Thank you so much. Thank you. Good evening. 7:45 in the evening. I admire you for agreeing to listen to me after a long day. Nitin is an old colleague, friend. So, very happy to be here.
I want to cover mostly the impact AI is having on the world. But, I think it's quite ironic that I have to speak on this topic when we are witnessing AI, we're actually in some perverse witnessing natural stupidity in its most extreme form all across the world. Right? So, I wonder whether we should be discussing artificial intelligence or natural stupidity. Right now, that still seems to be winning and causing extreme pain to unintended people all over the world, right? Including many of us in the room who have businesses to run.
See, there are technologies and there are inflection point technologies. And in a lifetime, any of us, even at the speed technology moves, will probably see two or three inflection point technologies being born. I'll give an example to illustrate what I mean.
The web was an inflection point technology, right? When it came. It came before my consciousness. Like, I don't remember when exactly it came. But, I do remember when the iPhone came in 2007. Right? That was an inflection point technology. Because it fundamentally not just that product didn't come, it changed the way everything got done. Like in a manner which nobody could conceive. I think that was one example.
Another I'll say is Google Maps. Like it won't be a travesty to say that Swiggy would not have been born if Google Maps were not born. Or Uber would not have been born. Because fundamentally the location map technology being as accurate as it became laid the foundations of many many companies to be built. There are hundreds of companies. And the way we travel, we buy food, we buy grocery, everything is enabled by the simple fact that map technology exists of a nature which was not possible before. So that is a foundational technology.
AI has been spoken about for many many years, right? And I think what we are talking is generative AI. Like there's a fundamental difference between the two. I in fact remember the first use of AI I had in my life that I did for a business purpose was way back in 2018 when I was in Max Healthcare. And we were looking at can AI replace radiologists or supplement a radiologist? And it was a fascinating problem to see because an X-ray or an MRI is a very defined piece of dimension. It is sized X, your X-ray of chest is placed in my location. So practically anything that a human being sees and interprets can be taught to a machine to do so, right? And it was fairly accurate even back then. That whole process.
Now, it's another matter of where human acceptance of that is much lower. And human acceptance of whatever a machine does is psychologically inferior, and I'll come to that point because it's very important. To give you a simple example. There are thousands of accidents unfortunately happening on the roads in the world every day. But if one autonomous car hits someone, that'll be global news. Whereas data will show that the autonomous cars are far reliable than a human driver is. A human driver sleeps, can consume alcohol and drive, can be irritated, can be after a fight. A machine has no such emotion. It can malfunction, but chances of malfunctioning is unidimensional in nature than multidimensional which a human being can malfunction.
But a lot of us will not trust the autonomous car because of the fact that we are preconditioned for a human being to be driving a car. I have sat in one in US, and it does feel quite bizarre psychologically when you are sitting and suddenly the car starts moving and they have a fake steering because honestly there's no purpose of steering. And I asked them like one of the manufacturers who were trialing, I said, 'You just kept it there to show motion.' There's no real purpose. The steering is doing nothing.
And I can give you multiple examples where our cognitive abilities do not recognize an inflection point technology when it happens. I truly believe generative AI is an inflection point technology. Which will change fundamentally everything.
I'll give you a personal example and I'll come to the corporate one later. Unfortunately, at the place I stay, there was an electrical issue. Some blowout happened in the main switch. Right? Called the electrician, gave a long bill, a quotation. Almost running into 1 and 1/2 lakhs with 15-20 components. I'm a commerce background person. I have no knowledge of electrical engineering at all. Right? And I asked him, he said, 'Sir, I say okay, it may happen. I need all these 50 components.' I just for the heck of it fed into ChatGPT and say, 'I live in Gurgaon. There's a quotation received for this issue. Right? Please verify transparency of prices and quality of brands and products being quoted in this.'
Within 5 minutes, trust me. It gave me a piece by piece benchmarking of prices for location, for Gurgaon. Suggested the sub-component brands which I should not be taking and should be taking something else. Looking at millions of records of safety ratings, Google ratings for those sub-components. Right? And I called him again in 15 minutes. I can promise you even if I'd taken that to electrical engineer, I firmly believe I would have gotten an inferior response. No disrespect to anyone. But which electrical engineer, however so qualified, will know in the Gurgaon market which sub-component is exactly available at what price and which is a better one? Unlikely. Very, very unlikely.
Another because I've led a healthcare organization for 5 years, I'll give you a simple example. With a cough and cold, go five doctors on the same day within 2 hours. Okay? Simple? If you get five identical prescriptions, I'll change my name. You should be getting it, right? You're the same human being carrying the same symptoms. Right? You'll not get it. Because we human beings love thinking everything is not a science, something is an art. It's in our DNA.
Well, obviously, how do we self-identify? How do we self-worth, self-identity? You know, saying I am unique in some ways. Right? So, everybody wants to put the color in especially India. Like China is different. China leaders says, follow. So, AI is actually going to impact India far more because it's going to give us a lot more data-backed answers that don't exist today. Well, they exist. Everything is there. It does actually exist. That's how you get it. But, it doesn't exist in a digestible form.
But, coming back to the doctor example, I firmly believe... Five to seven similar words. I can challenge that this is at level one. And AI will do a far superior job to the 90th percentile medical... Please do not record it. Please do not quote me. I don't want to get fooled by anyone, right? I'm just speaking as a person who understands some data. It's no disrespect to anyone, honestly. I can say this for every profession, including my own, which was consulting for a long period of time. Which frankly at the base level is dead. It's officially dead. The analyst job in consulting is gone. It actually gone 6 months back. People are just discovering it now.
So, is it hype? I think for the... Whenever this confusion exists, now, because I will be put to that because you will be put to that. So, you know, people my age, men, once you come here, I will advise you to leave. Because it's not baggage enough. Well, you better watch it again in life. They think they are out there. I will get out of here. You know, in companies, the intern is asking, 'Yeah, what to do?' And the leader is asking the intern, 'What did they ask?' Right? It's a... So, let me put a stake in the ground and tell you that it is real. If any of you are wondering if it is around the corner, is it real, is it happening, not happening? I officially confirm to you it is real, it is coming after you. Or it's coming with you, it's your choice. Right? It can come with you or it can come after you.
So, I am already... Things are going to change radically. Now, next level, how do you navigate a world where... And everybody will do it differently. The one thing I personally believe in is if we don't do something with our own hands, we always have some suspicion of it or lack of knowledge about it. So, 5 years later, not using AI tools will be like not knowing typing. Right? So, best part is to use everything yourself and not just talking ChatGPT. If you're in business, if you're like a learning legal firm, you must be looking at tools which are specialized in that domain and playing along with it. At least for curiosity, if not for implementation. It is absolutely must. Do not delegate this. If any of you are leaders in the room, do not delegate this.
Just to make it clear in our organization, June cycle appraisal, top 100 leaders, if they are not moving fast on AI, cannot get the highest rating. You can be the superstar, but you're not getting it. You'll be stopped at the next rating. So, it's a scale of 1 to 5. You can get maximum four if your AI performance is not good. You have not adopted enough or pushed enough in the organization. So, we are forcing it. We are putting a force function right now saying, we value it so much that frankly, if you're not adopting it, you're not qualified to be a leader in this week.
We will almost stop hiring anyone who doesn't have AI knowledge. By AI knowledge, we do not... And the beauty is that everyone can have it. I think the advantage is that the leveling playing field has suddenly been flattened. Because simply speaking, what has happened is what was the power available to a computer science graduate, which means high rank, which means hard work on education, has suddenly been leveled overnight. If anybody who's 21 years or 22 years old, hardworking, wanting to learn, can become as effective in 3 months' time. Right? That is a wonderful world to be in. Where I may have made a choice XYZ, but suddenly I can be anything I want to be. Right? So, the arbitrage on degrees has just collapsed massively. And the only arbitrage will remain, and that fight will begin soon, is when professions will safeguard through law and through mandates. Saying, 'You must be XYZ to do XYZ.' Who still have that? That fight will begin very soon.
Because electrical engineering is not needed anymore to evaluate a quotation. I'm not sure. Any other thing? Power plant banana to cut the little thing. Very good thing. But the electrician is not there. So, you will come with a hot day every day. So, some things will get more valued, and I think those prices will rise as we see globally the whole lot in there will go on and on.
But coming back, one is do it yourself. Second, from a leadership level we said, we are force-feeding, we are also helping a lot. Initial three to six months we started the journey about a year back. Initial three to six months a lot of help, a lot of excitement, rewards, you know, let a thousand flowers bloom. We could be sort of getting a lot of those to celebrate the whole of that. We create the motivation, enthusiasm in the whole company. But now it's come to accountability. Right? And no culture in any company flows from the bottom. Bottom-up culture is a myth. I've never seen it. Culture is always top-down. Right? Have you ever seen people in the beyond the company with the culture but it's a good person that the culture is good over there. Is that going to be so nice? Right? The CEO culture is that you will go on and on and on. Right? So, we are force-feeding it from the top 100. Saying, 'You have to move.'
That's one. Now, question is to move on what? Is also very important because look, for better or for worse, there's an AI PR machinery moving on the globe, which is also real. Which is telling you every day saying... And there are millions and millions of dollars going into that PR. So, there's also a lot of noise. Right? And I think that's we have to be careful of as a company and as a responsible institution. Saying and also not lose out. So, it's a careful balance. How much to adopt? How much... What is wasteful money? So, we have taken a very simple framework at least. We said there are things which are already proven being done better by AI happening. Okay, kind of the new model. Both of you have said to me we won't even question. We'll just double down and do it.
Examples contact centers. Chats. Voice AI led lead gen. I'll give you a very simple example. We recruit about 40, 50,000 people a week as delivery partners. Just register the number, 2 lakh a month. Right? Now, for that recruitment, the lead gen is let's say a million. Some number like that. A million people are contacted for us to land 2 lakh people a month. Right? Now, somebody was calling so far saying 'Suresh ji aap ye kaam karna chahenge kya? Aapka ye incentive milega ye hoga.' Whatever the... Now, Suresh ji, the language spread is 42 languages last we counted. The city spread score 750. So, if you just do a grid, it is a crazy grid to manage for you human being net processors.
Today, 20% experiment level already happening. Every time the person is calling, her name is Disha. Right? Disha react negative. Because Disha human being negative. But if I... I wish I could have played that conversation to you if I did not tell you that fact, you will say this is the best conversation I've heard from somebody who's trying to recruit. We've already moved to three, four languages, and I think we can add a language every 3 days now through training. Yeah, I think it's English, Hindi, Tamil okay. The team is adding three languages every day. So, that is already proven, already happening. It's just a question of how fast we can do it, and what is the cost structure, and get the right vendor, right agency, right?
There, in terms of using agencies and vendors, we've taken a very clear approach saying we don't know who the winner is. So, we are not going to bet on anything. We have built our own system where we can plug in and plug out agents, bots, vendors, partners, because we don't want to get stuck with the wrong technology at this point in time. So, that's one where things are working.
Second block is where things are promising enough, and the area is important enough for us. It's a simple two-level decision-making, very important for us, and promising progress being made. There, we have not gone all in, but a certain portion we have moved and taken in at least three best vendors that we are partners. And say, 'You work with X, you work with Y, you work with Z. Clear timelines, two to three months. Want results. Or we'll decide otherwise if it's working or not working.' But we are playing. We are not certain, but we are playing. And the moment we get success, we'll scale it up. And I'm not even asking teams to give me anything any cost benefit immediately. In that second block, even if I get cost neutrality, I'm okay. Because neutrality give us cost benefit to nickel and then I'll do it.
Third block is the wild west. Because everything is happening. Like today we get pitches from legal AI, tax AI. Something or the other, right? People are saying everything can be done. We don't know enough. Right? Or we know something and all kinds of things. There we are trying to do two things. One is we have a small central team whose only job is to plug into the AI world. And keep telling us what are they saying. And is there anything interesting? Bring it to us. Right? Because we just can't process because running a day job is also there. We just can't process that much.
And second is there the belief is that things have not proven at all. But there's a lot of experimentation. And it's very expensive for us to do that at this point in time. I don't mind being a one quarter later follower. If something out of that funnel comes up, I will catch up in a quarter and that quarter is okay. I don't need to be the first of the block in everything. There are no prizes for being the first mover advantage is complete myth. I've seen many second mover companies win. Many third-mover companies win even more. Right? In India especially, first-mover advantage in such a competitive space doesn't exist. Unless and until you're developing some real IP like a drug. Just on patent for 10 years or something like that. So, that fear is not there. Whatever happens, if something happens, we'll catch up.
I think there's an underlying key theme of how do people see it here? At some level everybody feels my role profile will get impacted. Positively or negatively, I don't know. But, it'll get impacted. And that people human beings like certainty. You like knowing what is happening to your life in the next quarter, two quarters, etc. There I think we're saying something very simple. If you learn how to use AI in a deep way, you'll be valuable in the world. Even if you're not valuable here.
And I truly believe that whenever transformation technology come into play, people who understand and use the transformation technology by research, by data become richer than others. It is just true every time it happened in history. You go back to industrial revolution, you can see it then. 1840. Every time the same pattern plays out. Somebody loses, but people who lose are usually people who never adopted that curve. It has never happened in human history that people who went along with the curve on a new technology which is foundational in nature lost out ultimately. They may lose a job. They may get... you know, you make a big deal of it in India but frankly if you're competent the whole world is waiting for you.
So I think develop that confidence with people saying this is something which is really really you're lucky to be learning it. I personally consider myself incredibly lucky to be leading an organization where this is the focus. I'm blessed. I also get articles from my friends saying the CEO job is going to get obliterated in the next 6 months. It's okay like so be it but at least I'll make an informed choice of that happening to me rather than that happening to me because I was ignorant.
I think last but not the least I think it is going to be a process of a lot of pain for a lot of people right also. But that change will ultimately settle well. Especially for a country like India we always need leverage. I mean we have so many problems that we have to have high leverage solutions to use. Like for example Aadhaar is a high leverage solution. I think UPI is a very high leverage solution to how cash was used in the economy. I think these are absolutely scale moves, right? For me I think medicine in India could be a high leverage solution if you adopt AI significantly, right? Where you just don't have doctors in villages and all but this can really really transform how we look at health care. So the possibilities for a country like ours is endless. Right? And it's okay if you are not building foundational models. Actually we never built foundational tech in most cases. So I think there's nothing to be worried about there.
We didn't build the web, but we are the second largest user of the web in the world. We didn't really manufacture the mobile first. We still don't have a national brand which is a global mobile brand. Can you think of one? We don't. But, doesn't matter. We still are probably the second most prolific user of mobile in the world where there are generations in India which have never touched a computer and used the mobile straight away and continue to do so. So, I'm less worried about that. I think we'll always be a large market for applications and applications we have core strength. We'll again play to those core strengths, build really well on top of that. We have enough bright minds who'll do incredibly well.
So, I personally I'm very excited that I'm honestly a student all over again. And the environment is giving me an opportunity to learn something totally new. Like, for example, I have a personally a 25-year-old Gen AI program manager who's my reverse mentor. I spend 45 minutes every second day with him. Where I'm learning... I've never coded in my life before. I'm a most unusual tech company CEO. I've never coded in my life. The first time I have been able to code and I have first delighted. Right? Because simple reason is... So, just to give you a very simple thing. When you are typing on ChatGPT or typing on Claude or anything, I'm super simplifying this. This is not 100% accurate, but you'll get... It is basically taking that English language to a SQL layer and going into databases. So, if you're a coder, you know how to write SQL or any query or technical language. It could be SQL or the other things. That's the core of what is happening.
What you are not seeing when you're typing even saying to Google, 'How's the weather in Gurgaon?' The systems are not reading, 'How's the weather in Gurgaon?' It is going... You would have heard this word tokens. Tokens are simply like each piece of that is actually codified as a computer language. Which it is taking. That's why the electricity usage is so high. You sometimes hear because this is massive electricity... That's why it's happening. And if you remember, so one way to save electricity is when you're prompting, brevity is important. Please don't say good morning to your ChatGPT. Please don't say please. Please don't say thank you. Everything you write costs some electricity sometime. So, learning to prompt with brevity is your contribution to sustainability in the world.
On that note, thank you very much. Thank you so much. If you have any questions for 5 minutes, I'm happy to take them.
