Google has been at the forefront of AI revolution. In fact, they were the ones who did a lot of interesting stuff, starting with deep point. But with chatGPT, do you think, I mean, is this a really a race or is this a two schools of thought? Or is there a reason for them to be worried about?

Srini
you mean Google? Yeah. Yeah. No, this is a difficult one. I, I honestly don’t know enough about internals of Google to say that. But, see architectures, like for example, like Lambda okay. Have been around for quite, quite a bit of time for now. Okay. And, and then they’ve been very, very good. If you look at architectures like Palm, for example, which is a different architecture from, transformer based, architectures like, like Lambda or like GPT or whatever. Google has been doing a lot of work on this, but I think they have been taken off guard a little bit. Okay, got it. In terms of, how quickly this has been kind of taken up. And, and the, and, and the other thing was, I think they were a little, I, I don’t know the scared is the right word, but they were probably a little more apprehensive about really releasing it into the wild, the way open AI did it, open AI just put it out there, right. And, and, that succeeded. But imagine what would’ve happened if, they would not have succeeded. They’ll put it out there and there was a lot of bad news, then they were also hit open high heavily. So maybe Google was a little circumstantial, but I will definitely not write them off. No.

Sairam
No way.I don’t think I, my personal view is you can’t write them off on, this is not about whether who the race, so does a generative AI, I mean, obviously human creativity right, is at the center of it. I was reading something where if you go and ask a prompt and tell that, give me something like a Yeah. And if you ask it to give something that Da Vinci has done right, the outputs are fairly different. So with that as an example, do you think Generative AI has got the power to replace anything on human creativity? Or is it going to be an assistive technology all the way?

Srini
Right now it looks assistive. Mm-hmm. just because more from a risk perspective. Okay. We just don’t know what it doesn’t know. But I’ll also say something else Okay. For a long time. and if you, maybe I’ll take you through a bit of history here. See, for the last five, 600 years, if you actually go through, you know, scientific advances in human history, about till about 600 years ago, we used to think that we were the center of the universe. Right. And then Copernicus and Galileo kind of proved, that’s not the point. Okay? The, actually we revolve around the sun, and then Newton kind of proved that the, that these are actually the same laws that govern the heavens that govern the earth. So the special, this, the specialty was taken away, then I think Darwin came about and basically said, you know, I’m not very sure if God made man in his own image. We just evolved from chimps, So he took away that specialty too. So I think it’s just the same path going on today. We think a lot of human thought and human language and human creativity are somehow special. Mm. But, that’s being brought into question right now. Okay. Because you’re kind of saying now that whatever we thought were special or maybe not that special, maybe you can actually automate them to say next. And so I think it’s just the continuing, scientific advancement that, that keeps going on. So in that sense, I would say right now it isn’t. But am I to say that it’ll never replace anything human? I wouldn’t go that far. Got it. I, I don’t know what can happen.

Sairam
What would you think would the impact from a human computer interaction, is there an impact that generate?

Srini
Oh, abso ab Ab absolutely. Right. Because, I, I read the street from Andrea Karpathi recently where he said the most important programming language is English now. Hmm. So, so, yes. Absolutely. So, so now you are, you are communicating with computers, using natural language that wasn’t possible. Definitely not to this level, beyond this day. So, Elon Musk is talking about Neurolink type of stuff now. I don’t know what that will do to, to machine human interfaces. So it’s definitely revolutionary. They, they, there saw two ways about it.

Sairam
Yeah. You mean I was, I was looking at a, a marketplace of prompts. Okay. Yeah. Yeah. It’s already getting in.

Srini
Yeah, Absolutely. Yeah.

Sairam
Right. And, and then the kids have started, they’re saying that why English? I could actually quiz chatGPT prompt in local language itself.

Srini
Oh, Absolutely. I think, I’ll not be surprised if, if the next couple of years prompt engineering will actually be a cause in colleges. Absolutely. Well, they’ll actually teach you that. Yeah, because, cuz I think it is a real skill to be very honest with you. It is a real skill. I’m learning it. I’m, I’m nowhere as good as I probably should be, I’m learning it. and, and I think it’s a skill that all of us have to learn.

Sairam
Yeah. I think linguist, conversational designers, these are going to be the roles. What are the limitations that you see from a chatGPT perspective?

Srini
See, one simple thing is, at the end of the day, all of these are auto aggressive models. When I say auto aggressive, essentially what they do is they kind of look at text generated till that movement and then try and generate the next word. Right? So it’s an auto aggressive model. Okay. So it has no idea, for example, of what will be the end output till it generates the entire output. So, so any question to the fact, like for example, if you and I, we are talking about something, even before I say the next word, I have a vague idea of what I’m going to be talking about. Mm-hmm. And then I say words in order to say whatever I want to say, A chatGPT can’t do that. A chatGPT has no idea what will come out till all the words come out. Sure. So, so, so that’s a huge thing. So for example, and that’s why it’s easy to Mel to manipulate it, if you know how to manipulate it, you can manipulate it because it, it can’t really understand, the, the final output before it generates the whole thing. That is a huge limitation of auto aggressive models. And that’s sent out, that’s a problem with chatGPT. And the biggest problem is you cannot solve this with this architecture. Got it. It’s not about more data or more parameters or whatever you do. Fundamentally, this is going to be a limitation with this family.

Sairam
So Whether it is GPT3 or GPT4, 5, they fundamentally change somewhere.

Srini
You can’t really solve that problem. And that is one of the biggest issues.

Sairam
My understanding on other perspective is enterprise conversational AI platforms, right? Yeah. Which is which feed on enterprise data. Okay. So do you think, what’s your opinion chatGPT feeds into that? it, it can add more value or conversational platforms are on top of chatGPT. Where do you see this whole?

Srini
Yeah, yeah, yeah. I think, I think models like chatGPT will become foundational bottles. Mm-hmm. but they, they, there is still a place to fine tune. Okay. so I think conversational agents will live. Right. but these will be the foundational models on which they will be bit, it’ll, it’s not gonna be, they’re not competing with each other. I don’t think so.

Sairam
Got It. Got it. So how do you think businesses could safeguard their data? I think this is, that’s the whole thing that’s running whenever such technologies come in. Yeah. Privacy, security risk from an architecture standpoint.

Srini
Also, See, at the end of the day, really, when enterprises, when this goes full-blown enterprise point, okay, when, let’s say two years time or three years time when this happens, people will have to train their own versions of these models. They will have to fine tune them, okay? They will not be able to just use APIs any, any, any longer. there may be some problems for which the API sort of things will work, which are the more transactional things. For example, if you run a call center, maybe you don’t want to take the effort of doing all of this, but if you’re trying to use this within your core r and d or core business processes, I don’t think an API based approach will, will work in the, in the longer term. Because there are things like you can hack prompts, okay? You can do adversarial, prompt engineering, you can do all sorts of things to make the model do or say things that are very, very dangerous. Got it. So, so, and, and today there are no safeguards today, there is no way you can control that. If there is someone who knows how to do it, they will do it. Understood. And so, I, I honestly think these architectures will be embedded within organizations mm-hmm. and there will be a way where open AI or Microsoft will actually have an entire business line where they will fine tune their models for you. Got it. So, so that’s the way it’s gonna be work. It’s not gonna be a pure API based business anymore.

Sairam
Got it. So how do you think in that sense, you know, from a customer experience, delivering customer experience, is there a role chat GPT can play? Can we take advantage of whatever it does?

Srini
Yeah, absolutely right. Na, natural language mm-hmm. So today people have to, today, for example, today’s chat bots, you have very, very controlled conversations, right? You a few options and you select this or that or that You have entire IVR products, for example. Okay. Which kind of contact the Transformation to contact. So, you don’t need any of that, right? This is pure conversational AI. You can just stop, and you can just get your stuff done. What’s more natural than that? So, customer experience is definitely getting changed. Of course, it’ll come back to the whole thing of about security. Mm. Okay. And, and, and that’s a problem that we will have to solve. And, but I also think looking at the advances between three and four, there are so many, it’s, it’s already tremendous, right? With three, you can do a, you could do a lot of stuff with chat GPT to make it do wrong things. With four you can’t, yeah. With four is, it’s already a lot tighter, much More tighter. You can’t do that. So, as I, as I’m seeing five, six, it’s definitely gonna get a lot better.

Sairam
Sorry, Understand. Well, here is a radical statement that I read. Businesses that ignore ChatGPT could find themselves left in the dust. What’s your opinion?

Srini
How true is this Sitting where I’m sitting right now? I think it is fairly true, in the sense that I keep going back to the industrial revolution example, because that’s the classical example. And, countries, we, for example, in India especially, have missed that revolution. We’re still trying to catch up 300 years later. True. Okay. so I, I honestly think companies that that kind of, do not take this up today will not end up being leaders in their domains a few years ahead. So if you, if you want to do that, you gotta take it up, of course, with all the caveats, understanding all the risks, but then you gotta take it up. Okay. The answer is not to ignore it because this is a transformative technology. This century is going to be called the century of AI and generating AI. So we might as well get on board with it.

Sairam
So do you, do you believe this is a seminal moment that we are looking at?

Srini
Absolutely. This is, this is because say people don’t necessarily understand the importance of this. The, the point is still now human decision making was something that there was no way to automate. Right? If you look at the entire IT industry and the, the whole technology industry, what did we do through the eighties, through the nineties, we used things like SAP and Oracle. You automated the business processes, right? But wherever the decision was to be made, you still had a human sitting there. True. Today you are fundamentally taking, they saying you have a technology mm-hmm. but can in part assist the human in making a better decision. Right. So it’s fundamentally transformative. Okay. We’ve never, the hu humans were the most intelligent species on this planet. Right. We never had anything even remotely comparable to us. Correct. Today we have a technology that can actually work with us. Right? Okay. and the hope is we will work with it too in order to make this, world a better place. So I think it’s absolutely seminal. Okay. This was, this has been a dream for 78 Years now.

Sairam
It’s, it’s a natural human twin, not just a digital twin there. Yeah, exactly. Well, to what extent generative AI in that case would, would be able to change the whole perspective of paradigm of user interface design. They say no UI is the UI these days.

Srini
Exactly. Absolutely. Absolutely. Because when you and I are talking, for example, we are not thinking in the context of a user interface. Here are we, we are just talking. Correct. Because that’s an actual tool.

Sairam
Why is the interface?

Srini
Yeah. So similarly, when when you have these sort of technologies, you don’t have an UI anymore, you just, you just behave as if you are, it’s already happening.

Sairam
Anything in terms of, speculative design and future thinking that generative AI could come in, meaning, new things that could come up.

Srini
New things, yeah. Yeah. New things that, that could, Well, anything.

Sairam
I mean, from a future perspective, that is something that we couldn’t think of. Is there anything that you, you have in mind that generative AI can, can play a role?

Srini
Yeah. absolutely. See there was, there’s always this one thing, right? We, okay, may, maybe I’ll take a little bit of, detour here. But see there, there are problems in the world that are considered reducible and irreducible. What, what do I mean by that? Okay. There are certain problems where there are patterns that repeat true. So you can actually learn and then you can understand how this will unroll. For example, human language for, for all its complexity has patterns, what’s in and semantical. And you can kind of do that, right? But for example, now you are looking at something like weather, let us say, right? Okay. now weather is a chaotic system in the sense that you cannot even given a set of initial conditions and given the laws governing weather, you cannot predict whether perfectly it’s just not possible. But with this scale of ai, where it can actually roll over and it can generate multiple instances, okay? It may be finally possible to start looking at long range when weather forecasting. Now this is a problem That it’s been there for forever, Ages forever, right? Humans have always, right from the ancient Greeks to the hing, ancient, we wanted to, sow our crops, we wanted to know how, how the weather will look. We, we couldn’t Right. Today maybe, and, and again, this is a controversial statement, okay? So, there are people who believe this is possible. There are also people who believe that it is so irreducible that you actually can’t even do that. But, but then I, I hope that, sometime in the next, in near future, we should be able.

Sairam
Yeah. In fact, I was listening to the futurist to recently set that while countries like America and China have taken a leapfrog, in the AI implementations, India seems to have a greater advantage in terms of climate, climate change, weather forecasting. Absolutely. If they, so in my opinion, and this can be a great technology for digital inclusivity as well.

Srini
Absolutely. Right? And actually, the one thing we Indians and, and, and, and I’ve, I’ve loved, you know, since I’ve come back here, I’ve always, one thing I keep saying about India is we are least resistant to new technologies, right? We adopt very quickly. And that’s probably going be the biggest advantage, right? In a lot of western countries, there’s a lot of resistance and, and, and, you know, this resistance. Some, some sometimes plays against them. So, at India we have a big advantage.

Sairam
Wonderful. So, I think thus far, Srini, we’ve, we’ve just had some real solid time in terms of talking about future Yeah. Architecture, genesis. Let’s get a little, little closer to what we do for, for winning our bread, right? Yeah, sure. What is it that’s currently brewing in the sign labs as far as adoption of generative AI and chatGPT? Are you looking at couple of use case? I know you work on frontiers of digital assurance test automation, a little bit of digital engineering as well, but are there use cases that you think this is being applied, which is more pragmatic, that could be relatively easier even for a developer and a testing audience to understand? Any thoughts about that?

Srini
Oh, absolutely. I, along with my colleague Rajesh today mm-hmm. we’re working on, what we call the test case generation use case. Okay. Which is, which is actually, we’ve been working for the last three, four weeks. Right? So essentially what we are trying to say here is you can just put in your use, user story or your code or whatever you want, specify the framework, and you’ve got the test cases out automatically. Mm-hmm. So, so that’s something we think is a huge productivity boost for whatever we are trying to do. Right. In fact, our goal in this is to automate as much as 60 to 70% of software engineering lifecycle itself. Got it. But, but that’s going ahead, right? Right. That’s probably a year or two ahead. So, so that’s one big use case that we kind of have today, in Cigniti, that we are all trying to work towards. there are, there are a couple of other things where we are trying to see if, we can use this tool for documentation generation mm-hmm. Okay. So, so that, that’s another thing that’s kind of brewing, which also is there, we’re also using this a lot, for synthetic data generation. Right. Interesting. Because for a lot, because we are an assurance company, right. And at the end of the day, assurance needs a lot of data for you to check what’s going on. Sure. And it’s very difficult to get the data. Sure. A lot of times it’s very expensive too, especially if you’re trying to do something like defect. Yeah. you don’t have too much defect data. Right. So trying to create synthetic data mm-hmm. using these generative principles to get kind of, improve our testing capabilities and Testing.

Sairam
I think goes well, to the fact that Cigniti being an AI led both Assurance and Digital Engineering company. Yeah. So coming towards the end, what sort of impact that could have on concepts like pair programming and developer productivity. Right. We, we obviously test data generation, automated test case generation. Yeah. Yeah. These are classical problems that we’ve been trying to solve. Yeah. Now, can you throw some light in terms of developer productivity and pay programming, things like that?

Srini
Yeah. We’ve been running a small pilot with our, digital teams Okay. About close to two projects. We’ve been running a small pilot here. what we saw and, and what gives me a lot of happiness is we saw close to 40, 45% productivity in the core development activity. Exactly. Using it with, things like GPT, it was recently a paper published by Deep Mind and Microsoft, together, which actually talks about 58% productivity improvement using this. So we’re kind of getting there. So it kind of, when I looked at those numbers, I was happy that whatever, you know, my, our teams kind of created was close to that. Right. So, so, I think, and this is coming, right? I, I, I told, our CEO very recently saying that within the next three months, we will end up writing proposals with two developers and two instances of GPT Excellent. You know, that’s actually coming and we are getting ready to, to get to the world. Yeah.

Sairam
So that’s where I love the word when you said it’s a productivity enhancer for the software lifecycle, not just a developer or a tester.

Srini
Right, Right. Exactly.

Sairam
Okay. Towards the end. So I, I also draw some parallels like how an open source evolve. Like, like today, let’s say imaginatively, I start checking in code, I write some code, I take some code, you know, all that stuff is possible with the, you know, chatGPT. And we have co-pilot with GitHub. What are your views? Like who owns the code in such cases? And, and how do, how do you draw a line between proprietary code, non-preparatory and customers, code? Because we are in the business of helping customers through, you know, IP led delivery, right?

Srini
Yeah. Now that’s a very, very difficult question. And so in the short term, of course, everything remains same, largely because I don’t think the regulatory, approaches have matured to, to, to such a level. But in the medium to longer term, all code is largely going to get commoditized, right? So ownership of code will become meaningless in that sense. Mm-hmm. what you own are, is the data and the requirements driving the data towards that code. So that’s where the IP will be. The actual code itself will largely be commoditized. Right. Because I don’t really see that remaining shorter term. Of course, it’s all gonna remain the same. Yeah. But longer term, even Sam Ottman was saying GPT 5 is all about the data now. Right. See, the architectures are commoditized. Right? We, we all know that data is where the issue is. Okay. Right. And so the data that will be used to generate the code will be where the IP will, will kind of Rest.

Sairam
And my, my view before we end this conversation, there will be vertical specific GPTs. For example, Bloomberg came up with this already. Yeah. The large model for financial GPT, do you agree things?

Srini
Oh yeah, absolutely. It’ll get verticalized, sub verticalized. It’ll, it’ll, it’ll get all over the place. And so, and that’s why I’m saying they, there are foundational models mm-hmm. there are API based businesses, right. And then there are these type of specialized, fine-tuned models and there are businesses downstream. So, it’s gonna be a whole second businesses. Right. And it’s probably gonna be a 100x value the further you go.

Sairam
So in that parlance, could I conclude that it’s gonna open up newer risks of opportunities for IT services companies? Absolutely. As long as you are ready to Innovate.

Srini
Absolutely. See, this is the best time to be in this business, to be very honest. Right. Because, I cannot see of a word. I cannot, like I said, it is no longer an enabler. It is your business now. AI is no longer an enabler. Okay. Okay. So AI is your business. So I really don’t see how you know this. A lot of people tell me, oh, this market is going, going to go down and this and that. I don’t think they’re understanding this. Got it. Okay. Fundamentally, this is going to grow many, many, many fold over the next few years. Okay. And so, got it. It’s a great time to be alive, man.

Sairam
And this Thank you. Thank you, Srini. I think it’s been, quite an insightful morning today. I’m, I’m very sure the audience would’ve found it very insightful. Just to, just to sort of summarize, I think as we heard from Srini, code is getting commoditized, but it’s all about data. And I can only tell you that this is first of the many series of these interesting conversations we’re going to bring you all, made me just kicking in alive. And I’m, I’m very hopeful that this is going to open up multitude of opportunities for all of us. And it’s very interesting to see what’s getting brewed in Cigniti labs and going back to what you just said, it is just bigger and the best is yet to be. Yeah. Excellent. Thanks, Srini. Looking forward for more time, and let’s get going. Thank you very much.

Srini
Thank You very much.