Will AI Kill India's IT Industry? | Adit Jain (Leena AI)

[MUSIC] Thanks, others for being here at Harvard India Conference. There was a wonderful panel of VCs and operators that you were on, which discussed the India-US corridor. And I'm very excited by that because that is exactly why I started this podcast as well. Maybe we can start with an introduction of you. If you can talk about your journey from probably like childhood as well. And what brought you to the US, what brought you to this industry? >> Absolutely. So like, Delhi-based, what enraged my dad is ID Delhi. So do not get a lot of options when it comes to choosing a career. I didn't study a lot till like 10th grade. But then dad was like, are you going to be, I hope it's okay to use him. >> Yeah, you can use him. >> Very knock it by, I'm like, okay, I've got to study. So I studied really, really, really hard. Two years of like 15 hours a day, 500 days flat. I know people used to say, I never studied for ITJ examiners, I had a bullshit, I studied a lot. So got through joint IT, I read extrail, 2011 to 2015. Met these two guys, Mayank and Al-Numan, now my co-founders. We know it's the last 15 years. Same hostel at IT and then that's how we became friends, we won. And it's been just an incredible year, knowing those, those very good friends and then of course co-founders for the last 10 years, it's been a long time. We completely undergrad, had a different job offer. It was like, hey, fuck it, let's try and build something of our own. So started a company called Chatteron, which was a horizontal platform to build chatbots. For people who were following conversationally I since the very beginning, there was the first way which started in 2016, '17, IBM Watson, Google Dialogflow, bunch of a chat fuel, a bunch of other companies. We were the same Arabis to compete with them. You can come in and build a chatbot for anything, your website, Facebook Messenger page, Twitter handle, et cetera, et cetera. Did that for a couple years? We scaled over 30,000 chatbots created. But interestingly, we only had 20 customers paid us. So there was crazy conversion by tricks like really bad. And we looked at those 20 customers. 16 of them were using us for internal IT, HR, chatbots. And at that time, we had zero professional experiences of what the hell is happening? And what he overestates again, there's 20, 70. 20, 70. Yes. We were like, what the hell is happening? So we sat down with them. We realized as organizations grow, there's so many systems, people, applications, departments, finding the right information, and getting your job done the huge time suck. So that's the birth of Lena, late 2017 or 2018. And then we applied to YC with our journey. And we had applied to YC earlier as I got rejected. I didn't know the second attempt. We got accepted. And we got into YC some of 2018 in the level experience. Did it $2 million seed around post YC? And then led by Ilal Guilard, led around. Then 2020 COVID happened. And just the accelerated growth, because a lot of IT HR teams are not in offices anymore. Right. And they act to service their employees. Like, hey, I want to reset my password. I want access to Zoom, etc, etc. All right. All my change, my benefits, right? So those kind of tickets, we automate using our chat but that's the business back then. 2020, acceleration happened. Then 2020, two, LLMs happened. 23, we were looking at our product. We had to give you a little bit about a stack for the technical audience here. A tech stack was like any hour a bird for an intente-density recognition. And it would basically be as a layer on top of a low-code, no-code workflow builder. So we had built these integrations with work days, service now, sales, for SAP, etc. So we basically saying that, hey, if I identify the intent by any hour bird, we'll be able to trigger those and say, hey, if you want to change your benefit, we'll be able to trigger that workflow. Got it. Got it. If that's not the case, we'll go to search. So we had built a search stack on top of a SharePoint, Drive, etc, etc. And then the last was, there's the ticket. So the idea was that, hey, you can automate 40% to 60% of the tickets by this stack. And it worked really well. We had lots of customers. But this was mostly-- we were starting to grow in the US and we had already grown up in India, right? Back then. And then we realized that 23, with LLMs, that this is not going to survive. So we first slapped ragly on top of our knowledge. 24, we were like, shit, this is also not enough. So we kind of just stripped everything about. We built everything from scratch, fully agentic, right? There is no any hour bird. There was nothing like fully agentic from the get-go. And did that 24. We realized there were a lot of learnings throughout that. There was a lot of-- we had to go through an organization and transformation to be able to achieve that. A lot of the engineers were just not OK using AI to code at that time. And we just had to have hard conversations with them. And either they adapt and change to the faster ways of working or move out. So that happened in '24. And it's like a new product, new company. And the growth has been incredible now. Like last 12 months, our business grew like 3.5x. It's been incredible. But it was a very, very hard journey taking your people, your customers along in that transformation. Amazing. Thanks for sharing that. I think there's a lot to unpack there. And I would like to start with IIT. There was some conversation even on the panel today where you were talking about the difference in institutions in India and in the US. And I also was troubled by this. I was part of Insight, which is the student media body at IIT Bombay. And I'd go on and ask, I don't want to name who, but like someone in the management of the school. Are you not worried that 90% of the people are not going down the engineering route and are doing business, or are they going to be doing business? I think that's the biggest problem. Who should the first move be if this word to be changed? I think that's the very deep question. I'll try and distill it very, very quickly. I said this at the panel and go to say it again. that way, but think about it for a second. And there she was. working in New York, never had to like, or always had the opportunity of going home like every weekend or wherever they want, right? And here we are, right? Like so I think if there is opportunity, people will not want to go away. Like think of all of us, right? We have left everything behind and we are here. And it's for the opportunity. But I'm saying maybe the opportunity does not need to be this big. Like it does not need to worth billions of dollars. If you give me a good respectable life and allow me to do my research, absolutely, right? And then give me the resources to do the research, absolutely. So I think there's a very clear answer. Everyone wants to, like nobody wants to, like just why would somebody want to come so far away if they can get the same thing there? That it needs to start with the government. Of course, then corporate will join because they come with projects and use cases which they collaborate on. But like it has to start to thrive with the government. - Makes sense. You've been in the IT industry space, studied about it for a while. It contributes probably seven to nine percent of India's GDP. It is ripe for disruption. A lot of the largest listed companies in India belong to the sector. And first principles, I'll be honest. I don't know too much about this space, but I'll along with you hopefully unpack a little bit of it. But first principles, I'm a little worried about this because all the arbitrages that India has had be it labor which led to the IT industry, cost which led to the manufacturing industry or distribution which because of low internet was possible. A lot of those arbitrages are closing today. And that makes me worry a little about what, not just this industry, but as a whole where India's position will be in the AI wave. Given that there's not a lot of work happening in the infra data models or orchestration layers, a lot of the actions still is in applications which is understandable today. But I want to along with you unpack what the IT industry looks like. For me, there is the physical layer which is chips, then there is cloud, then there's applications and then there's services. But I'll pass on to you to understand when someone says IT industry, what does that beast look like for you could you break it into parts and then we can get into what parts you're interested in and what parts you're trying to capture well, you know. - Absolutely. So I think from this conversation's point of view, when we talk about IT industry and from an India point of view as well when we talk about the IT industry, it's mostly IT services. Let's talk about what IT services are and why they became a big thing in India. So we had a population that was great in English and great with problem solving. - Correct. - And of course the cost arbitrage, right? Like an engineer in the US would cost what? 150 grand back then, like I'm talking 20, 30 years ago, 150, 120, 150 grand and then, you know, here it would be in India it would be nothing in comparison and let's say the efficiency was at 80%, right? 75, 80% in India. So it made all the sense to offshore and outsource all of the coding and engineering to India, it made all the sense. - Correct. - Of course there are many other things like, if you unpack that, it has multiple and many other things including, hey, building the software, maintaining the software, which is support services. And then, Anseli services around that, like, hey, managing the cloud of that. So, in India, IT services does all of that, right? Managing a cloud for a customer, building the application and then, doing support services on the application and then also security services. So that's the Anseli, I'm talking about. The security services to make sure those are secure etc, etc. So, that's a big piece that absolutely is going through a lot of transformation. If you look at, if you look at all the big IT services player with a lot of presence in India, Vipero, IBM, like, name them. - Yes. - A lot of them are not hiring at that speed in for sales, etc. They're not hiring at that speed at all. And they used to, thousands of people getting hired out of tier two, tier three colleges in India, engineering colleges in India. Any branch does not matter. We have mechanical, chemical, whatever you've done. You've come cold and we'll train you how to cold. And that's essentially how the thought was back then. That work is gone. Cloud can write what an entry level engine is going to write in 30 minutes. - Right. - In a full month, it's right in 30 minutes. - Right. - That's just the truth. So, it absolutely requires a deeper thought. Because we're betting big by the way. Like, if you look at our projections, we are saying that the Indian IT services is my revenue generation, going to hit $7,500 billion. And I don't know, 2030, 20, I don't know what that target here is. But if you want to achieve that, like, how do you get there? Because the model is a priority which will work for 30 years. It's just not going to work. So, I think we need to kind of relocate. And the IT company is they need to relocate the model. They are in a lot of pressure, of course. But the overall ecosystem also kind of needs to relocate. What is the IT services of the future? Or how does it evolve in the future? - Understood. And let's break up the IT services space now. So, there is the building software part of it. Then there is ERP, which is applications. And then the CRM part, which is more customer and support. Do you, is this exhaustive? What is the? - Oh, no. - What do the large chunks in the services space look like for you? - Got it. So, from our perspective. So, we only focus on the back office. - Correct. - The CRM, etc. We integrate with Salesforce, but I leave that for some other day. Let's talk about a large company. So, what do they need to run their company running? Is, of course, an ERP, the CRM, ITSM, HRIS. ITSM is information technology. So, you know, that, and HRIS is HR stuff. So, now, you've got these four core, but then depending on the industry, you'll have many more applications as well. So, you typically have, not depending on industry, but this one is like common across is knowledge environment. So, you use a SharePoint or something like that. So, it's going to be some KM work. And then depending on the industry, you will have many, many applications. You're very surprised to know. And average enterprise in the US uses around 250 plus applications. This data submitted by OCTA, because they have all the data. But as you go to those, enterprise, large enterprise, that number will go to like 800 applications. So, now, where does the IT services come in all of this? Right? IT services, or the Indian IT services basically help, a lot of times they will like say that, "Hey, I have a partnership with Service now. I have a partnership with Salesforce." So, you don't even, and this is actually coming with Service now. So, like, you don't even need to have the Service now license. You just outsource all the IT help desk to me. It is called the MSP Business, Managed Service Provider's Business. The Managed Service Provider's Business is basically saying that all your IT support across your applications will be handled by me in India. All right? Now, people will raise tickets on Service now, which is also their license. It could be customer's license or the MSP's license. And human agents in India will receive those tickets. They will solve those tickets and they will together think done and you pay anywhere between 10 to $70 per ticket depending on the severity, etc, etc. That's a very typical model. Now, if you think about just that, I'm talking MSP right now, I'm just talking support. This is not implementation. This is not, and this is technically just support. Right? So, we're just looking at one model that itself, and using our product, you reduce 70% of the tickets coming in. So, basically, if you have 100 people in that MSP, supporting a customer in the US, only 20 or 30 are required now. Right? They are now basically being promoted to be a manager of agents and that's 70% of work is in my age. Yeah, agents, right? Or yeah, colleagues as we call them. So, I think this disruption is here. It is real and it's coming for all parts of it. Yeah. And one other way to look at the, I would say foundation of knowledge in companies is there's a system of record, system of knowledge and system of action. I think the system of action/agency part is what you're trying to hit at, which is not just find the information, but also do some, probably analytics and like, do some action on it and get some work solving. Yeah. When I look at the space, there's a lot of the income, but some of which you mentioned is the SAP Salesforce, etc. Glean seems to be somewhere in the middle where they are trying to get knowledge and probably now getting into some part of the agency/inside space. And I met Dev, Dev, every recently, I think they're also trying to hit the agents part of the space. But what part do you think you want to own? Does someone who isn't the space have to own the records and knowledge and insights part as well before they can actually do action? Or can different players within that value chain own separate pieces? If I were a company, would I work with a SAP and a Glean and a Lina? Or would you want to eat into the complete pie and also own the data at some point? Given the backdrop that it's not very easy to get out of SAP or Salesforce. These are very sticky tools. So when you see this value chain, at least add to the value chain if I've missed something. But what is it that you're doing today and what you want to do going ahead? >> So let's kind of draw a mental picture of, you know, what a system of record really does. So if you think about a system of record, there's a database at the bottom. >> Right. >> There's a BPM layer, the business process management where you're writing the business. That if a lead comes from the US, it's assigned to like a SDR in the US region, right? Like that simple workflow, written in the BPM layer, which is called the business process management layer. So there is database, there is a BPM. There might be an integration layer where you are integrating it with multiple other data sources, like in the same case of the Salesforce, you're integrating it with let's say, marketing campaign, that's running. So data is flowing from a third party, you know, leads provider into Salesforce, right? So there's an integration of sorts, thin layer, but important for business. And then at the top is the user interaction, basically the SDR working on Salesforce, or the sales that are working on Salesforce. So it's called the, like, is the engagement layer essentially, right? So we've talked about these four layers, right? Inherently, if you think about it, they're sticky because, initially they were sticky because of database. So Oracle became becoming a big company because of database. But then they became stickier in the age of cloud because of the BPM layers. Because if you have set up an ERP, SAP Sforhana, for example, or SAP ECC 6.0 and above whatever, it is just like, it has 10,000 rules written in the BPM layer. Who's going to replicate them? And documentation is bad, and all my single lot of times. Why did we do this? To empty God damn years, I don't know what he's about, right? So that's the second thing. And integrations, I think a lot of these companies do a really bad job at anyway being, that's why I said it's a thing. Because all of these big companies, I think, they have a very internal view of the world, right? They're like, we are the biggest, we are the best. They have tens of billions of dollars in their venue. So they do a, I'm not saying like all of them are like, most of them do a great job at the integration layer. And then the engagement layer is also pretty bad, right? So what we are seeing right now is that agents are already eating out the engagement layer, right? Where you are saying that, hey, like instead of the ZEP or the SDR or the sales rep, actually going into sales force is just talking to an agent on Slack or on Teams and getting things out of here. I had a meeting with the CEO of Apple. This is a conversation go updated, right? That's already, it's like, was done two years ago, right? What we are seeing with AI colleagues right now. And in our case also, like instead of going to work day, we see work day or SAP success actors or service now or sales force, like from that perspective, we see that our customers employees, they just interact with the leaner on Teams and Slack. So basically the engagement layer is gone, like that's not even a conversation anymore. And the reason it's gone is because you can orchestrate a cross multiple applications by being a leaner, because see as a SDR or an AI, you're not just on the other hand, as any employee if you want to do anything in IT or in HR, you don't want to, you don't even know, hey, do I go to work day for this or do I go to like a benefit portal for this and then, oh, for my laptop, I need to go to service now or like a software. I don't want to remember all this. The engagement layer, because going to a leaner and we are integrated with so many applications, you get that advantage of just like easy access, which reduces tickets. So it's a win-win. End user is happy, customer is happy because tickets go down. That's why the engagement is gone. Like it's, there's not even a conversation anymore, right? Now from a system of the core perspective, you have the BPM layer, which is the most sticky layer, till date. But what we are seeing more and more is, as AI has evolved, right? And as we call them AI colleagues are coming to life, they're automating end-to-end business processes across multiple applications, across multiple business functions. So let's go deep into that, right? Let's talk about, I'll focus on the back office, but this is generally applicable across the enterprise. If you think about the back office, you have individual contributor job roles that are running around five to seven business processes, 80% of their day, literally do that. So think about somebody residing passwords, and just making things up, but residing passwords, helping and depending on your benefits plan. Very simple, like repetitive stuff, right? And then even let's go to finance and accounts receivable, right? You're chasing now invoices. $20,000, hey customer, I have you not paid, right? So it's like the same thing again and again and again. But if you think about, if you do a work study on any of their workflows, you realize they're touch at least anywhere within six to 12 different applications in every business process. So it is extremely a cross application, workflow automation play, that will give you the efficiencies to be able to automate all of this. And that is, I inherently, so, so, so hard. And at Lena, we've been able to achieve this for the back office, only for the back office, like in IT, in HR, in finance and procurement. Now, what's essentially happening here is, if you look at being able to automate an end to end business process, you have to tell the AI colleague, what is your business process in your company, right? Like, still order has five approval that a million dollars and above it purchase orders, as a mission, right? Just an example, sorry. So, and PNG has 15. Now, that needs to be given to the AI colleague or the agent, or else, how will they know that, "Hey, for this company, I will do this," right? And that's something we introduced to the world called AOP agent operating protocol, where you're literally putting in a written document in English about your business process. You get your best employee in that business function to write that document down, and or, you know, any presentation, any PDF, or you just upload it, and that's what we, the round the AI colleague in your company's specific business process. Now, why did I explain all this? It's because this is the God damn BPM layer. Right? So, what's happening is, you already got the engagement layer abstracted away, and then slowly, instead of the BPM layer, is also getting abstracted away. So, here's my thing to see IOs, right? I'm saying that whenever you are in the process of building new business processes or updating your existing ones, you got to write it in the agent layer, in the AI colleague layer. So, while in the next couple of, quarter and a couple of years, all of these big enterprise applications, like the SAP is a service now, the sales force is well as safe, but it is a real threat because that BPM layer is going to get eroded away into the agent layer. Right? And that is the threat. That is why you see the market, like all of their stocks dropping 50% and so much chaos, and it is a real threat. You've seen this with our customers. If the business process moves to the agent layer, then it's basically a database. And I agree, it's not a threat, like immediate threat of churn, but like, by biggest thing to see IOs, and I do all these concepts, is that, hey, if you have renewals coming up with any of these big applications, you cannot do a heavier renewal. You've got to do a one year, two year, maximum three year renewal, so that it gives you the ability to see out how it goes. And during the transition, do you still need to depend on players that are doing BPM today? Well, you still need to depend on them. And then how would that, and is it an easy transition then, to move from the players who are owning that layer today to someone like you, because you mentioned multiple rules and multiple connectors that would be there, is it, what does that transition look like is something I'm trying to understand. They were sticky, you are saying that they're not going to be anymore, but I would assume that transition phase would not be easy for enterprises to do, but I might be wrong, so I want to understand. You're absolutely right, the transition is going to be a longer transition, but the transition is going from the BPM in the system of record to the BPM in the agent declare. So here's how we go about. We go to see, I always say, you set up a COA center of excellence, which is basically going to be building these AI colleagues for the enterprise, for all the functions. And there's job issues, say, any new business process or updation of old business process is going to happen in the agent declare. Now, why do they bite in the first place? Because there is value out of the boxes, right? Because we are replacing humans, not just the application, we're replacing humans that do the, there is a result, tickets in IT, or in HR, or in finance, etc. And then other business functions as well, business processes as well, right? So you are getting, that's why you're investing at the top, but if you get that COA in place, and the center of excellence in place, you can have them continuously building out the business processes and they get updated in the agent declare and then over one, two, three years. Interesting. You've replaced a couple of smaller applications already at large enterprise customers, not like the big ones at this point, but I think that's going to happen over the next 24, 36 months. Understood. So what I understand by what you're trying to capture today is probably the intent of an employee on what they want to do. And then there's obviously a lot of orchestration and connection happening behind to do tasks. But where I'm headed with the question is that the capturing of intent part, there is a lot of people trying to do that, I think. I don't know if this might be a correct comparison, but Emma is building any employee. Gleene is trying to get access of, okay, what do you want to do and I'll do the rest? I think Devreva was also talking about that to some extent. That capturing of intent layer, what gives someone a right to win there? Is it being, is it distribution and convincing them that you know use other. before someone else comes to you or is it quick value, time to value, essentially for a customer, once you show them a pilot or integrate something for them and show how much better you are as compared to their arcade systems probably or is it something else? You need a question. So if you think about our business, right, we build AI colleagues for the back office where we land the customers on yes, ITH are ticket reduction, right, but a lot of our customers are using us for end-to-end business process automation where it is nothing to do with end-employing at all. It is automating a business process like APN is accounts payable, right, so you get an invoice from vendor, it's in a PDF format. Now you've got to match the invoice with goods received, received, with a contract, you've got to triangulate all that, it's called three way matching, you're doing three way matching and then you're going to post, get an approval if required and then post that payment from the bank to this is nothing to do with end-time, right, this is like an end-to-end business process that we automate with our AI colleagues. So it is already gone to the level where this is a solved problem like capturing the intent like we've been in that also if I want to just specifically talk about that there is a value in depth because the as you said time to value the big thing here for a CIO if I can reduce like 70% of their IT head desk start or HR head desk start this stuff in like two days by not it's real RY millions that are saved, right, so absolutely customer buy for that today but like we ourselves are not just doing IT HR head desk automation, it's like end-to-end business classes automation and that's the future absolutely but if I were to specifically talk about your intent layer question I think clean is more knowledge, right, to do agent with it probably all over the place right now like coding, marketing, sales, I don't think that's the way to win, definitely not. I've done that before we chatted on, more horizontal you get, the more you get lost, right, am I the same, do customers support agents and internal agents, so all the best to them, absolutely, I wish them the best but like I don't think that's the way to win the market. We've talked a lot about how the industry is getting disrupted, how you can win, I would love to talk about the social angle of it as well which we briefly mentioned in the start and someone working at a local capability center or back office based out of India, BPO, KPO, a lot of these decisions of replacing them are talked down, it's not an augmentation technology per se, it's a replacement technology. When I said to think about what they could do or would be doing one year or two years hence, I don't think I have a good answer yet. Does that, like in the consumer internet wave, the gig economy came about in India which has employed a lot of people that probably would be here to stay because people still need last mile delivery, but the IT industry is something which I feel like I haven't I haven't thought enough about, I care about it but I haven't reached hypotheses or answers yet of what it might look like. Does that worry you? If someone who is at an IT back office is listening to this or someone who is managing that team is listening to this, what would your advice to them be? Some people have just set up scaling, retrain etc. But if we could get into a little bit more deeper than that as to I myself don't know what I'll be doing six months down the line. Yes. My job as a knowledge worker as a consultant or an investor has been changed and if I am feeling this sort of anxiety then I'm sure someone who's tasks can completely be automated will be feeling more than that. So yeah, would love to hear your thoughts on that from a moral angle also when you build this, how does that make you feel and what is it that as I asked these people should be doing should be focusing on? The great question I think. So we did a model of our impact in our customers while there are different types of bad at the end of the day it's about cost saving and efficiency and that comes from reducing humans in the process because AI is just faster and way less cost in a bunch of these back office processes that we automate. To tune off like it's 15% of the cost and it's like 10 times faster. It's a no brainer and it does lesser mistakes. It is a no brainer to use AI. I know there's a moral angle to it which I will come to in a second but like it is inevitable. Like you saying that it is morally wrong to do it and that's how I'm not going to do it. You will get beaten by the market. Correct. And forced by the market forces and you will have to do it in a capitalist system. Correct. Correct. I was going to say it was socially society, my answer would be different but it is not right. Correct. So that's the assumption we did a model with our customers. A bunch of our customers at part of the part of it. A lot of SIs, BPO's, etc etc. And of course large enterprise customers here in the US. One of the things that we realized is AI colleagues will need guidance and help in edge cases. So you can say that 20% of the displaced back office individual contributed job roles they have an opportunity to get risk and we've got a full LMS. Like they can take that course today. I'm talking about myself right now. I'm in Alina here right now. They can take that course today and get certified and build the AI colleagues and maintain them and improve them. Got it. But that's 20% of the work. So let's say there's 100 people. Yeah. Any business process. Let's say IT help desk. Right. Because of AI gets down to 30 people. Yeah. Or 20 people. Right. So you have 80 people who are displaced. Out of them, 20 can actually be taken back in creating that managing those AI colleagues, improving them, making sure they're within the guard rails. Correct. Doing spot checks, audits, etc. Right. So that's great. And they'll probably get both the 20 that remain because they will be doing like really L3 level stuff and the 20 that get consumed into building AI colleagues and making the better, they'll be paid a higher dollar per hour rate as well. So it is if you are in that 100. Correct. You should absolutely raise towards using more and more AI in everything you do and probably automating your own job so that you either become the L3 or the or the 20 that get the job. So 40 basically we'll end up having job after all of this is done. Right. Like now. The question which is which is very hard to answer. And this is where the you know morality comes in place and this is a big problem for India is what happens to the other 60. Yeah. Now if you talk about economics and people who are talking about AI taking over the world, etc. You'd say that hey, the the market or the pie or GDP will expand so much that like you just have more. So there are more things to do. So for example, you won't need 100 people because now you need there are many other jobs that are correct. Correct. Correct. But I think there's going to be a transition period towards that. It is a big India problem because like we're definitely not going we're going to probably the fastest in the world but we're not going like 20 percent a year. Yeah. We're going to be going to be going at whatever seven eight nine percent a year. And a lot of these AI different efficiencies that are helping growth in the US increasing the pie in the US because that was anyway a cost center. It was not a revenue center. So this is where I do think there is an opportunity given all of this backdrop. Right. Like economists say this but I don't agree. Here's what I think we should like a restaurant. It's also people like me and our organization and our partners and the government everybody in the ecosystem who has the work on it. The other 60 for the India opportunity they have to start thinking that hey instead of just doing the cost arbitrage because that's irrelevant anymore. Yeah. We have to do outcome based work for in the same field for companies in the US. So we have GCC's global capability around 1950. 10950 of them in India right now. Yeah. We have a lot of BPO's. We have a lot of IT, IT, etc. IT services and IT, etc. I think all of them have to actively disrupt their own models. Like a GCCO has to say I'm not doing a cost arbitrage. I'm going to do outcome based pricing for there's a CEO that technically part of the same company but like they need to go back and they need to say I'm going to change my pricing model because until you change the pricing model your mindset cannot change because if you are if you're going to build a hundred people and 18% margin. How will you use technology that will probably automate bunch of their work away? That's not going to happen right. So I focus on that because I think if you start thinking like that and then you open the canvas. Okay. Now given that AI is there and it can do so much which human possibly could never do. Yeah. What are things I can do for my parent organization and this can be in business process that you already run right. I'm sure there are things that you've never been doing. I'll give you an example. This is incredible right. Helping our organizations in the US are required to keep updated records of their patients right. There's some regulation. Nobody's able to do that because they have like so many customers. How can I check like data for every customer. So technically there's a risk if someone comes and checks they might get a million dollar 10 million dollar fine right or whatever. So I was meeting the CIO and he said you know then they actually used a Lina to build this the building AI colleague that is now going and checking every single customer record oh this phone number looks wrong because I can make that image like triangulate a bunch of data but it has three zeros it's a wrong phone number it's a wrong in the lighting space. and they're now able to do it. Imagine they were spot checking like 10 out of a thousand in a year. Now they can do it every day. So this is just one simple example if we invest in that 60 people and have them find opportunities and drive outcomes for customers, right? Now that customer could like, this is a GCCC or this is an IT, you know, CE or the IT service is a CE or this is a BPOC all have to think this way. They have to disrupt their pricing right now. Like they have to decide, it cannot be people based, it cannot be tickets based, it has to be out of space straight. And once you do that, then you start to think, okay, let me automate as much work as I can on this side, but then also look at opportunities that I can add value, my customer, my parent organization, etc. And I think that needs to very, very deliberate, very, very quick change in thought process. The unfortunate thing is a lot of you meet a lot of folks. A lot of them still want to protect their piece of the pie and they are not incentivized because of the contracts, they're five year multi year contract with them and they will kill themselves. They will not get renewed. But what I'm saying is it's not all glue. There is a lot of opportunity for, but we have to just think differently. - To make sense, thank you for that answer and thank you for being here. I think leaving it at this food for thought for a lot of people who are employed in this industry, not just as employees in the back offices, but also leading it. I think this is a good place to start at least the things you were mentioning about being out more outcome based and contract based and disrupting their own models. And yeah, best of luck to you for leading the change in this industry, best of luck to Lina AI. Look forward to seeing more of you and Lina across the Indian industry and more. And yeah, thank you so much for being here. It was a pleasure hosting you both for this and at Harvard India Conference. Amazing. Is there anything else you'd like to tell anyone listening to this from your experience? Anything you saw or heard at Harvard or otherwise in your journey? - Oh, I think first of all, thank you for having me and it's been a pleasure. This is my first time on the campus. It's beautiful. Like if you were a person in the back office, BPO ID, anywhere, just be maniacal about understanding how you have worked, like auto-medial and entire work yourself as soon as possible. Really be maniacal about automating yourself away. maniacal, that's gonna go really far. And if you are a leader listening to this, I think you've got to look at opportunities beyond just selling hours. - Got it. Amazing. Thank you very much. - That is great speak, Muzdev. - Thank you, man. (upbeat music)