Creating an AI-First Culture with Sanjay Srivastava, Chief Digital Strategist at Genpact

It's about trying new ways of doing things. And that was good. And it served its purpose. And it actually benefited many of us. But now I think the journey is very different. I think the journey this year is about pilot to production. It's about realized outcome. It's about delivering economic value.

And there's a real pressure point. There's a real focus on how do we take, everything that we've learned so far and translate it into projects that will come through and deliver economic value. And I think that's very different this year. And I think the timing's right. I think we're finally at a point where we've tried enough of this.

We made some mistakes, we learned a lot along the way, but I think we're ready and poised for that. I think that's the journey for this year.

Richie Cotton: Excellent, yeah, it does seem like there's a limited amount of time you can just be playing around with new toys before you have to start delivering something. And it seems like there's a lot of different aspects to this, so maybe we'll start with the cultural aspect first. So, what parts of a company's culture are going to affect its ability to harness AI?

Sanjay Sravistava: I think it's amazing that you started with that question because it really is the most important part of this entire journey. We all get caught up in technology and large language models and the next and the latest thing and so on and so forth. But the reality is tech is never the long pole in the tent anymore.

It is right down to culture. It's around people. It's around process. It's around design and experience that actually drives success. And you picked the right one to start with, with this culture, because fundamentally this is about people and this is about getting enterprises behind it. I'll give you our example.

What we do for a living is we take business processes and we drive it for our clients and across the Fortune 500. And in so doing, we'll take our business process, we'll put data and tech and AI and, sort of operationalize it, we'll make it much more efficient, we'll redesign it to the latest benchmarks in the industry, and then we'll run it on behalf of our clients.

We get paid on a transaction basis, they benefit from really, focus on their job number one, and we do the best job we can at this. Now, if you think about our company, we've gone through a process of saying, What is it that we need to do about our culture and our mindset, actually, and the mindset changes what's really critical.

So we've decided we're an AI first company. But what does that really mean? And for us, what it means, and we've spent some time thinking through this is we're saying we used to work in the following fashion. We take a business process, we'd automate as much of it as possible, and then we'd manually drive the rest.

And every single year, we'd actually increase the amount of automation. Jen. up by five points, 10 points, whatever made sense for that year. We felt really good about it. And we're in this continual process of increasing it so we can get the process eventually fully automated. We've decided that our mindset change has to be that we invert it.

That when we take a business process, when we run it, our goal, our absolute norm is 100 percent automated, completely no touch. Now, we don't, we're not there today. We just don't, we're not at that point. So let's say we're 70 percent automated and it still has to be 30 percent manual touch. We call that a defect.

Every time we touch a transaction, it's actually a defect. We label it, we get the metadata around it, and we will continually strive to automate it using machine learning. And so in many ways, it's an inversion of a business model. It was that we would put productivity into it, we'd automate it, and we keep increasing every single year.

The baseline is what it is and the increment is how you improve it. We've thrown that out of the window. We're saying the baseline is 100 percent automated. Anything that doesn't get us to the baseline is a defect, and I'm using software terminology over here. And so just like a software company would, we have scrumless and bugless, and we actually spend time on sort of working down those bugs.

And you might say, hey, Sanjay, you're doing the same thing. I mean, you're really just, putting more and more AI and automation into it. And I'd say, no, it's not. Our fundamental mindset is different. It's the way our culture is designed is very different. And I think that's super important.

And I just put that as an example, but all of the companies that I work with that actually get the culture right, get AI right.

Richie Cotton: That is absolutely fascinating. I think that's perhaps one of the more extreme automation approaches that I've seen, where you're saying every time a human has to touch a process, that's a bug or a defect. That's pretty incredible. And do you have any examples of processes that are sort of amenable to this kind of high level of automation?

Sanjay Sravistava: Well, you know, we spend our day in day out working on business process. We do finance and accounting, regulatory compliance, fraud and transaction integrity. We work on supply chain automation. We, we just, we work across businesses and industries and take core business processes and in any of them, right?

and this is new for us, by the way, this isn't something we had five years ago. This is our evolution of our culture and our mindset to be AI first. And we've now made the pivot and we've turned around. And I'll give you an example. The two things that we hold dearest to our heart. One is this mindset of start with 100 as the baseline.

Don't start with 70 as a baseline and feel good about 70 and then see how you can go to 72 percent automation. Start with 100 as a baseline and feel bad about the fact that 30 is still not there. Like you still have a 30 percent bug list. And then start working it down. And that's a very different mindset and that's really the point I was trying to make.

But a great example of it is then what's the next thing you do? Every touch, do not waste the label. We have this mindset that every touch is an opportunity for a label. And we must label it because once you label it, you actually have the basis to start running machine learning on it and then be able to automate it.

In the old days, we wouldn't think like that. We wouldn't approach it that way. And so that's, like, a concrete example to your question of what has changed. Because if you don't capture the metadata around that touch, you'll never really be able to automate it.

Richie Cotton: That's fascinating. So even the bits that are currently manual, you're then preparing them so they can be automated in the future.

Sanjay Sravistava: One hundred percent. We think that's where the world is going to be, and we'd like to think that we can help lead it to that end point.

Richie Cotton: Okay. And so some of the examples you just mentioned that you were talking about core business processes. Are those the things you'd want to work on first? is there a natural order to which

Sanjay Sravistava: lead to invoice, the entire process of kind of going through a sales cycle and converting it into an actual deal and revenue.

You think about that whole process, right? Typically, if you have to automate it, you take the end to end process, you break it into its chunks, take every little chunk and you say, well, geez, how do I automate it? How do I digitize it? how do I make it faster? You do all this for all of the pieces, then you aggregate it up.

Once you have it. You have an end to end process, it's faster, it's better, it's cheaper, it's probably more scalable, like a lot of goodness in life. But that's not artificial intelligence. One of the things we're learning about AI is that when we are done with AI, at the end of all that, the work that is remaining for our human colleagues to do, is actually entirely different.

And so you actually have to think about change management, you have to think about process, you have to think about the operating design in a very different light. And you must reimagine what it takes to get it done. You can't just apply the automation techniques. I've been spending some time with a colleague of mine, a dear friend in the automobile Parts industry, you know, they make the sorts of things that go into your car.

So for instance, the sort of things that detect, a crash and they'll inflate you know, the airbag system as an example. Now, this is a very precision manufactured component. Do you have to get it exactly right to quality standards? Because not only is it super critical in, you and me driving cars, But if you ever had to do a recall on it, it's a really large economic impact to the corporation, right?

So quality is one of the most important things. And they had this process where, you know, part would come on the assembly line. You take a part and you sort of inspect it visually. You turn it around, you look at it, and you sort of just make sure that it's the right quality. And then you move it to the next step.

And they went about the process saying, how do I automate this? How do I get more efficiency, better quality? And I think a lot of time was spent around, Applying artificial intelligence techniques on computer vision we haven't talked about computer vision yet, but CV is a great way to sort of actually replicate some of the visual identification processes we as humans go through.

You know, the long story short version is that they weren't able to get to the right level of quality, that the human inspection, albeit long and more process incentive, et cetera, was actually a better way of doing it than doing it with computer vision and therefore AI technology. And the way they eventually solved the problem was very interesting.

They said, actually, let's go back and redesign and rethink. And when they started rethinking, they said, let's not think about visual anymore. Let's not try and apply computer vision, because just because that's the way we do it humanly, let's try and apply That goes out of the window. And so how else can we approach it?

And they take a Tom and you hit it with Tom and it creates a sound, right? And the sound has a footprint or a fingerprint to it. You analyze the fingerprint. All right. And it turns out that if you analyze the fingerprint of that sound, when you, hit this, you know, part coming out of the assembly line with something that sound actually is a much higher predictor of the quality and the accuracy of that part.

Then actually, even the human visual inspection, and I just share that with you is a fantastic example that reminds us every single day that you have to have think differently about business processes. They have to think fundamentally, reimagine, reengineer how you're thinking, because if you just take the same lens.

And so let me just take what we did and we can just try and automate it. Yeah, that worked in the yesteryears. No, it doesn't work for us going forward. And I think that's key and critical to achieving success with AI, more and more as we try and apply it to the, to the last mile, to the outside edges of, things that we haven't been able to fully automate today.

Richie Cotton: Yeah, that's an absolutely fascinating example. And I do think it's important to just start thinking about how can you do processes completely differently end to end. I also love that example because it's a natural thing when some of your electronics is broken, you just want to hit it and see if it works or not.

So, yeah, I like that idea of hitting airbags. Okay, so I'm wondering whether different types of businesses need to approach things differently. Is there a difference between small businesses versus large enterprises?

Sanjay Sravistava: Yes and no. I mean, I think, look I'm sort of in a very interesting role, and I really get an opportunity to work across the spectrum. I think in the work that I do with very large corporations, it's mostly Fortune 500 companies. We as a corporation, by the way, serve probably a third of the Fortune 200.

So it's just sort of deep in the very large enterprise space, and I spend time in that business. And I think there, the real challenge is You know, how do you bring in emerging technology like AI in an existing sort of four walls of a corporation with a book of business and sort of wheel that is turning at high speed?

And that's a very different challenge in many ways because a lot of that comes down to how do you end up being an outside in and an inside out person? It's super important to be outside in because you understand emerging tech and you've been in the venture ecosystem. You don't understand, the long arc of how technology is evolving.

Otherwise you get left behind. But the reality is the work we do today is very transformative. It changes businesses, it changes processes, it changes skilling and resource requirements. And therefore to do any of that, you need all of the stakeholders behind you. You need the championship of the entire organization.

It's kind of hard to achieve if you're not an insider, right? So it's a very different challenge because you have to be an outsider in person. You have to be an inside out person and you're bringing an emerging tech into the existing book of business. You fast go to the other extreme of the businesses I work with.

I spent a lot of time at startups in the venture world. And in the data and AI space, the real problem that I think that, they have is the problem of perfect landing, wrong airport. And by that, I mean that they're so smart. They'll build the best possible thing that you can imagine.

And yet. Is it at the right part? Is it the right place? Is it the right fit? Is it the right go to market, right? That's where I think their challenges come in. Because I think there's just no problem in building great technology. It's how you put that into practice. And then I think somewhere in the middle, we have these medium sized companies.

And my biggest observation of medium sized companies and the challenge they have is how do you transform yourself from being a large, small company To a small, large company, and that sounds trite, but if you think about it, right, you're just going through that inflection point, which you don't want to do is become slow.

You don't want to become large in the negative context of, not able to be agile and move fast and modular and so forth. And yet you are a large company now, so you have to plan for scale and size. And so my counterparts in those companies, the ones I advise, they're straddling the fence on I need the enterprise at stack.

I need the scale and architecture. I need to be able to design for growth. And yet, I don't want to lose the agility, the mobility, all of the things that got me here in the first place. So they're very different problems. They're very culturally different for different places at different times. Obviously, technology is one underlying componentry that goes across the spectrum.

So the many ways similar, but I think that's super different on how you apply technology to solve the problem.

Richie Cotton: That's fascinating, and I hadn't really thought that medium companies have their own real class when they're in that awkward teenage phase. Yeah, not quite grown up yet. So, okay, maybe we talk about the large companies in more depth. So you're saying a lot of it is just about overcoming inertia and existing ways of doing things.

Do you have any advice on change management, then, for larger organizations about how you might go get into adopting AI and automation more?

Sanjay Sravistava: I think change management is the critical component that drives success, particularly with AI. You know, One of the things we're learning about artificial intelligence, particularly with generative AI, is that it's actually a little different piece of technology in the way we need to interface with it.

So for instance, over the years we've been working with artificial intelligence and actually before that automation, and all of that is binary end result. In other words, it's black or white. So I'll give you an example. You go to the wall and you turn that switch and the light either goes on or it goes off.

It's binary. I think with generative AI, what we're finding is that that's not what we're working with. We're working with the fact that the answer is that there is a 96 percent probability this customer was charmed. There's a 36 percent probability that this part will need to be fixed when the flight lands in Mumbai.

And so you have this probabilistic answer, which is actually a great insight that we didn't really have before, so very thankful for being able to get all of this back to us. But the reality is you need to think about how you engineer that into your application and how you drive change management and how we work with it really differently.

And so this is the thing we've learned, which is when you apply generative AI, it's no good having an insight. It's no good having a recommendation and just being sort of intellectually looking at it and saying, well, that's, very interesting. To get economic value, you have to engineer that into the line of work.

And the way that gets engineered is very different from that light switch scenario, because it was black and white to this new scenario, because now you need a human in the loop to be able to review it, to augment the decision making, and it's And then do something about it and how you design that is key to success because remember the work that's remaining for our human colleagues to do once you apply AI is fundamentally different from what they used to do.

And so you have to think about the experience, you have to think about the design, and then you drive the change management that gets the adoption. of this new capability. And that, to me, more than this LLM versus that LLM, we spent so much time last year thinking through prompt engineering and thinking through the best LLM for something and customizing it and training it and tuning it.

In reality, what we've learned is, yes, that is important, but really what's critical is actually the change management.

Richie Cotton: That is fascinating. And it seems like, well, from a data scientist's point of view, just basic worrying about probabilities, that's quite a fundamental thing. But for a lot of people in an organization, that's going to be a very new mindset. So does some kind of training have to go alongside that to educate people about probability and about AI?

Sanjay Sravistava: I think there's a lot of training that's required and actually, to be honest, a lot of skilling gaps that need to be filled. I'll give you one story from a colleague of mine in the consumer goods industry. And, you know, every company has their annual events. and as a sort of thing where you bring all of the top leaders in the company and you sort of kick off the new year and you talk about all the things you're going to do differently.

And, and it's a great way to kind of energize people around that common future we're trying to build to. And in their case, they were having this in Asia and Singapore and they had like all of these presidents of all these, and it was a very large company. So. Yeah, a good set of people and CEO and a great set of discussions.

And the CIO took an afternoon and asked for an afternoon and said, I just want to train everyone on generative AI, at least expose them to generative AI. And they ran a very simple process. They said, let's break the whole group into like six teams of four people or five people each. Every team gets, you know, access to, open AI on some Azure kind of projected thing and mid journey to design some stuff.

And now these are senior executives. Their president said, why don't you guys go off and design a new product as a consumer goods company just in the next four hours and present it back. And then your peers will kind of rank rate everything. And you can imagine like, this is an amazing thing, right? Like, In the early days of generative AI, you're getting your fingers dirty, you're getting on to it, and they're kind of brainstorming what scent perfumes do, 20 to 30 year old women in Western Australia really like, and you're trying to design this new product, and, and a picture, and mid journey, and all that, and, and there's all this back slapping, and kind of, thumbing up, and people getting excited, and they come back, and they present it, the best group wins, and everyone gets really energized, because now they've personally experienced the value.

What this can deliver literally in four hours and therefore kind of the whole generative AI and the AI tracking that companies exploded since then. But here's the key point to get that four hour afternoon thing set up. They had to break this into groups and every group they put one person. They said, you know, Sanjay and so many other people.

You're just one other person in your group and he's the chap who's going to help you or she's the woman who's going to help you, take your questions and put it in the system and get things back. And so you can just sort of talk to him or her in English and they'll kind of figure this out and and they'll be part of your team.

And, you know, you and I know that's prompt engineering. That's about saying when you ask this, you actually meant that. And let me rephrase the question in a way that get to the answer and, you know, all the stuff that goes into it. But the point that I want to make is at the end of the session, he said, can you look around the room and say, who are the people in the room that don't wear our company badge?

And indeed, there were like seven of these people that had been put in one into each team and said, that's prompt engineering. And guess what? We don't have those people inside the company. And so that's a great example, right? Like to take on these new things, you've got some skills that are missing.

You have some skills that need to be trained and you have to start getting very thoughtful about your long arc of where you're headed and then make sure that you have the people and the skilling and the skill sets to be able to get you there. Super important.

Richie Cotton: Absolutely, and just well, I don't want to say problem engineering is simple, but it's a fairly fundamental thing if you're going to make use of generative AI. And so those skills do need to be in house in basically every company in the near future, I think. Okay, so, We talked about the enterprise case.

We left off bit on talking about what happens with startups. You were saying that startups often struggle with the go to market aspects. They can build something amazing with AI, but they're not sure how to get that to market. Can you just explain a bit more about what startups need to do?

Sanjay Sravistava: Happy to do that. There's one other thought on the large enterprise that came to mind on the point you were asking about skilling and I have a great colleague of mine gave me this kind of example and stuck my mind ever since. And it's kind of this little story about back 30 years ago.

There's a set of five people that used to meet the second last Friday of every month. And they sit down and they talk about, and this is a sales forecasting meeting, by the way, and they just meet. They're the five regional VPs and it's kind of for the parts of the U. S. And they sit down and they talk about temperature and consumer focus and what the market's looking like and who's the competitors are.

And, and therefore, what the forecast for this air conditioners and heating and cooling of equipments in that case would be. And, you know, two of the people in the room were like really good at like bringing competitive information and so and so's doing this or the weather's changing in Minnesota or whatever it is.

And a couple of the people were like, fantastic at arithmetic, like, the conversation be going on, they just like, do do do do, computer in the head and say, okay, this is what the number is, right? And then they had this one person and she was kind of like, you know, what about this?

But you know, but what about that? But wait a minute, you know, what if this happened? And it's like, let's just get on with it, right? And this group was working together and they're getting sales forecasts and so forth. The reason I said it's 40 years ago is because then came Microsoft Excel.

And so you can like quickly see how this meeting changes, right? Someone comes with an input, you enter it into the spreadsheet before you even can say the next word. The whole thing is re computed, out pops the number. Now think about what happens about scaling and the people in this exercise. The two people that were bringing in all these fantastic inputs from like clients and customers and industry and competition.

It's still very relevant in the discussion, albeit some of that got automated, Like, we can get automatic feeds on temperature, we can get automatic pricing off the internet. Like, a lot of that stuff has now gotten automated. It's still kind of relevant. But those jobs kind of change over time. The two people that were like these fantastic mathematicians, without whom, by the way, this meeting would have been super ineffective because we'd be sitting around crunching these numbers and so forth.

Guess what? All of a sudden, completely irrelevant to the discussion because Excel recalculates it before you can even, like, finish your sentence, right? So think about what that means. By the way, this fifth person, who's kind of like a little bit of a, seen as a drag on the conversation because she has to be asking all these questions and it's like, can you please just, hold off till we get this done?

All of a sudden, she's in the prime role in that business because she's the one who's asking these tough questions. She's doing the what if. But that, but what about this? And that's the whole idea back to your point about prompt engineering is like you have to rethink what are the skills that are Necessary and this is a simple example of sales forecast and we can see this play through in so many different careers and businesses It's how you approach the problem What are the skills that are required and the more I think we think about ai and how the world is going to come to Be It's the people that understand the business, it's the people that are willing to learn, unlearn, and relearn.

It's the people that ask those questions. It's the people that want to interrogate the data. It's the people that want to sort of go, what if? And what if then? That's the skill set. That's the mindset. That's the skilling that we really need. And prompt engineering is one version of that, but there's so many different examples of it.

And I just, every time I sort of think of that, I come reminded of, how all of us in our own careers need to be thinking about learning and actually completely unlearning it and then relearning it again with a different lens and that journey is perpetual and it's just so much so important for all of us.

to answer your question on the startups, I think, look I'm just in an incredibly fortunate place where I spend part of my time with large enterprises, and I'm helping and guiding, CIOs, CTOs, and Chief Digital Data Officers of large corporations really rethink business transformation.

On the back of data technology and now increasingly, and that's incredibly satisfying because, we're getting to actual outcomes that are a scale that they're truly meaningful for the corporations there at, but at the same time, you know, when I work with startups, I'm on the other side of the spectrum.

I'm in venture capital. I do a lot of due diligence and data and I firms that work with startups. CEOs of young startups advising them, mentoring them on their boards. And I think the fundamental challenge that they have is they're able to reimagine technology in the ways it can drive productive value.

But the challenge is obviously around how do you that with real enterprises, with real demand, with where the puck is right now. And capital at times can be super cheap, but really, it's actually expensive. The most expensive part of it is your own time and the opportunity cost. So anyway, you cut it, you only have a few slots and chances at it.

And even if you had all the capital in the world, and so getting that question answered, right, which is what is the right intersection about this amazing thing that I can build. But the real life applicability today, here, now, right? And all of that isn't technology. It's about business processes, it's about change management, it's about data.

A lot of things that people miss is like, you need a tremendous amount of data and a foundation of data and quality data and government data. to be able to do anything with AI, And so you have to think through how all of that comes together and then intersects with what you're building.

And so most of my time spent with large startups, well, with small startups, with small companies or venture backed startups is actually helping them crack that problem. Now, in reality, it's actually, what I love about this is that both of them converge because new emerging tech, Existing enterprises. One is looking for the other.

The other is looking for the other for the first one. And so it's a really interesting ecosystem that is full cycle, but in some ways, slightly different problems that they face.

Richie Cotton: Okay, so it just seems like there's a lot of things that need to come together. So you mentioned like it's not just about having new technology. You've got the data, and you've also got the business processes that need to be worked on. So a lot of things there. And I know in addition to your work at Genpak you're an investor with Masa Group.

And are there any particular companies or areas that you've been looking at through Masa?

Sanjay Sravistava: So, you know, look, Richie, I do three things. I obviously do a lot of work in digital transformation at Genpak. We're a business process transformation company with a now with an AI first strategy, super compelling value proposition. We only take a few things and we do it very well. And so that's been very involving.

I also have a startup background. I built four startups back in the day, sort of dabbled in my own set of technologies, really inspired and energized by that ecosystem. So the venture part of my life today is a little bit about and it's a very small footprint. So think about data companies and artificial intelligence really just the two areas.

We do a little bit of tech, but, beyond that, I sort of, lose. Relevance very quickly. And so there's just significant, if you think about the landscape that's changing our artificial intelligence and how you apply it. And then particularly companies that can work through data and build the right stack and qualify it and engineer it and make it high quality, like there's a lot of opportunity and a lot of gaps to be solved.

And then I spent a third of my time actually running a think tank, which is actually super interesting because we've got almost 90 CIOs, CTOs, CDOs from across the world for large enterprise like Fortune 500. Some fortune 1000 companies and we use that think then to actually call out the collective intelligence amongst us.

And this is pretty. This super important, because, if you think about it, like for someone like me in my CDO role. The way you get things done is, you try things, you incubate other things, you pilot it, you experiment it, some things work, some things don't, you learn through it, you kind of get it right, right?

And that's in some ways, if you think about it, it's kind of expensive, it's inefficient, right? And yet, like, it's emerging text, there's no rule books, there's no opening manuals, sometimes you bring in consultants, but sometimes the gap between what they're saying and the reality of day to day is large.

Sometimes you go to analysts, but it's like way too much data. And so we all struggle with this. And so I started this mechanism where I just call up a colleague and I'd go, Richie, I just ran into something. I haven't seen that before. Have you seen it? What would you do? how would you come at it?

And those one on ones really just helped me in my day to day job evolved into two on ones to three on ones to 12. And I started this thing called the executive technology board. We're now close to 90 members across the globe, mostly North America and Europe, actually. it's across industries, so banking, consumer goods, capital markets, life sciences, high tech, manufacturing, so it's a good set of people.

And we basically come at it and say, listen, in a very confidential setting, in a closed room environment, in a peer group of curated members, and I do all the curation, We'll pick a few topics and we'll basically unpeel it together. And what we're going to do is force the discussion, not unlike what's the future of mankind, what are the sentient threats of AI, but very specifically in your corporation, what are the three things you're doing in this area?

What's breaking? What's working? What have you learned? What can you share? And so it's a true learn and share environment where we're trying to get fast forwarded through this normal cycle of try, experiment, incubate, learn, da, da, da, da, to how can we pull out the collective intelligence and put it into play in a way that we can be more efficient and more productive for our corporations and kind of lead them in the right direction.

And it's just been an amazing forum because what it does is it puts you right on the front street, where the rubber meets the road. Except that you're looking at it across industries, you're looking at it across functions. And it's just been a super engaging, intellectually satisfying activity.

I'm super passionate about that. And so, the cross section of being in very large enterprises with trying to deploy digital transformation that we're accountable for, by the way, at the end of the day. Working with young companies, startup CEOs, and helping them actually think about how to apply emerging tech in the large enterprise space and then bringing to that a think tank and a way of collective intelligence, a way of culling out the wisdom across the group.

So you can actually not. make the traditional live and learn kind of mistakes and fast forward through our journey through the collective wisdom. it's just been great sort of combination of things. And I've enjoyed it and I enjoy it. I love it. And I love what I do.

Richie Cotton: That's wonderful. That social aspect to learning where you just chat with other people who are facing similar problems is incredibly important. Are there any learnings from that group that you can share?

Sanjay Sravistava: I think the group is very much focused on granular, specific things, right? So, for instance, one of the things we tackled on is how do we think about enterprise architecture, these are large corporations, in the new geopolitics.

Now, as recently as even three months ago, things have changed quite a bit, right? You think about the conflicts that we're seeing across the world. You think about the regulatory evolution across different countries. You see the new regs that have come in, for instance, from China and other places. They drive some implications for technology and architectures, right?

And so, IT, for instance, is always built on the foundation of leverage, right? I mean, think about large global corporations, CIO, you build an organization, you build a set of things centrally, and then you run the globe or your operations across the globe through it. Well, that's changed now, right? Because invariably, data has to sit in its sovereign.

More and more, we're starting to air gap and segment things in case things go wrong. We're able to kind of manage the, conflict better. We're now starting to separate licenses and other things, and separate the systems completely for other regulatory compliance and other evolution of rules that are coming through as we all think through our own policy of the corporation and the implication of enterprise architecture.

Now, that's a great example. This is happening as we speak. Thank you. It's never been done before. There are no operating manuals. There are no rule books for this. And the only way you can come to a good answer, well, there are many ways, but I think one efficient way of coming to a good answer is to actually work with a set of peers and really put the pedal to the metal and really start to think through, well, what about this?

Well, what about that? Well, we tried this. Well, we're trying this. Well, here's the exposure. Here's something we thought through. And that collective knowledge, that ability to bounce ideas amongst peers really drives a really great outcome. So that's one example of one of the many topics. We spend time on generative AI.

We spend a lot of time on data cloud platforms. We spend time on organizing for success. We think about talent. it's what's on people's minds. I co create an agenda with the people that are attending a meeting, and then we spend a day really, really getting the depths of it. And so it's just a great great experience, a great learning for me.

Richie Cotton: So, we've talked about how startups and how enterprises can approach AI. What do you think are the biggest opportunities in the next few months?

Sanjay Sravistava: Wow, Richie, that's a great question. I mean, I think you have to say that we're fortunate to be living in the time that we're today. Everyone talks about the pace of change, and indeed, everything we've seen over the last year or so, it makes you believe that, oh my God, the pace of change is amazing.

But here's the truth, the pace of change, as fast as we think it is today, is actually the slowest it'll ever be. And that's the mindset. That's the reality. That's what you have to sort of think about as you kind of go into the next years. And I look across the spectrum, right? I look at things like medicine.

And the work we're doing in precision medicine, it's not just about drug discovery and the latest proteins that can take a new crack at solving old problems, it's actually about how do you transform an entire company, so you're going from mass manufacturing medicines, which is what most companies do in the pharmaceutical sector, To precision manufacturing.

so the thing that's going to go into my arm isn't actually stocked at the CVS down the street. I'm going to send a sample of my DNA. And when they receive that, they will actually mix the cocktails specific to me. It's precision made for me. It gets delivered in 24 hours and it goes into my arm.

That's the world we're going into. And how do you transform a pharmaceutical company that does mass manufacturing, distribution, retailing at the, at the neighborhood pharmacist. And then I kind of show up with my little prescription and get it. To a world that is entirely different. And that's kind of the future where that's going.

I look at sustainability. So many of my colleagues spend so much time thinking about sustainability and energy efficiency or, for instance in some of the companies like water usage or other precious sort of ingredients and really optimization, routing, so many components of that that I think we are going to completely shift the needle on to As we apply more and more data and artificial intelligence to it.

And then the one that I feel the most. sort of closest or that inspires me the most. A little bit of this is I grew up in India and I look across for instance in that country and I think about these farmers in the fields and the villages, right? That perhaps haven't had the opportunity to go to a school like I did and they're out in the farm tilling away and they have to apply for a loan and they have to apply for a government plan and their ability to be able to speak it out in English or in Hindi or in their native tongue.

And for generative AI to pick it from there and fill all the appropriate forms and get this process moving, you think about the long term impact of social inclusion and economic growth. We're talking about a technology that is just about to scale and kick off a whole new era for humanity. And so it's just got to be super inspiring looking ahead.

All of these use cases that, we get to see every day. And I'm just inspired by the work my colleagues do in this world. I'm inspired by the technology we have. I know there's a lot of things to work out, by the way, this is never a dull day, but it's just an absolutely fascinating point in time that we're at.

Richie Cotton: Absolutely, and I think precision medicine is especially exciting, just being able to have the right drugs for the right people. You've got more of a chance of them working. Also, I do like that last point about just not having to fill in forms and have have AI do that for you. That sounds like an amazing thing because nobody likes filling in forms.

That's a universal thing, I think. Alright thank you for your time, Sanjay. That was incredibly informative.

Sanjay Sravistava: It's been my pleasure. Thank you for calling on me.