 for this panel AI is interesting so I can't deny the fact that people are just wanting to you know listen about it and hear about it more so AI has been in existence for decades maybe like good 67 years 60 70 years right but this chart GPD sensation just lateral to the world right and I remember in 2012 December you know when it was just launched chart GPD was all people spoke about for the next six months right and you've seen a lot of startups create their own GPD versions right for multiple purposes and all right but one thing I absolutely love about AI now is that people are seeing how AI is impacting their real day-to-day life right so that is pretty interesting to see so without any further ado I'll just ask my speakers to briefly introduce about themselves and what AI means for your business Hi thanks for having me here I am Siddharth and co-founder of Bharat Agri Bharat Agri is a B2C e-commerce company for farmers where farmers can buy all kinds of seeds, fertilizers, pesticides from the platform and we deliver to their doorstep in three to four days we are an advisory-driven e-commerce where farmers first come with their queries about what they should grow, how they should grow, what kind of fertilizers or pesticides they should use our system recommends them what they should use and then they can place on and what AI, what impact has AI had on Bharat Agri? Multiple ways, ordinarily we have a couple of people on a captive who forced us to build something in AI in early 2023 when chart GPD just got famous so we have used AI in two ways the first is once a user places an order on the platform since these are rural users they have almost never used Flipkart or Amazon so this is their first online order they get very anxious where is my order I have made the payment or they will start asking about okay my seed is delivered how do I use this seed and we invested heavily on customer support where there were humans who were answering these queries fortunately we started building a model where a lot of these queries are now answered by AI we have divided it into three parts the first part is saving costs naturally where if AI answers some bunch of questions we save money in the team the second is advisory we have to answer questions about their crock problems they are growing a crock they don't know what to do and generally these answers are given by an agree doctor and now stage two for us is where AI is able to answer some of these questions I would really welcome you to try it go to Bharat Agri website or have maybe ask some questions that my tomato is yellow and I don't know what to do and the the board or the AI tool will suggest you what you should do that stage two and stage three is where they recommend that the products and help them in ordering those products that is where we have not yet reached so that is something that's still under development where we want our system to recommend the products and also help them in placing the order so that's something that we are trying to do right now thank you sir you're done my name is Anjuman I'm founder of a company called Scaler which is this new age educational institute for higher education in tech I mean in India everybody wants to become an engineer or a doctor and unfortunately in spite of that we believe that the state of education and engineering is kind of broken so just for the field of tech which is computer science AIML if we believe we could be the source of a great source of talent and I mean Scaler as a company tries to nurture that talent if people aspire to go and box the likes of Google Facebook etc then Scaler is sort of that stepping stone for that so we currently run programs one is online for people who have already graduated so it's a year long rigorous program completely done online without having to do a job and the second program is if you just completed your class 12 then instead of going to a regular college you come to us for 40 years and then that's graduated with a best quality degree as well as probably better just out of that last year so we'll figure out how to do that as well as great thank you that's what AI I mean look prior to Scaler I've been an engineer myself I used to work with Facebook I lent the team a little bit Facebook messenger and I always felt that AI was not at the stage where you could leverage just to do something meaningful you could probably just hardcore your logic and that will perform better these LLM models and charge activity there are now certain kinds of parking statements which can which probably I can solve as good as I do so as of today we've actually replaced a lot of our teaching assistants who would help doubts when they are solving problems with AI so we don't completely replace I mean humans with AI but just the fact that now PA is able to do a lot more because now I can be assisting five different people at the same time just because AI is doing 80% very similarly like things around like what kind of problems to suggest as you're learning if you're a beaker and in your learning path so then customizing the learning path that is again like one other case where we'll end with you great thank you thank you hi everyone my name is Abhiru I'm the founder and CEO of a company called Velocity we started Velocity three and a half years back to extend financing to new age brands new age businesses I was a VC in my previous life and I felt that a lot of companies even though they have good cash flows they have difficulty so we try to restrict cash to financing products that we also as for AI I think we at Velocity at least have taken a goal that in 2024 we need to transform ourselves into an AI first company and for that what we have done is that we have a simple task force across multiple things that we do as a company that team is like looking at use cases building a solution building applications internally to address those use cases we talk about a few of them it can range from content creation we are creating content which is the first version of the content typically gets treated by the AI and then it would be a person who would sort of do ensure that the quality control is in place but I think that team is something that we are seeing across the board that even in terms of the way tech development happens a lot of voting can be done by AI but then you would have a developer who would sort of be leveraging that to ensure that it is all piece together nicely and overall there are no mistakes there so we see a big role that AI is going to play in not replacing humans but significantly improving the productivity of humans across across multiple areas that we do including customer service and sales also so we are building multiple applications which will be like pretty impactful in private productivity movement at R&D. Great thank you and I feel that one of the most interesting benefits AI provides is the personalization customization you know the competency of it so in your you know specific industries what role does it play and what do you think are the hidden gems or the hidden gaps which you know potential startups or you interact if you wish to expand can you know fulfill so yeah I will start with you. I was at basically a conference from Jeff Lee who was one of the total chief scientist he said like there are two major breakthroughs that are going to happen through AI in the next five years one is in the field of education and the other is in the field of education. He mentioned that look I mean if you look at the kind of education that aristocrats used to have in early times was that a king would make sure that their sons or daughters have a personal coach all throughout the time however that does not scale the regular masses. One thing that AI can do right now and again like I'm with him on that like AI cannot replace human beings but one thing which AI could do is that over a period of time could evolve to become this personal coach initially assisted through human beings. However like in the long run it could probably become autonomous and the reason why that's a big big business opportunity in the information space is because once you have that model which you possess it's an extremely sticky model. Even AI model knows more about me and how I learn what are the choices that I should do based on who I am then I mean I have no incentive to move to a different place because that other place would know nothing about me and have to build these types of them again. So hence I mean one of its hypothesis was that look the next trillion dollar companies are going to come from these places that can own this data and can create these hyper personalized models on how people learn and how people take care of their well-being educationally one of those industries. You are beautiful. At least on the personalization front that is at the heart of what we do right because we started with the belief that no two businesses are similar. The cash flow of an e-commerce business for example would be peeking out during the festive season but the SaaS company typically would have more recurring cash flows most people rate. And the reason why we started Velocity was because we felt that a one size fits all approach does not really work in financial services but that is what was happening externally. Everyone would just provide a business term loan which is like an 18 month term loan with a fixed GMI repayment and we felt that there is a lot of potential in building financing products which are in line with their requirements in line with their business cash flows and the way we see AI playing a significant role in that is that as of now still that internally we have some frameworks in terms of how we bridge that gap, how we internally figure out and recommend the right product structure to them but a lot of that initial triaging can happen based upon an AI agent. We are building voice agents which will have the initial conversation which will be as an example building upon the bank statement of a business understand the cash flows and also talk to the customer a little bit to understand their requirement but then propose a product which is really in line with their business requirements and in line with their cash flows and that we believe will take the personalization to the next level. It will definitely have a big impact on improving customer satisfaction. They would have a product which meets their requirement much better and we believe that a lot of financing products should eventually move to something like that where we are not really trying to think product version and trying to fit that on the customer need but we are really able to build products in real time which are suited towards their requirements. Personalization is a very interesting aspect behind AI and I will just explain this by an example about how we are using it. So, let us say there are two different farmers one of them is growing anian and another is growing gambling and we know that both of them need the same pesticide. If I just tell them that this pesticide is applicable to both the crops they will maybe not buy it but if I tell them that this pesticide is specifically needed for anian specifically for this disease then they have high chances of buying it because now it is personalized for them. Interestingly AI is helping us do it right now we are able to create images in the catalogs based on what the farm is growing and when the farmer sees an image when the farmer initially tells us that he is growing anian and then the entire app has images about what pesticides are needed for anian they have high probability of buying it. We are looking forward to a stage where we will also be able to generate videos that are personalized for the user. Right now we have read the stage where text and images can be developed and can be shown to them. Maybe in next three to six months we will reach a stage where even videos will be generated automatically based on the input that the user has given to us that would be really next level where the farmer will feel or the user will feel that this video is specifically for me. It's like a doctor speaking to you then the probability of the farmer buying that particular product is very high and that's possible only because of AI because we can't we can't create hundreds of videos every week. You remember. Great, great. So I start this question will be for you. Is that you know considering the fact that the global agriculture market is you know it was supposed to be around 1.7 million USD dollars in 2023 and it's about to obviously grow like you know spend this day. What all what would be your one you know like just a piece of advice for potential agri techs I think in this particular field the biggest advice that I would like to give may not be related to AI but would be that staying with the customers for a long duration of time interacting with them as much as possible really helps in the early stage. One difference between agri tech and any other segment is if I am in back law and let's hypothetically assume I am building a e-commerce or a quick commerce business I can interact with five customers within 100 meters of my house and I can do customer surveys and I can try to see if what I am doing is right or not but that fails when you are building an agri tech starter where you have to maybe shift yourself to a village but just going to the village over weekends will not help which a lot of people do before starting the company. You will have to shift there live with them for at least six to twelve months and learn from them that won't be my advice. Great great and Anshuman this will be for you you know considering the fact that in today's time ID code obviously they are known for it and there's just a constant upscaling and reskilling which is required right so what on a scale of one to ten ten being the most mature what AI pool in India ID pool do you rate right now people are going to beat me for it but our quality of talent actually so India is two parts there's a mass India and there is this niche top 0.1% India the niche top 0.1% India is absolutely ready reason is that AI is actually built on contributors of computer science if you're really really strong on computer science then it's becomes easier for you to then also become an AI engineer if you look at the talent pool for example in open AI or Facebook AI labs or Google AI labs none of them have done any boards on an educational program which is around AI is there just been proper CS engineers who had extremely solid fundamentals and hence were able to learn everything that makes sense then there's this mass pool of India which unfortunately associates themselves with tools if you go and ask a random person they'll say I'm a Java engineer I'm a Python engineer which is absurd no no mechanics in them a screwdriver mechanic and because of that because they stick to that portion they don't I mean their minds aren't actually designed to look at patterns and then like understand AI as much and hence your mass India is actually not ready for for building models as a product there is a big gap that's probably why we exist this poor you are giving 0.1% quality 9 out of 10 the other idea for 9 totally out of time great you know this will be to you you know into like there's this just one but this you know like a boom of AI startups right and I mean everybody is interested you might let's go start an AI startup and almost up right but obviously you need to have a mindset where you will be patient about it right because you have you will have like hundreds of competitors in the same space right and obviously you need to have that patience which you know certain sectional founders might not you know have that patience or they just want you know like quick results and all so what would be your advice to them and how are you planning to you know adopt more AI into you know further processes of velocity so see I think the approach should not be to build AI products first the approach should be to solve a customer need gap or a problem leveraging that technology right and I think as you've already mentioned that a lot of companies are being built around AI across the globe and across a lot of areas the competition is going to be global right so every use case that you're thinking about there is someone else who is sitting in Sillapan Bali who has potentially raised more money than you always building a similar solution right and I think the way to address that and build that into your plan is to think your user first your use case first so if you have a deep insight about your user if you have a deep insight about the vertical that you're operating in that's where differentiated use cases can be built upon so instead of building like a thin wrapper on top of an open LLM it's better for you to sort of pick the vertical that you understand well at least in our case we understand financial services well and we therefore understand that typically when a collections agent is calling a particular customer in India exactly what is the kind of funnization that needs to happen exactly what are the boundary conditions that need to be built to ensure that it is compliant regulatory advice right and when you go that deep understanding your user and understanding your vertical specific context better that's when you can build a differentiated solution and I think that's what my advice would be to people who are thinking of building AI based applications that they should go deep towards a specific niche customer segment and they should build upon understanding of a vertical that they already know well that's where the unique insights and unique use cases are going to come from and while addressing that problem they should definitely think about leveraging all the developments which are happening in India but it has to fundamentally be focused on a very sharply defined customer persona okay thank you and you know in just a few days back we had you know Mr. Chandrasekhar in talking about the AI advisory and it's just you know like just jolted the entire segment but he's saying it's not going to affect the startups particularly it's more into the you know like giants tech giants and corporations so what is the to the three of you what is the current sentiments towards this AI advisory in the space so I think we are fortunate enough that a quick advisory came that it is not able to start up so we are safe because we don't want to run behind someone to take up yours but I think the logic behind that must be that AI can also be misused to create maybe some deep fake videos and deep fake images which may change the sentiment of large amount of population that might be their intention but as long as it's not impacting the startups directly it's good and it's I think it's very very difficult to regulate so they are in a very difficult situation where they have to regulate it to avoid the misuse but it's so difficult to regulate it that I am sure as of today there might be thousands of startups in India who are trying to build some solution and I don't think government has the tools to understand and regulate what they are doing so it will be challenging but as long as they are clear that it won't impact the startups but like the AI startups are you know contained with the packets not going to touch us right like there's no 0.1 percent doubt as well at least for Namaya that's the optimism you want to live in so it's definitely a relief to see the clarification coming in and I think the intent of the regulator is to ensure that there is no like systematic misuse of technology and that's where the focus on big tech comes in from and I think in general we have a regulator which is very much innovation friendly right so I think we are definitely optimistic we believe that there is enough room for us to like operate play experiment and build something and while the initial rumors were a bit concerning I was glad when the clarification came out. I add one more thing to that again the the clarification helps but I am personally at least I am very uncomfortable because when you give AI autonomy over let's say a particular product AI today still is in a state where it's a great tool to increase productivity if it is assisting humans that you have be it developers or be it let's say let's say for a particular function if there is somebody doing x how do you make it my effects then x AI still might not be as mature even when it comes to gen AI or other models to to be given autonomy over a particular product when you expose it to consumers that's where it could possibly cause more startups to be honest they are not building their own like in the models they're using interesting models right and they all have to see it they all go into frenzy so so I mean I think the way for even startups to look at it is like look somebody else building models I'm using the models let me just use it for cases where I'm where I'm sure it's not going to hallucinate and just maybe make it a tool for me to become more and more productive I strongly agree with what he's saying there are some use cases which that does not impact the customer directly it's it's a tool that may help our internal need to improve efficiencies I would just like to give a quick example as an e-commerce company we need to change products on our homepage every day and as of today all the products that we see on the homepage are completely AI generated the images of the products are AI generated now it's an internal tool that our team figured it out it used it and has flooded in our app so that when we want to promote a new product every day the images are automatically created now this does not impact the customer directly so it need not be regulated okay okay and this is my last question before we open it to the audience what apart from your specific industries what would be the other industry or sector which can be massively head to with AI just one one sector apart from yours I think marketing yeah I said healthcare before right today we have so much data I mean I know every single second what was your heartbeat what was your on-season levels I have blood reports today no doctor when they get their diagnosis they are considering all of that data and if there is let's say a system which had access to that then I could when today diagnosis is for symptoms you have a fever I'll give you a paracetamol so that your fever goes away never is a diagnosis on the root cause of your issue right if you if I have so much data then doctors could start giving more accurate diagnosis of the root cause of your issue and that is again another trillion dollar company that will be made in the space because they have so much data have you talked to you I would say one opportunity which is clearly here in now already is to to replace customer service with the AI agents and I think that's going to impact multiple industries a lot of customer service is like basic SOP implementation I think 70% of customer service is basic query resolution like that and that is something that can definitely be done by AI in fact I think last month there was an announcement by a fintech company basically Europe called Clarna that they have replaced their customer service agents with the AI agents and 70% of the queries are are handled through that they've been able to raise to save some 40-50 billion dollars based upon that the total response time is much faster than human agents overall customer satisfaction is much higher so it is not something which is going to happen the next five six years I think it is here and now already and it is up to the enterprises to adopt it thank you so much and now I will open the entire panel for audience questioning no wish okay that's it thank you so much everyone for joining us thanks so much