S2 Ep5: Data to Decisions with AI at Conglomerate Scale ft. Pankaj Rai

The best questions are not asked by mathematicians. They're asked by philosophers. >> How would you explain data analytics to each of them in one sentence? >> Hard question that data science people are uh finding it hard to answer saying >> from finance to [music] fashion, from mining to trucking, transportation, trading. He is Mr. Pank. Thank you so much sir for joining in today. What do you think India's biggest win is going to be >> in the world of abundance? The decision-m model has to change

>> in one of the stores of a large chain and these guys put a black tape on the cameras more information consumption mental illness. I think the value that the startups bring to the table is far higher than that. >> A lot of people who know you take a lot of gap unless you add commerce and humanities on top of that which your parents sometimes told you is not worthy. >> A lot of people are having a top- down push for deploying AI. Ultimate intent is not calculator first or AI first, business first.

>> What is that one big problem that you're trying to solve for right now that that you're completely obsessed with at this point in time? >> Are two industries that are very dear to us? Our health and our money. We don't want to part with any of it. >> Are there any key attributes because of which bones have become faster in terms of dispersements. [music] This show is brought to [music] you by KRUDs, your partner in digital transformation. From cloud innovation to intelligent automation, we redefine

enterprise growth, empowering organizations to shape the future with technology that thinks ahead. Hello everyone, my name is Ah Tandan and welcome to yet another episode of the innovators and disruptors podcast. Today's conversation is very interesting for me because we're going to deep dive into the world of data analytics and specifically talk to you about how a lot of business decisions from banking to tech to the world of a Indian conglomeate is being taken using data analytics and how that becomes such a pertinent part of our lives today and

for that we have a amazing amazing person joining us today our guest who I also refer to as my mentor as well he is Mr. Pankar thank you so much sir for joining in today >> thanks a and it's wonderful to use this holiday in this way for me [laughter] >> no I mean and and for the for the audiences let me also introduce you panka sir you know has in his journey been an engineer from IIT uh IT Delhi then from I am Ahmedabad you did did your MBA and then you have been you know changing the landscapes across banking and tech by deploying a lot of

technology you're teaching technology to lot of nontech managers of people in the world of business as well right and uh across companies like ICICI uh Dell Gap to uh then Wells Fargo where where I got had the opportunity to meet you as well for the first time about 8 years ago and now with Arcter group as the chief data analytics officer for the whole group so that's such a fantastic journey and today sir I wanted to talk to you a lot more specifically about art the builder group of course I want to talk to you about your leadership aspects. How do you look at leadership? How do you look

at technology today? But uh also dive deeper into how business outcomes are a function of how data and AI levers are being tweaked on a day-to-day basis in both business as well as consumer brands of Arthur Builder Group. It's a very large conglomerate. I think if I'm not wrong, it's a hundred billion dollar market cap plus conglomerate right from India and it ranges across a lot of different diverse business vertical from finance to fashion from mining to trucking, transportation, trading, uh paints, uh cement and I don't know what not, jewelry as well, right? There's so

many more I'm I'm sure. Uh so so I wanted to touch upon some of these aspects if that's good with you. >> Sure. So I'm a I'm a big fan of Simon Chenek and I'm sure many people follow him and he has this wonderful video called the golden circle where he draws three circle and there's a why in the middle and there's a how in the outside and then what outs totally outside >> and usually we see the what of everything and everyone and we try to make conclusion based on the what but the real meat is in the why and the how. So I will use that analogy to say why

and how of my own approach because a lot of whats that you spoke about I have often reflected and I have been asked why do you do that. So my reflection is that you know while I studied engineering and MBA uh my first job was in management consulting >> where [clears throat] the muscle that was developed was to say what should businesses be and what should businesses do. So this was back in ' 93 when India was liberalizing and a lot of consulting assignments were to say what business should we invest in and a lot of foreign companies were coming and Indian

companies were also trying to diversify. So my first muscle development was which business should be uh you know entered into what is the ROI what is the opportunity and all of that right so first was the consulting muscle of what should the business be >> right >> next I got into IC and G capital which was about again lending money to businesses to say what kind of returns can you generate which businesses will do well which won't do well so again a a commercial angle so the first angle was the strategy angle in the consulting

world and then there was the commercial angle how do you make money and why do you lose money because there were some companies which went sick and you know in IC and all we would also to look at those things. So I I learned that and then I got into this world of data analytics and offshoring and so on through Dell and others. So I think uh my primary metric to look at anything has been strategy >> and then finance and then technology. So I think uh for me whenever someone talks about saying what will AI do or what should happen here and all my

natural inclination is what business are we in >> how does it make money and now let's figure out what contribution can technology make. So I think that's just a reflection to keep in mind. So when people talk about AIA strategy or you know what should data do I said uh you know let's first talk about strategy and finance and commercial and then we can figure out how to you know use this data and analytics to make those things meaningful. So that's just a just a context in which I think about things. So, so talking about Alibaba Group, I've

been here close to 4 years. >> Uh, first time working in a Indian conglomerate, great brand, great reputation, diversified businesses. I I tell people that this is a microcosm of the Indian economy. In fact, one of my welcome speech to some of our interns who just joined last month was to say that, you know, by the way, our current chairman uh took over about just about 30 years ago. uh you know when he took over the the company and the group I think it was about $200 million market cap and now like you said about $110 billion so in these 30 years the group

has probably averaged about 17 18% of CAGR market grab growth right so I tell people that you know you guys are joining here now >> think about the next 30 years if we were to apply the same 18% CAGR to this 110 billion what will it become >> it become over two trillion >> right >> and I tell people that this 2 trillion is actually not difficult to imagine today we have two trillion market cap companies a few of them out there in the US. So there's no reason when India becomes a 30 trillion economy as it

should there can't be a $2 trillion group. So I just tell people and that's when nothing changes the same CG >> we should actually try to do more than that. So again those are the context in which I talk to people to say this is the business ambition that we ought to have. Now in order to pursue that ambition, what can data, AI, all of these things do is how I would frame this problem to say what do you want to become and then how does these technologies allow you to become that either by making your current businesses different or launching new businesses

which use the power of these technologies and the group is trying to do all of that. So that's just a broad context. Uh just the other flavor on the group is that a lot of people and me included when I joined we had heard a lot about these various brands the consumer brands Louis Philip on the fashion side and you know Masaba and many others and also on the capital side but contrary to popular perception the >> market cap contribution >> of these consumer business is not as large. So like I said the 110 billion market cap our you know fashion retail

business probably contribute three uh the the financing businesses to listed companies probably contribute about 8 to 10 or something like that right so in the larger scheme of things uh the big daddy is really Artic CN close to you know 354 billion dollar market cap >> the next one big is Hindal coalinium which also owns a company called novelis in the US right >> third is grassim some of us growing up have heard of grassim as a as a conglomerate in itself and then of course there are you know phone idea and then BA carbon and some unlisted

companies trading business and so on so it's a very and of course new businesses in paints and jewelry have been gone a couple of new digital businesses like Bila pivot which is a B2B e-commerce platform tomorrow which is a bunch of D2C brands which have been consolidated and many new businesses in the offing green energy is one area being looked at very seriously and so on and so forth so therefore in some ways The Ayataba group represent is the microcosm of the Indian economy and therefore anyone working over here >> can get the exposure to the full breadth

of what our economy has to offer and therefore it's an exciting place to be. I'm sure I mean I I just think about the fact that you know you have access to all data points either which are existing today or the data points that you would want to extract out of it to you know build a lot of decision frameworks on and that for a lot of people would be like a candy land of course not for you you've been living that life and there's a lot of responsibility is on you on you to deliver those outcomes but for someone sitting outside you having a overview

and like you said of a microcosm of the Indian ecosystem is so exciting ing and so so relevant today right to to get into deeper insights and that's what I wanted to also tap into today two amazing points that you shared punkage was one was that you know today a lot of people are having a top- down push for deploying AI and you said that's a very counterintuitive mechanism or method of looking at solving a problem you spoke about strategy coming in first to figure out what's the business that we are in second what is the money-m opportunities out there what is the way we currently

monetizing what the different ways in monetization and saving money and so on and so forth as well the top line bottom line impacts and then you talk about how technology can come in that's where AI and data play a role to either create topline impact or bottom line impact or NPS and so on and so forth right so that's a fantastic learning for a lot of youngsters who are trying to build something today right because uh you don't have to be technology first in everything that you do you have to think about what's the problem why does it need to be solved and then think about

what's the technology that comes in uh >> I think just a flavor over there when people say I am AI first or I'm data first all of that I think the way to think about it is that you know whenever you are thinking about that business do consider the relevance of that capability because many a times you know we are very used to using whatever we were doing before right >> I mean imagine when the calculator came in and people wouldn't use a calculator because they felt they they're better without it right so you have to tell them maybe calculator first so that's

fine but the ultimate intent is not calculator first or AI first business first makes sense >> so that is sometimes not said >> but that that has to be imagined that eventually you have to improve your business. You're not doing it for the sake of doing it because many a time people do end up utilizing things. I mean if you own a car I mean you need to know why you want to use it otherwise it's of no use. So any technology the user has to know why they bought it and what they want to do with it and is it worthwhile. So many a time when a lot of

companies say xxx first >> people just start using it and it's not bad to not use it but then eventually the value will come in when you merge it with your business outcome >> and that requires a slightly different muscle to do that many a time people are hiring a lot of tech people so I was in six sigma people hired a bunch of six sigma people and said go do some improved processes now some of the people had no context of the process how could they improve process they just knew tech sigma so just having a technology mindset is not good enough

that technology has to change business. So that's why I said the strategy and commerce and then technology and in fact we have a framework internally we call it the ways of thinking ways of working and ways of doing. So I say the doing part is the tech part >> ways of thinking is the strategy part and ways of working is your commercial acument. So unless you merge all three uh think of yourself as a full stack human being vertically downwards versus horizontally. I think that's the impression that we are trying to create among our analysts to say while you

learned engineering and science that's just the base layer necessary but not sufficient unless you add commerce and humanities on top of that which your parents sometimes told you is not worthy of being done that's where the value is so today the value is actually in the uh the the the humanities and the commerce side because tech is sort of going to be democratized and much more easily available in fact after the LLMs came in and by the way my daughters are also economics graduate they are not data science people and bunch of kids that I know was scared of maths and they would

in the past feel that if this data science is a is a great field but we we don't like math so we can't succeed. So these days my my message to them is that now these LLMs have democratized this thing to say you have to be a good prompt engineer and what is the prompt prompt is really asking question the best questions are not asked by mathematician they're asked by philosophers so if you are a economics person or or a liberal arts person maybe you can ask better question than the engineer so don't think that you are any less than the other guy as far as usage

is concerned now of course creating those technology there'll be some scientists who do it and they obviously need to know all of that but as a user I would argue that maybe the the humanities people can ask better questions. The engineers can may not be as good by having >> fantastic point. >> So those are the ways I'm kind of challenging people to think about using these technologies and asking better questions because a lot of non-engineers have been scared of using data science and I'm telling them that this LLM is

actually giving you an advantage because you know English and this is linguistics >> how to think about the world, how to imagine the world because nowadays people are saying you know you have to contextualize the world. Who's better to contize something than than an artist? >> No fair. That that's a very very interesting point. In fact, on the same note, I wanted to ask you uh since EBG is such a diverse group and you must be talking to a lot of very different sets of personas right across the group. If you had to talk and explain data analytics to let's say a plant head on

one side, a wealth adviser on the other side and a merchandiser on the third side, how would you explain data analytics to each of them in one sentence? See everyone understand their business and their function well. >> So you have to just talk to them in their own language. So I use three words to explain data analysis. I said whatever you're doing you can be a marketing person whatever >> this data will either make it do make you do these things faster because sometime you know you can you can automate things and all that better

because you can bring all different uh sources of data and you know more richer analysis so analysis can be better and then newer the things that were not possible and you can do it. So I would say these three words summarize the outcome that anyone can get in any business that they are in. Faster, better new world. What else do you need? Everyone needs faster, better world. So I think that's how I would tell anyone that whoever you are, wherever you are, you are at the bottom of the pyramid, top of the pyramid, you are the top of the pyramid. You are consuming a lot of

reports. Maybe they're giving you >> I mean you can get it and by the way and there are lots of examples of these examples. So there there's a specific use case. There lots of reports were coming and people are figuring out what to do. Now all of those are coming at 9:00 a.m. all automated. They're also highlighting the issues that are there, the anomalies. So people can start acting on it from 9:00 a.m. versus from 9 to 2 they were figuring out what to do and start acting at two. So you are faster, you are better and some newer ideas are also coming up because some of

the insights are new. So I'm just saying and you can apply this theme anything to anyone. So I think whatever business or function you are owning >> you want it to become faster better newer with better decisions and that's what I think one one has to think about it and then go back and seek the right data and the seek right technology to get there >> true I mean I remember in 2017 uh I started running some experiments when I was working at target that's where we first met as well so we uh use NLP and NLG to one look through a lot of data

points identify what are the data points that meant were more meaningful for us extract information from that use NLG to then create reports and with L&Ms today that work is so much more sophisticated because now as you said right there's a lot more richness of data from varied other sources that can come together and give you a lot more specific information through generative AI as well right now charts of different form factors of manufacturing chart would be very very different as against a merchandising chart and it's like you said available on their tabs or uh the laptops at 9:00

a.m. in the morning for them to immediately come out on it or work on it. Uh in fact uh that brings me to my next question. What is that one problem because you work with so many again diverse groups. What is that one big problem that you're trying to solve for right now that's that you're completely obsessed with at this point in time >> you know. So I think uh the the most uh interesting problems which are useful is very are usually very boring in nature. So data quality >> is a very boring topic. No one would say get excited by data quality but that's a

huge problem. We have so many plants, all kinds of machines, all kinds of data points. So we are now trying to create a unified data quality engine and it's not you know difficult to comprehend you know think about we have sensors. So making sure all the sensors are in place >> they are giving the data the data is in the range that it should be if it's temperature it should be in this change pressure this range it is coming at the right frequency and therefore the model that runs is the correct model because if these things break down >> you would know what went wrong and that

can be applied to your healthcare model or whatever model. So I think this whole data quality piece is very very useful especially in the realtime situation but not often people spend too much time about it. So we are focusing a lot this year to say and we have 200 factories how can we create a data quality engine where all the factories can have data available of the right quality and you know then the models can run. So that's one uh second is we we call it the asset taxonomy framework. Now a a motor or a compressor is there everywhere. Now in every place those things have not been

labeled properly. So a model if you make a digital print of a compressor in one factory those those points are called something. We are now creating again abstracting it out and creating framework to say if you create a model for a compressor here it can be transported there. So many of these things are behind the hood not very well known but extremely powerful. So we are trying to focus a lot on the fundamentals because we get excited by the outcome. Oh we created a great twin but how do you translate that twin to 20 other places? That's hard. So some of

those foundational layers is where we are you know getting excited about. >> No that's fantastic. In fact, more often than not, asset libraries or common terminologies become so important uh whether it's a CDP implementation or anything else later on for a large group uh with so many different types of businesses that it becomes very important that you be all talking the same language across the group, right? And and that also means or brings me to my next point which is there must be use cases where you must be able to leverage one of the models that you've created to

solve for a particular problem in one business and that can be applied or ported to other businesses as well. Could you share some examples of that and also maybe an example of something that you tried to make it work but for some reasons it did not. >> Yeah. So I think uh model models around digital twins they are the easy ones because you know if you make a digital twin of a of a equipment in one place >> wherever else that equipment is there. So those are the easy ones which we are trying to trans which is the reason this this need for uh you know asset taxonomy

came about because while the equipment is the same it's not labeled in the right way. So once we create the right labels the transportation is likely to happen. So that I think will go smoothly because those equipments behave in the same way. In fact on on that one we are now starting to engage with an interesting startup. What they're saying is that the traditional digital twin was just using the temperature pressure the just the the the the mathematical numbers. What they have done using LLM then we have to figure out whether it works or not. They have utilized the

physics and chemistry of the process to make the twin even more comprehensive. So they saying we have taken the textbook based on which these things were made the physics of that material the chemistry of the process and they have made a more comprehensive twin. So we we'll do that. So we also try to engage with new startups with new ideas to say can we make it even better. On the other side where things haven't moved as well which we had expected to is the area of logistics. We felt oh we are a great manufacturing company. We have so many trucks running across

businesses across the board and if you do some logistics optimization in one business it'll work elsewhere. I think what we failed to acknowledge was that sometime the trucks are of different sizes uh different capacity uh different uh products go in different sizes of trucks. Then there are you know local nuances and you know sometimes you know mafias and all in different location. So there a lot of other factors which are not allowing the so while the technically the model is the same but the other factors we fail to recognize which is what I sometimes say that the

whatever we do in the real world you have to look at it holistically it's not just the mathematics of it there's a whole lot of other variables which are hidden which we call change management related to other methods. So that those are the things that we are learning and we'll find a way to see how to transport that >> makes a lot of sense. Thank Thank you so much. That's a very very interesting example. In fact, in 2018, I learned a hard lesson uh myself was running an innovation project for anab bush web and we wanted to use a combination of a few

sensors and a couple of cameras to capture a lot of refrigerator related or vizzy cooler cooler related uh analytics, right? Be it temperature, uh be it luminosity, be it even the movement of a asset, right? So people were stealing assets. So that had to be replenished. So we capturing a lot of these analytics and we tried it in you know our demo room environments and we thought that you know it makes a lot of sense. It's capturing a lot of information that's going to be very relevant for the business. Cut to deploying that in 10 different point of

sales environments, wine shops etc in Bangalore itself and we failed at it badly because we realized that you know the real world scenarios were so different. One wine shop owner in the outskirts of Bangalore was very disappointed when we tried to do this at 12 12:00 p.m. afternoon and we thought that these are these are not peak cars right so this would make sense and he threw us out my team out saying that you know you know if you want to take your cooler you take the cooler with you I'll bring another one from kingisher but this is a time where the it's a peak for

me because in the outskirts more people shop during the afternoons for beers hypothetically which we we are sitting in in you know the central part of Bangalore we did not know this supposedly it was our fault in another case we deployed this in one of the stores of a large chain and these guys put a black tape on the cameras and at our end our systems were showing that the sensors are working the cameras are working they're operational but the images that we were getting were black images >> and we had to actually go there to

deploy and change make changes right so >> the learning was very simple and straight for me at that point in time as well that you know sometimes sitting out of corporate offices we are far removed from the realities of uh the world out there right like you me mentioned there are so many cultural nuances that come in at each of these mining sites as well where trucking solutions could be very different from one location to another. This is so so critical today for for people to learn from. >> So the generalized idea that you refer

to is something that you know I've been telling. So we we also have a centralized team to drive training and learning and all of that. And recently one of the senior leaders was talking to me to say how do I create a digital transformation or AI mindset among leaders right >> so so they are the thinking is that the leaders don't get it quote unquote and we need to teach them >> take your leader of that wine shop right >> someone say he doesn't get it look he's he's threw us out right and some cases people do don't get it but many of times

I think >> we don't get it >> we need to understand their business context to say how do they approach their business and within that how can we make a difference. So I think uh the technology people uh need to take time to understand how the real world works and within that how this technology can be uh you know subsumed. I think we tend to believe that our technology is is the superset. >> Everyone else has to come and comply with it, >> right?

>> I think we all know that that's not how it works. They are interested in making their business better >> and they are not necessarily interested in all the nuances of technology and we tend to talk about an LG this that and the other. I mean I mean imagine you have a car and a driver and you tell him you have to go from X2 and he say oh I'm going to do this and now the carburetor will fire up and I don't care what's going to happen. just take me there right and tell me what I need to do and if I don't need to do anything that's fine so I think somewhere some of the

folks who are in any field >> they tend to imagine the world in the way they have perceived it in their discipline they need to enhance their perspective to include the real world of business in which these technologies can make a difference of course we need to teach the others also about these things but I think a lot of teaching has to happen to the specialist for them to understand the taxonomy of the world in which their technologies can make a difference >> right contextualiz ization and awareness become so critical in these aspects.

>> Uh so I'm going to get into some of those uh business groups within ABG as well. Talk to you about that. And for audienc's uh perspective, let me just call it out. Uh one of the bigger groups uh is around mining as well. Right. So Hindalco runs captive boxite mines. Ultra Tech which you mentioned as from a cement standpoint has about $40 billion market cap or more operates limestone mines that feed India's largest cement network right from that perspective if I were to ask you about a specific question uh around some of your data decision making as well data points we

know we we have noticed that monsoons and again heat waves have also brought in a lot of uh you know disruptive holidays as well right so in these scenarios what are some important important metrics that you look at would uh you know would there be certain data points that will help you plan your trucking your trucking systems your loaders and how do they sync with each other better what kind of KPIs do we look at >> yeah so mining is a great one by the way what you didn't mention but we have we actually have a separate company on

mining called SL mining limited >> and they actually do some contract work for our own a few tech mines also and they may end up doing more because they were originally set up as a helping the coal mining as coal is the biggest mining [clears throat] in India but now they're diversifying to do other things. So with them the SL mining company now we are working on a project of a specific mine where we what we calling it a connected mine >> where you know everything will be connected. So I think the way and and I I have to say I haven't uh you know been

an expert in mining but a lot of mining work right now in India is very very localized uh not very high-tech because it is seen as a sort of a not something that you you put in a lot of high-end equipment but that is changing a lot. So we are specifically taking one specific mine where a lot of new high-end expensive equipment is being ordered and we want to start kind of doing connected mine from day one. So while the mine is still being worked out, our team is already creating mock-up dashboards to say, you know, the way the mine is operational as a CEO, what would you

like to see on day one? >> As a, you know, the operator of a certain machine, what would you like to see on day one? So I think we are already working on quote unquote the change management to say some of you have done these things in a certain way. Is there a new or better way? And you tell us, we'll work together. So instead of you know having something being done and then trying to change it we are hoping that from day one that mind will probably work with a totally different mindset where everything is connected to each other and we are trying to optimize

at a more more global level. So that's actually a very exciting project that our team is very excited about because this is not something we have done before. So and we hope thath we can utilize some of these knowledge to many of the other mind that you spoke about what we trying to show. So so one thing that we do is pick up one area do everything very well take those learnings abstract it the abstraction can happen in the form of code which can be translated it can be in terms of some best practices some management rituals and maybe some reports. So in this case

we think that some management reports can be created which can aid better decision making for various practitioners within the mind. So because what is happening is so much new things are available now many of the people over there they have to also change that oh if this is available then I could do this those things have to happen in collaboration with them because everyone is used to running a thing in a certain way so I think we are having those conversations to say when you go into this mine maybe you can run it differently so I think that has to

happen and that probably has to happen everywhere >> wherever now new type of information is available people can do things differently I'll give you another example Example there's a there are lots of market intelligence teams in many companies and we have many of them. >> So they what they do they will bring lots of reports some are paid reports some are open report there are three four analysts they'll sit down create a nice deck for the boss and say you know what boss this is happening this month this can happen next month all of that

>> very manual intensive >> in one place what we have now obviously done is automated all of it. So not only does it generate that report probably much better than what those four five analysts could do, it is also able to say what should you do next one based on that >> which is using again external data, internal data all that. Now the role of some of the people who are creating reports is going to become that now you can act on these things. So let's say those reports were saying you can do this in procurement or this in marketing

or this in supply chain. these people who are data people can now become supply chain and marketing and other people to say I can now use this insight and do a higher order job to drive action versus create insights. So I think we are also looking at some of these problems as a human plus process problem because many a time the other question being asked is oh if you automate everything what will these guys do right >> oh they will become higher order people they'll act on those things because in the past those insights were going to

the CEO who may or may not have the time to instruct his people to do whatever but these people can become his extended arms to act on those things and by the way based on those actions they can train this model they can say oh by the way so you can actually create a better model using the real insights and keep training it. So, so I think the nature of work and the job will change and people will get upgraded also. Not everyone but many of them can. >> It makes a lot of sense. In fact, that was a lovely point where you also added the fact that you actually close out the

loop with these guys will then based on the actions that they're taking will also feed back into the engine the the LLM >> that is the proprietary knowledge of that business because I think the focus that we are having is what proprietary knowledge that you are creating and what new knowledge are you creating. Frankly, this whole world of A and all of that should translate into better new knowledge being created at every node and I think it's our duty to figure out for us in that business what is that new knowledge which is meaningful for us and

how can we keep accumulating it and make it better and that's the endeavor that we are all trying to go after >> for sure. So thank you so much. Now let's move let's move from mines to aluminium right so again hendalco and nollis you mentioned in US as well so for uh the perspective of the audience let me just highlight that novelis is the world leader in the aluminium beverage can sheet and the largest recycler of used beverage cans pushing higher recycle content which focuses of course on circularity as well a lot more in fact the leader sustainability

targets include 75% average recycle content and uh keeping all these things in mind You know we understand that recycle content is related to the big energy events right and you did at the start of the conversation also talk about energy focus sustainability focus what are some of the key metrics that you look at from this land standpoint >> so I think if you look at um aluminium you know in India hindalo is our flagship company novel like you said is there two very different businesses two very different context

>> so Halo has a lot of upstream you know where you have boxite mine then you take alina and all of that and you make you know products which are you know it can be coils it can be rolled whatever right >> there it's the end use which is the can >> right >> that goes to a final end user here it is sort of an intermediate which can go to this now of course we are also trying to diversify downstream in hindalo also to maybe get into making things for the EV battery >> maybe Apple or whatever wherever downstream so downstream is an important

focus area even for our domestic business but the global business is largely over there because the market is very mature. Recycling is very mature. >> If you think about recycling in India, it's still not a very mature market because we are so >> not as good in recycling. Our our market is not mature. So, we are still a bit primary aluminium focused. >> Those countries are far more recycling focused. >> Now, ESG as a as a goal that you mentioned is a is a is by the way for

all public listed companies. There are BRS reporting as they are called they have to all publish it. It's like a it's a mandate in a company. In fact, our team is involved with our group sustainability team to ensure that these things are reported in a right way and all of that. But I think compliance is one thing to report it. I think the bigger value is saying what insights can you gather and what can you drive in the business to make the business better. I think those are the areas where you know you mentioned energy while energy is the highest usage in in a place like

aluminium. So there you know we have worked on a project to say that you know there are large baths in which you know the the aluminium is formed aluminina is formed. So over there it's a 36-hour cycle in which the stuff is getting melted. Now in that molten form sometime the temperature can fluctuate and it's a machine. >> So the model that has been created is to look at all the sensor data look at the pattern to say will the temperature fluctuate and if the temperature fluctuates energy consumption will go haywire and your quality may also go

haywire. So now there's a model which is able to predict a temperature fluctuation going. The antidote to that is there's a there's a catalyst called almonium fluoride rods. So you can dip the rod and you know control the temperature. So net net these types of things are being done to say can you project something if something's going to go wrong can you proactively take a measure so that the energy consumption can also happen the product quality can also be controlled. So in general a lot of energy related projects are anyway happening for cost reasons and by the

way if you even if it is for cost reason if you're conserving energy you're also you know sustainable in some ways so [clears throat] many a times we tell people that don't think of sustainability as a as a as a extra thing that you have to do and over and over everything sometimes you do better logistics >> you will save cost but you'll also save you know become more you know sustainable so I do think that a lot of work that we do whether it is in transportation whether it is in energy conservation People who are doing it are

not thinking that they are contributing to sustainability but they hugely contributing to sustainability and we make them see that that what you do is actually showing up in the ESG scorecard as well. So don't think that ESG is some after thought that you are doing and we are trying to bring the two together. Safety is another one which we are focusing a lot on and over there actually we doing some interesting work on the behavioral science to say it's not only about data how do humans think and how can we you know make the factories more safe.

>> It makes a lot of sense. In fact, uh they say in manufacturing reliability beats hero days or the top production days, right? So is there any way that you've been able to because I'm sure that you know when you're talking about manufacturing of aluminium uh there are a lot of other secondary and third degree vendors or partnerships as part of the whole ecosystem where uh where there there are a lot of dependencies for example when are the trucks or raw materials coming in so on and so forth. Have you been able to deploy any algo which has thereby reduced downtime of

manufacturing because that I think in most industries becomes a very big blockage in production. >> Yeah. So I think uh predictive maintenance or asset reliability is a is a classical manufacturing use case to make sure that your asset doesn't fail and then the metric that is used is called meanantime between failure. So >> between two failures if you keep increasing you are better off. So that's a standard metric that manufacturing companies have always used. So there are obviously there are algos like I said that they are able to so for example

there's a sodash making process >> there lots of these membranes >> if a membrane fails then your process stops. So are you able to predict which membrane will fail >> and you'll replace it at the next normal shutdown. So you know all of these things have shut down. >> So if that shutdown period you already know some of these things are going to fail in the next before the next shutdown. You can replace it now versus wait for that and failure to happen. So many such things and there are outside companies a company called uptime AI and

many others. So we also engage a lot with outside startups who have come up new algorithms try to develop our own algorithms also but asset reliability is a classical you know case which we continue to focus on and make sure that the machines work for as long as possible and whatever has to fail can be predicted in advance and done at the annual shutdown phase. I'm assuming since you've come in and these these alos that have come in more mainstream uh you said that this has been there for a very long period of time but have you been able to create a delta in this

space as well you know that's always a hard question that data science people are uh finding it hard to answer saying what did you do and by the way that question was there 20 years ago when I was a six sigma head in the G capital days because many a times what an enabling function has uniquely contributed to is sometimes times hard to quantify. So I think we look at it in a collaborative way that we are working with the business teams >> and collaboratively say this is the past data you had this this thing >> after we have worked together what is

the future and is there a delta change. Uh so in the past many a times we would do one or two things only >> now we have learned to say in that area try to do all the things together >> so that you can make a bigger impact versus you just did one thing some three other people did something else no one knows who did what. So we are saying let's work together pick up a theme let's say predictive maintenance if there are five things to be done on predictive maintenance let's do all five >> now which of those five contributed we don't know but we all know those five

together contributed something so those are the new approaches that we are doing which makes it easier to you know quantify the impact that people are making >> that's fantastic >> uh and because people work with so many vendors so many other things all of that so that's the reason for doing that >> got sir so now from aluminium let's move to trading so art builder global trading right so again from the consumers perspective. ABG's trading arm ships uh 27 plus million metric tons annually across 80 plus sourcing countries to 95

plus markets with about $6 billion of revenue. It's headquartered in Singapore. Uh I wanted to talk to you a little bit about this. Uh what are the key metrics that are daily basis that you track when it comes to trading? What what is the most important metric that anyone who wants to get into or understand the space should be thinking about on a daily basis? Right? So first of all, it's a privately held company. So the numbers are not publicly available. But if you think about a trader of anything, right? Uh their job is to you know predict what's so so they

have to basically read the market where is the market headed and find opportunities to do arbitrage. >> In the past it was all about information asymmetry. Even in the stock market someone got some information someone else. Nowadays that information asymmetry is gone with the advent or everything. So I think having a better sense of the market and you can call it reading the market is where the value is now how do you read the market I think data can help you because if you have a lot of data and a lot of insights you can bring it all together to have a

better read of the market than someone else because now you can't trade on proprietary information which was the case in the past. So I think therefore the the need to use data to better develop your acumen is becoming [clears throat] more and more prevalent in the world of trading. So in our case, we are not like day traders. Our our traders are taking positions on the commodities and you know the volume of trading is not as if you are morning evening sitting but they're not taking long-term positions. So I think the thing that they would like to predict is

>> where is this moving? Is it going to go up or down? Many of times it's not the absolute number that they're interested in. They're just looking at the direction. >> So so so our teams are trying to create mechanisms and models using all kinds of data. These days data is available trying to figure out which are the more relevant data and over the last one and a half years we have been working with traders to again fine-tune the model. So we'll tell them we think this is going to happen then something happened feedback why did a delta happen so we'll

never be 100% right but can we be sort of give the confidence to the trader that when you look at this model most of the time uh it is better than your gut feel or if you're a new trader we you have just come into this business this is the past data you can use this as a companion so in many ways I'm saying that all these AI models and all should be seen as a friend and a companion and as a co-pilot not as a competitor Because many a time the starting point is what will this fellow tell me that I already don't know and this hard for any business person to believe that someone

can be a companion. So I think the the selling point is to say think of it as a companion and that's the main thing that we are trying to work on. >> Got it sir. Now I'm I'm going to move to a interesting thing that we spoke about you mentioned about right at the beginning which is a lot of new categories shaping up a lot of extra push into newer categories including paints below opus and then uh jewelry as well with India uh on on the Indria front I read and for consumers to know uh art builder group is infusing 5,000 crores I think someone mentioned that

they have seen a store in Bangalore that's already started the 100 stores likely to come up very soon and the plan is in the next 5 years uh India should become uh the top three jewelry brands in the country right that's an aggressive approach that EVG is taking on that front I wanted to talk to you about two things how are you making sure or on data points that consumers actually are able to take faster decisions and their cart size or average revenue per user are also increasing is it uh you know better appointments uh I do not know if jewelry stores have

appointments But I'm just guessing it out from retail standpoint. Are they appointments? Are they better trying out features? Are they able to use some ARVR methodologies or I mean uh technologies to to try out things better? What works out in this environment? >> Yeah. So first of all, whenever the group has entered a business and that's my personal observation is that they think of it as a generational business, not in a short-term way. If you [clears throat] look at any of these businesses that we have been speaking about,

>> they've been around for like 20, 30, 50, 75 years, right? >> People are not there to say, "Hey, we'll enter it and in some we'll make some money and sell it out." That's not the mindset which is there. So, one thing therefore one has to remember that while there'll be short-term targets of getting to some revenue number or margin number. I think the longerterm intent is to be a trusted partner for the consumer for the years to come, for generations to come. So that's the uh you know strategic thinking that one has to keep in mind for any business that one enters

and there could be some cases where we've exited business also we tried a few and didn't work out but in general the idea is that you want to create a trusted brand in fact uh uh uh you know inspire trust is is is a is a purpose statement value in our group this thing that you know create businesses which are sustainable and so on so forth right so that's the thing and the brand does inspire trust now specific to jewelry I think the what we have also learned is that you know when when you have to use data analytics >> start with something very basic and

simple >> not very complex because many of times the analytics people would love to do some complicated model >> right but that's not the intent so at least in jewelry we saying okay we are opening stores can we put just some simple scorecards over there >> and just start tracking data because many times if you don't have the right data you can't even make complicated models so the starting point here has been just simple data to say which store >> how many they're called JC's jewelry consultant how many of them are there,

how are they converting, what is selling, what is not selling, what is sitting in the inventory. So right now a lot of stuff is happening just observing that in the Krollbach store what sold when uh in the morning or in the evening or who sold it in the this store which has been you spoke about which has opened in Bangalore we'll again monitor. So I think we are trying to learn a lot from the market and using that learning we will start to take action. So I think like I said since we are in it for the long run right now the initial phase just get the right data in place get

some simple ideas start acting on it and the simple idea could be so so many things are selling in this season see these are very new businesses some things are very seasonal in nature so while we may have stocked it and over a period of time we say oh it actually sells only in Diwali or in marriage or whatever so why do we need to stock it through the year we can stock it at that point in time so I think we are trying to look for those patterns which can allow us to run the business better end of the day what do you want you want lower inventory you don't want to carry

a lot of inventory. You want faster conversion, faster. So the basic business metrics are exactly the same. The question is how do you get to it? And the idea is to get the right data, observe on it, do some experiments. I mean the data allows you to do experiments. You try something in some store, it works that becomes your ax which can be applied somewhere else. So I think idea is to learn from data and try to distill that to be used elsewhere as well. But right now it's very early days because every location is very different. We can't say that this works

in this stores. It will work somewhere else. Over time, I do think that we may find some meta patterns to say in these localities that and this is very similar. We can apply the same logic. But some of those things will happen later. Right now, we're just observing store by store what's going on. Over time, maybe we will be able to segment and come up with things which can drive our business better. >> Makes a lot of sense, sir. In fact, you mentioned a very interesting point. I wanted to follow up on that. uh [clears throat] you said that uh there

could be meta patterns that could come out of it that can then be leveraged across right for consumer businesses in the past for example when I've worked at TVS we noticed that you know there are certain data point anchors which are very unlikely anchors you don't think of them as anchors and you realize that you know there could be correlations with a lot of other patterns right uh for example in case of Swiggy we uh I came to know about it through some read that you know they identified that you know for example if a household has a washing machine there's a very high likelihood

or correlation that if someone has a washing machine they would uh you know order via or they can order via Swigi right so let's target those kind of consumers across I know you mentioned that consumer businesses do not contribute as much as some of the non-consumer business you know manufacturing businesses do or trading businesses do but have you ever come across any interesting unique uh data anchor that that has led to very unlikely correlations that most consumers will not know report. Unfortunately, I have to say I haven't

come across this because like I said, all of these businesses are quite independent of each other at this point in time. M >> I do think that at a group level there's a belief that certainly we can bring ideas from jewelry and fashion and you know the the capital business and you know tomorrow paints and all of that but I would say today they are all very independent and some initiatives to bring things together at a group level hasn't really you know succeeded in a big way but I do think that it remains a big area of opportunity I mean we have

seen that the you know Tatas have come up with the loyalty program and so on and so forth We don't have any such program at this this point in time and maybe I would argue that some of our businesses are relatively small and still in the growth mode. They're trying to just sort of grow and come to a certain level and then start to do that. But some things on a bilateral basis keep happening between two businesses. But I won't say that at a group level we have been able to harness all the consumer data in a meaningful way. that remains an aspiration and I do think

that that aspiration will get fulfilled in some time to come because people are seeing the value what you just said that the same consumer and the same behavior how can everyone take advantage of that that's an area which is happening in bits and pieces but I wouldn't say that we have been fully successful but that's an aspiration >> that that was one of the targets like we were discussing earlier also was put on us as well at TVS where uh your colleague and my ex- boss Mahes Khali had mentioned that you know average products per consumer.

>> Yeah. >> Should be much higher than one. Even though TBS is perceived as a longtail set of products, right? Yeah. >> But uh that's where a lot of these ideas were shaping up in our head as a conglomerate as well back in the day about how do we connect them and use data analytics to figure out how do we target different consumers at different touch points. >> I think what tends to happen is when you have large public listed companies with their own goals and their own strategies, uh bring them together at a

point in time is not always possible. So it it requires some effort over a long period of time to get there because they're all having their own goals. They're doing it in a certain way. Some of these horizontal initiatives are not necessarily you know coming together but I think I think the will is there and slowly I'm I'm expecting some of these things will start to happen. >> No I'm sure I mean with the size of the companies that you're speaking about it is obviously a lot more complex to come bring that together. That's why we they have a champion in you to bring that

right. Uh now coming to one of the other giants and other champions within the group which is ultrare tech cement uh >> one of the biggest uh companies if I'm not wrong of the group uh for a firsttime home builder who's trying to use ultra tech cement what are the metrics that are the most important does it does it lie in terms of on-time delivery of cement or the type of or the quantity of cement and concrete that they have to use or uh you know some kind of technology framework which talks about the guidance of how to apply it and so on and so what what metrics or

metric moves the needle the most when it comes to the first- time home builders. So you know one is of course the strong brand that already exists through lots of advertising and customer segmentation all of that to to make sure that people have trust in that brand but if you come to the other side on the back end you know the cost still matters I mean end of the day it is a quantity business and if we have to continue to sort of grow we need to bring the cost down. So reducing cost per bag continues to remain a big metric and the the the the components could be energy and logistics

and procurement and many others. So that's number one. >> Number two, you have a huge dealer network which you need to to sell. >> You don't want very high attrition in that dealer network. >> So how do you make it meaningful for them? So there's a mechanism in which the schemes that you are giving out, do the dealers feel excited? Do they make money? Are the incentives correct or not? Right? So how do you motivate your dealer network to stay with you? Are you able to even predict that some of these dealers are going to attract just like

we do at employee attrition? So what's going on? Are they not making enough money and what can we do to help them make money? Right? So how do you preserve uh you know your dealer network as well. So I think there are some of these things also very critical on the back end that you need to run your efficient very efficiently to continue to reduce cost per bag. Set up newer projects again in a faster time frame at a lower cost. get your dealers with you and of course on your marketing side continue to make sure that your trust that you have uh is preserved. So I

think all of those things clearly there are many other things that have been there. There's an app you as a consumer can go to the app you can do all condo design all of that but unless we get the right cost we get the right dealer incentive and the right plus branding some of the other things may not happen. So I do think that all of them have to come together but some of these are not very visible to people outside and that's why I'm highlighting them. >> No makes a lot of sense. In fact coming to uh dealer excitement uh one is incentivization

uh but uh could there also be a important factor of inventory planning because uh for example in paints industry I've heard that you know it becomes so important and uh to identify which dealer will sell what kind of sq in terms of boxes or cartons of paint on a particular day of that month so that that is being stocked out and accordingly the inventory planning in the whole supply chain logistics of it is working out that How important or pertinent is that in the in the in the cement industry as well?

>> Very very important like in any other industry. So between manufacturing and sales >> there's there's an inbuilt constructive tension if I may to say you know they will always say that this is selling and you should make this and they will make something and they will say they are not you know doing it. So I think this constructed tension exist. I would say still we have a lot of opportunity to bring these things much more seamless and therefore there are some projects which we are working on to ensure that our factories are producing exactly what

is selling. In fact recently I was you know talking to someone in the group uh and he said that you know the the the the founder of the group uh he used to look at only two metrics when he used to come to a factory. So he would come to a factory you are a factory manager he'll say are you at 100% capacity utilization >> you'll say yes are your warehouses totally empty yes that's all he says you make full and you don't keep any inventory so if you think about it we talk about metrics and measures and all of that uh but the late GT Bila and some of the founders at that time also with

no data had a very simple metric in their mind to say am I operating at the highest capacity and I'm not inventory it has been picked up and by the way there were a third metric that he would ask the employees are you happy working here think about it that was a balanced scorecard >> right >> and there was a perta system of a daily P&L that they had so sometimes I wonder whether with all the data and AI heavy over complicated stuff than some of the founders who could get it with their mental math on daily I'm making money

employees are happy factories are running at full capacity there's no inventory and all of that so so just an anecdote on that >> it's a good rule of thumb to just quickly assertate right but I think uh the the supply chains have become since then become hopeful so we don't need all of this I'm not saying we can run it that >> I mean inventory holding cost becomes a very important aspect today right and sometimes reverse logistics as well right these are very important cost factors that that you factor in uh

>> I now will move on to art builder capital there's so many group companies I'm sorry I'm bombarding with so many questions on these uh >> you know we we have started talking about uh loans becoming faster are there any key attributes because of which loans loans have become faster in terms of dispersements. >> See, first of all, capital is a is is a conglomerate in itself, >> right? >> So, there's a NBSC which does loans, there's a asset management company which manages your asset, there's a home

finance company, there's a life insurance company, there's a health insurance company and all of that, right? So, it's again a big thing. Now, data analytics is being used as you can imagine in financing business all over. Um I think the one of the unique part that I want to call out which I think is a bit different compared to others is the the health insurance company which was the last one to sort of come about within the group 7 8 years ago was set up almost like a well-being company not even an insurance company in fact the the the the head of the company he's an

insider uh he said that uh I don't want to hire people who are actually from the health insurance industry because it is sold based on fear I want it to be kind of created based on wellbeing. In fact, I'm a I'm a biased person because I'm a consumer also. So, I have an health app. So, if I do X number of steps, they give me discount. They give me some nudges on what to do. So, I think there again and we our team also works on some projects to reduce fraud, improve underwriting, enhance customer experience and all of that. So, I think a lot can be done. But again, like I said, the purpose here I

think is slightly unique and different to say we're not scaring you that you should buy insurance. we are encouraging you to have better health and hopefully therefore not use insurance. So I think we kind of many of us resonate with that purpose and doing a bunch of projects on that otherwise you know some of these industries everyone is doing the same uh you can do one person better than the other nothing wrong in that but the there's no fun in it I always say that we should solve a problem which has a unique brand of strategy and technology so if there's a new thing that you are

trying to do let's support that if someone is following the same strategy in in any business than the next person we can still help them become better but then the delta is only But if someone is finding differentiated strategy the delta can be far higher. So I think we try to also look at innovativeness uh apart from just the reusability in our projects. No it makes a lot of sense. In fact again it underlines trust right consumer trust which you spoke about as a very important principle of value uh as part of the group right for people from the consumer standpoint. Uh Arthur

Arabella health insurance company covers 22 plus million lives. I don't know how dated is this data but I think it's last year's data. 22 million lives presence in 5,000 plus cities with more than 140,000 uh agents you know spread across and like you said that you know they try to not fear monger they try to talk about the value aspects of the ecosystem of the insurance uh coverage uh so I want to talk to you about this in Europe I read about couple of years back there was insurance health insurance company rather insurance company not health

insurance but I I'll tie it back to health as well uh they use computer vision to let people immediately talk about what was the incident that happened with them and bases the retina scan and the facial scan they would be able to identify whether this person is speaking the truth or not and according the disbbursement would happen a lot faster which made the whole process simpler. I do not know how successful was this company but I think initially they were one of the first companies to have tried this out about two two three years ago

and uh for example I'm at the airport I left my bag on my seat someone you know steals that without me knowing about it so I can just open my app talking into the camera narrate this incident for about a minute or minute and a half and that's all that they need to understand if I'm telling the truth and if that is actually what has happened and in they'll approve or reject the claim And probably in an hour's time that money would hit the person's account is I mean some sometimes it's become as frictionless as that. How do you how do you think we can go a lot faster on the

health insurance claims as well? One me one aspect that you mentioned was gifying the whole thing for example insurance premiums can go up or down depending upon your habits and there'll be incentivizations through the premium payments and by by developing good habits for example steps counts or something else. uh how else do you think it can change? >> So I think agentic AI is going to play a big role and we are working with the team over there. See wherever there's a human agent involved in doing things we trying to see how we can automate or

bring that knowledge into a you know digital agent. [clears throat] >> So think about sales, think about customer service, think about any kind of other adjudication of your underwriting or whatever. So I think what we trying to say can we create specific agents which can amplify what we can do through agents and not necessarily rely on human workers because think about any place where you know humans are a concern you can do only so many calls or you can have so many interaction. Now if you can replicate one not only can you do more

>> you can even do it more effectively. So again going back to faster, better, newer. So I think there is an opportunity over there with this agentic AI to improve the customer experience at all layers of the life cycle because a lot of things are still manual with human dependence and there's a lot of attrition also in these industry. I if you look at just the financial service industry, people do move around from here to there. So the knowledge goes with them. So if you can create knowledge agents who are smarter and can work 24 by7 by the way there's no

restriction. you can I think achieve a lot more and do many things faster and better and maybe newer things as well. >> That's a very interesting point. I think that's a very valid application of aki that that we keep hearing about all the time these days, right? Uh I'm going to move into a very different segment. I want to talk about you chief. Uh a lot of people who know you take a lot of gan from you. You know you have you've showered a lot of wisdom to a lot of us including me. So I want to talk to you about how have you picked up this wisdom over the years? You mentioned right at

the beginning of the conversation about some of those episodes right but uh is there some very interesting fundamental learning that you took up at IIT Delhi or I am Ahmedabad that you still you know apply for taking certain tough decisions. >> See I think what what has happened is that whenever someone has asked me a question I have tried to see what have I done what have I learned or what would I do. So instead of trying to lecture someone on what they should do, I have always personalized it to myself to say what could I do and then shared that

saying you know what if I was in that situation I did that or if I'm in this situation now I would do that. So actually I have used those questions to provoke myself to think and form a model for myself and then I also say that you know that model may or may not work for you because you are a different human being >> at the same time this can be a valid input that you can take away. So I have utilized intuitively all those interactions to learn about myself >> and to me therefore each uh such interaction has been an interaction of

self-reflection. While people think that I'm giving them gan, I'm actually getting gan for myself and then relaying it back to them and with the caveat to say that this may work for me because I'm a different human being. But think about what will work and of course then I say in at your stage in life or whatever I know of you maybe this model has to be contextualized in your case and could be different. So to me these interactions are actually learning experiences and reflections. And to me this interaction today I'm thinking that a lot of thing that you asked me made me

reflect and think of things and I'm certainly I would have learned a bunch of things which I can't articulate today but somewhere some neural networks have been formed and made me better than I was before the interaction. No, I think uh what you have said is not just humble, you being humble, but I mean it's logical of course it's you being humble but I think that also very clearly shows uh empathy and emotional quotient of a leader like you because I think uh I know a lot of people who have good IQ right so we are able to rationalize a

lot of conversations we're able to talk about a lot of things solve a problems but a lot of problems today also have a underlying human layer to it right and that is where the way you've responded to this by saying that you know I think about you think about how would you respond to a particular situation how would you solve for a particular problem talks about empathy which is I think super critical in the world of AI that we are living in today >> I think more than that and this is true for anyone I also don't know what can happen to anyone

>> so how can I tell someone what they should do the world is very different from what it was in the past so even My my father was very sure that I should be an engineer because he was one and maybe he was right because in the days of uh scarcity which we were brought up in things are very less it's easier to make decision the take it or leave it these are two options right >> in the world of abundance the decision-m model has to change so that's what I have also learned and enra to say I can't tell anyone anything because there's so many different paths possible

and but I can help you think about the pros and cons of each path and maybe through your own experimentation you have to learn. So I think today the advantage is that we can through our own experimentation and reflection >> learn about things. >> I use the product management agile methodology analogy to say think of your own self as a product. Think of your own life as a uh you know keep adding features and launching them every two not two weeks every two quarters and then learn from experience and then keep adding. So there is no career path.

There's only a learning path or a or a human path where it will take you will whatever is working for you. Just like Amazon can't predict what five years later the website will look like. How can you predict your you know career stage five years from now? True. >> It's a wrong question to ask. So I think a lot of questions that we are asking today >> are relevant from the previous era and not worth asking anymore. We need to find new questions which are more relevant from the era that is going to come which is an abundance era where

different questions have to be asked. The other easiest one to describe is the two thing that we we consume as human beings is food and information. Right? M >> food information in a scarcity era you can consume as much and hold it because you don't have much so we would hold information >> we would also hold food we get a food we'll just eat it because we don't know when we'll get now today we are applying the same mental model in the abundance >> more food obesity more information consumption mental illness those two elements are happening because the old

mental model being used on new data >> wow >> I think the better mental model is you know patience so for me I don't eat unless I feel like eating. I don't consume any information. I live in the Joo mode of joy of missing out. I have no idea what happened yesterday. If someone won a cricket match, so be it. If I I can ask perplexity or I can ask they'll tell me if I need but why do I need it everything. So I think we all have to figure out our own mental models on how do we want to deal with that abundance. Abundance is great

>> but it requires you to take decisions on what you need to engage with and what not and that is hard. Everyone is engaging with everything because that's how we have been trained. So we need to change our mental model to deal with this new world and that everyone has to figure out their own method. For me it may be different for my daughter it will be different. I tell them that you can't live in a jumo mode that I am because I have come somewhere and I have achieved something for you. Maybe you need more information than I do a certain different type of information. I for

example can get away with says I am I don't care what happened to the market yesterday or what happened somewhere else. But in your team if your bosses feel that you have no clue what happened yesterday they may feel you are dumb. Uh so things like that. So I think everyone has to figure out what they need to do but I am trying to figure out what I have done. I share that with others and say use this as a you know GitHub code utilize [snorts] it form your own model. Makes lot of sense. Thank you so much for sharing that. In fact I'm going to borrow some of these learnings and

analogies that you've just given. Uh I'm going to go to another aspect of your life uh you know very important one which is that of Wells Fargo. In Wells Fargo, it was a very regulated environment and you were trying to take a lot of uh important decisions uh strategic decisions basis data. Uh although the trade-off was very clear that there highly regulated environments around that, right? Was there any interesting episode where you know on paper some strategies seemed very obvious looking at the data points but after interacting with the frontline

workers or associates etc. you realize that you know there has to be a big shift in what the data is showing you versus the holistic vision that you have always thought of. >> So actually I was not so much in the data field in my Wells Fargo tenure. So in some senses when I left Dell I kind of left the analytics field and got into more of a strategy and setting up the offshore center and I did set up a digital team but I was not in data but I think the broader thing that I have observed in both Wells Fargo and frankly healthcare and finance

>> are two industries that are very dear to us. our health and our money we don't want to part with any of it right so I tell people that you know think about it those two also are regulated very well the government also doesn't want those companies to you know part with your health or your money and therefore the pace of experimentation and change >> will by definition will be slow you can't suddenly do an experimentation money disappears or you fall ill right so I think the expectation that people should have from those industries in terms of innovation and there has to be

very different from what you would expect in a Google or something else where it doesn't matter if something's failed so So I do think that people working in these companies sometimes we did hire people from some some of these startups and they would like to do a bunch of things which was not allowed and not relevant and similarly if you go from here to there you have to change your mind so I think again like I go back to what's the strategy what are you trying to achieve what type of experiments you can do to achieve it without taking undue risk that's the

only meta lesson that I have learned >> sure sir thank you uh if I can ask you one more question which is around the startup ecosystem the innovation ecosystem how how do you think are the startup e ecosystem uh bringing that innovation value to large organization is that really happening or is it just more of a phase that we are seeing where startups are creating value >> I don't think that the uh startups are able to influence the large corporate in the way that I would expect them to uh and the reason for that is the large companies work at a very different pace

and their ability to do experiments and absorb experiments is very low so while the big techs are good in all of those things but the other companies I don't see them you know I mean they all have such programs and you have been part of them I have seen many of them >> I don't think the they they have figured out the best way to absorb the knowledge and the the acumen that these startup bring to the table so we use a lot of startup but again I would say we use them in a very tactical way to say oh you made a solution we'll use it and that there's nothing wrong in it

>> I think the value that the startups bring to the table is far higher than that they're able to do like I said the why and the We are using their what which is a very small portion of what they bring to the table. >> Their energy, their ability to think differently, their ability to challenge the status quo. Some of those values I think are far more impactful which I don't think we are able to absorb very well. We just use the final outcome which is not bad but I think the opportunity is far higher.

>> Right? I I personally think that you know sometimes it's also about cultural bringing about cultural change or cultural mindset shift right. uh uh and and and a bunch of these organizations have tried to bring in startup founders just [snorts] to talk about their way of leading lives or how do they run operation on a day-to-day basis just to kind of charge internal employees up a little bit more to understand the difference between uh the privilege that they're going through in life versus the hardships that you can go through outside right so you have a safety net

in the environment in a corporate environment which means that you at the end of the month you're going to get your salary in a startup environment Sometimes that's not the case and that's why the rush to ensure that you know they are innovating they are developing products they're shipping them out faster is a lot higher right and how that change mindset shift can bring so much more value to the corporate world as well of course they have their own complexities and that's why sometimes it's not easy to bring about those changes and sometimes not even viable

but uh that mixing of two worlds is sometimes very >> we have seen some some bad cultures also in some startup because the desire to get too much too quickly at the expense of the uh you know employee wellbeings is also something which is uh which is not very good. So I think everyone can learn from each other. Both have a place in the economy. >> I think we have to figure out how to best uh balance the value that both bring to the table. >> So very last segment short one rapid fire rapid questions rapid answers and

we'll end it on that. So I'm going to throw a bunch of these questions at you and please uh do share your candidate perspectives on this. I know that you always do but still. Uh [clears throat] the first one is on Altra Tech. Ontime trucks or perfect forecast which has a bigger impact? >> First one on time. >> On time trucks. Awesome. Uh Artabilla Fashion Limited. More sizes or more styles for growth? >> More what? >> More sizes >> or more styles or SKs?

>> Styles. >> Sizes. Interesting. Ba opus shade preview or painter matching bigger customer experience when >> uh painter matching >> painter matching uh >> I have to say I don't know these businesses as well so I don't know whether my answers are correct or not some of these CEOs might listen to me and say he's got it wrong so I don't know these are my personal observations yeah >> observations of course capital fewer steps or more explanations what cuts

drop offs more >> fewer steps fewer steps. Okay. One AI data buzz word that you would ban from DEX. >> Digital transformation. [laughter] >> All right. Metric you check most often? >> Steps. My steps. [laughter] >> That's amazing. Uh book you gift the most? >> The invisible women >> by >> I don't remember but it's it's the it's a seminal book whoever wants to think about gender diversity and it's a big topic.

>> I think it just changed the way I thought about that topic. I think the the discussion in the public discourse is very shallow. So I tell everyone that you should read it and if they ever call me to give a talk on women's day I tell them you should all first read this book. >> Brilliant. I'm I'm definitely going to buy that. Uh what is innovation in one word or line for you? >> You know to me that it's faster better newer which I said before. >> No makes sense. Makes a lot of sense. Tool you can't live without at work. Of

course not emailing but any other tool >> nowadays perplexity has become my companion. That's amazing skill every non- tech manager should learn this year. >> I do think that using these these you know any chat GPD type tool in a meaningful way I think it's a great companion because people are still >> so whenever now someone is coming and asking me a question I said have you checked with these tools and tell me what they told you and let's talk on top of that rather than you you me being the first protocol. I think it's a great

place to get some initial idea and then brainstorm on top of that. So I think they should become the base >> and then let's create new knowledge which is probably not available in chat GPT. >> For sure. No that's amazing. Uh one habit your teams should start tomorrow. >> What what habit? >> One habit. >> I think consistency is one area which I find hard in any anyone [clears throat] today because there's so much distraction there's so much happening that doing a few things consistently

people have to pick up. It could be going to the gym and doing an exercise. It could be a weekly meeting. It could be anything. But I do think that people have to remember the power of compounding and find certain thing that they have to do consistently in their life beyond brushing them teeth and having a bath. >> Sometimes they don't do that either. [laughter] >> Uh what is a morning ritual that keeps you going throughout your day? >> My morning walk, run, whatever. And conversation.

>> No, nice conversation. >> Yeah. So I don't do lone run. I'm talking to someone and the term that I have coined is conversation that we chat like this. So I could do this podcast on a run. >> On a run. >> Last last marathon I ran with who's the the head of analytics at Misho and we had a 4hour conversation on lots of topics including data and AI. >> Wow. Okay. I could keep up for 1 hour. I remember the last time I met you at your office for 1 hour. We were just walking. >> Yeah, [laughter]

>> that's amazing. I think I should improve my physical well-being. Uh what would be India's biggest win in the next ticket? presently when you're building in India what is your what do you think India's biggest win is going to be >> I think our digital public infrastructure is a unique thing that we have created and continue to create >> I think how to get more and more people to be aware of it and use it and do something about it in fact one of the thing that I missed in my GCC world was leveraging DPI because DPI can only be used if you are in a domestic business

so the a good thing that I like in my current job is that we are having an India business >> and a lot of things so whether it is the fast tack data to be utilized to figure out where is the traffic going, where is the city expanding, where should I set up the next uh cement plant or the paint distributor or if I can use some data to project something around, what SKUs to sell or where to locate a new store. So a lot of external public data which is now being available and these DPIs are allowing us to utilize this in many meaningful ways that I think is going to

be India's biggest strength and we are also as a nation exporting it to the world at least the UPI and some of these stacks. Last one. So, finish the sentence. Data helps when >> when you have an idea of what your strategy is. >> Fantastic. Thank you. Thank you so much. It was so enriching to get these candid perspectives, a lot of data points from you and I'm getting a very holistic overview of how a large conglomerate very diverse businesses also leverage a lot of AI, a lot of data in leveraging these data points to shape up the

strategy better or actually operate a lot more effectively when it comes to comes to cost or building the top line. So thank you once again sir for taking our time and sharing a lot of this information with us. Thank you. Totally enjoyable time spent with you. >> Thank you. Thank you so much. [music] >> [music]