AI investment insights: Top sectors and strategies for maximum ROI

From each one of your business background, do you think is investment in AI appropriate for everybody or there should be some guardrails? I think that's a great question. But one thing I do want to kind of tease out for people, AI is not new. What is new is large language models. So I think that is something I want to share with audiences. When you were buying things on Amazon and you were seeing things for what are the other customers buying. So you already had a recommender system and that has been existing since 2001. So AI has been in existence whether it's for fraud, abuse detection, for payments, what you're doing through Visa. So that has been existing. So I just want to point that out. So the second thing what you mentioned, doctor, is the advent of LLM is actually what makes it super interesting. And because of that, I am a big optimist and a big believer in that. And it's not the snake oil or the tulip mania I was saying, because we saw that with crypto and it's trying to find a use case. Similarly, with the Oculus Lens, people are saying people are just going to spend time on that. But this is real because it is actually the previous panel talking about the why. Because it's making a foundational difference to these businesses, how business is going to get done, how companies are going to operate. So I'm a very big believer in that and how this is going to transform the face of the planet. All right. So you're saying the focus should be on the solution. Correct. Okay. Nikhil? I mean, I would certainly agree to that. And just to add, because AI when it was started, even recent 2020 stats, there is 200 billion plus dollars invested in all AI companies, just AI companies, and majority of it went to legacy models. So while legacy and large LLM models are very attractive, not every AI company is going to be only AI. Now with DeepSeek and everything, everything is going to change in many ways. I would rather focus on actual solution driven companies, solution driven AI models or technologies that can make a huge impact in the industry. It may be cybersecurity, it may be general security, or it may be healthcare, it may be retail, so on and so forth. So I think that's where the true value is. Otherwise, everything's going to be like spike up and down. Just like as soon as DeepSeek released, trillions of dollars lost in the market. So it's going to happen again and again, as long as we are focusing on just the wave of the AI in the market. Thank you, Nikhil. Salil? So, you know, people will, of course, make money and people will also lose money, right? So yeah, we saw it with e-commerce first. Somebody spoke about AI first. So we've seen 30 years back e-commerce first. People invested, 1999, 2000, a lot of companies started. So many, you know, dot-com ventures, right? There were thousands and plethoras of dot-com ventures coming in. People made a lot of money, but people also lost a lot of money, right? There were more people who lost money than people who made money. And then for the last 20 years or so, it was all digital transformation, right? Everybody was doing a transformation, whether they were digitizing or they were doing a transformation and people invested, people made a lot of money, people lost a lot of money. And now it's AI first that is going on currently. I mean, there is, obviously, there are many frameworks for AI, but there's no one governance model, you know, which is out there from an AI perspective. So it's all, my colleagues are talking about LLMs, large language models. DeepSeek, you know, since December and probably last week at Davos, showed what it could do to the entire LLM. Satya was hosting about putting $85 billion this year, you know, and DeepSeek comes out and says that we could put a model for $6 billion and there goes the $85 billion for a toss, right? Same thing, Elon and, you know, everybody else who are trying to invest, you know, billions of dollars. Now they've started rethinking. I mean, think about the Stargate, right? President Trump, you know, brought in, you know, four entrepreneurs, you know, billionaires and wanted $500 billion fund for Stargate, and now they're already rethinking. So it's all about solutioning, I would agree, because obviously, that is what we do. We pick up a problem, right, a challenge, either repetitive in nature or any challenge that a business can have. So pick up either pharmaceutical, you know, vertical, or you can pick up BFS, Banking Finance, Social Insurance, the government, or utilities, right, and look at the challenges and then try to solve those challenges using, you know, these amazing models that are being put up. Look at these pre-trained models that can, A, put up a challenge, what is the ROI, what is the return on investment from there? How do you optimize, you know, from there? In case if you see all these sort of things, then yes, go ahead and think about investing. Otherwise, stay out, right, there's going to be next year or so, we'll start hearing about quantum computing. So right, so that's another area where we can also think about, or looking at right now to invest. So, right, so that's what I have in mind. Asif has already started planning an event for quantum computing, but you raised the point that I want to dig a little bit deeper. So this time, I want you to answer not as an investor, but as a business owner, right? As an introduction, it was said that I have a corporate training gamified platform. And nowadays, when I pitch for that, the first question I'm asked is, does your game use AI? And I say, wait a minute, am I solving a problem you have right now? Or should I just, for facts sake, say, yes, I have an AI module? So and I'll go this time backwards, Salil, since you have such a large company with so many employees, as a business owner, how do you contextualize adoption of AI in your own business? So many areas, from right from finance, to HR, to supply chain, to innovation, right? So many areas, we apply AI within our company, I mean, we are a forward looking company, so we do apply it. Simple example, HR tech, human resources technology. Simple thing of taking the fake candidates away from the market, right? Getting them through an AI questionnaire, making them a code, making sure that they are who they are talking about, clearing all the biases which are out there, taking out all the hallucinations which are out there, and implementing that. So that's a simple thing, right? But finance, reconciliation, you need to, prior to what we've been using now, there would be somebody who would be downloading the bank statements, then reconciling, looking at the screens there. But now, right? LLM does it. That's it, right? You just get a statement, you get to know what is it that you need to know, right? So those are the adoption, but my question is, do you decide to invest in them for your company or do you outsource those processes? No, no, no. We invest ourselves. So we have a huge innovation lab where, I mean, today, I think we probably may have over 450 plus AI engineers that we've invested in, and we are not only implementing things internally, but we're also building products for our customers as well. Fantastic. Yeah. So there is a forensics accounting product that we've brought in that is doing an end-to-end solutioning wherein, obviously, we've heard about the California fires that are going on currently. There are so many claims coming out from businesses. I mean, earlier, it would have taken probably six months to settle those claims by insurance companies. You know, with our solution, it's being done in a matter of seconds. Fantastic. Yeah. And that is something that, you know, we're offering. And there are many, many products, right? We've built a product for the water sector. We've done... We are building a third-party risk management. You mentioned cybersecurity. We are building a third-party risk management tool, which not only does, you know, a third-party risk assessment of third-parties, fourth-parties, fifth-parties, but we're also doing countries monitoring. We're doing operational risk. We're doing financial risk. We're doing cyber risk. We're doing ESG. And we're also doing modern slavery invest, right? So our agents, which somebody spoke about agent-tech AI. So I think our agents, you know, we have single agents and we have multiple agents also, which are going and scanning the whole world's MLM data, bringing it in, and then creating and building assessments from a risk perspective as well. Fantastic. So if I contextualize what you're saying is, you are adopting AI and investing in AI primarily for efficiency purposes. Efficiency internally, right? Externally solving a challenge for our customers, right? And also making sure that a lot of the repetitive tasks and also knowledge-based tasks can be replicated. Wonderful. Nikhil, your business is slightly different. Yeah. So, of course, we don't want to build everything. If there is a better player out there who is building a better model, we just use it. But we heavily use AI internally to optimize the process. Again, AI is a learning process for us. In the sense, what I mean by that is, in the business, there's always challenges. It can be from the client side issues or client conversations or customer service or whatever it may be. So we use AI to understand these patterns, to better prepare ourselves, to better equip our business. So we better be ready for this. Compared to Sunil, what's your workforce? It is AI, of course. So we have 30 AI engineers, not 450. So we have just 30 AI engineers, but we are focused on our retail solutions. When it comes to our HR or customer service, so on and so forth, we don't want to rebuild, reinvent the wheel. If it is already out there, we just use it. But we don't want to keep on adding more problems internally. It will just keep our mind out of what we are trying to solve, actually. So that's what we are focused on. And I think if the businesses are not using AI, they have to, because of the fact that it really helps to optimize in terms of efficiency of a business in much larger scale. There are two aspects of Factorial. One part is the investor and the second one is the advisory. So as an investor, we have been investing in AI-first companies since many years. So many years back, we invested in an Israeli company, which actually does the detection of deep fakes. Like you saw during the first President Trump's election, there were conversations about deep fakes. So that company does a really deep AI check on images, video, and it is used by large news corporations in the world now to solve that problem. So as an investor, I think what we look for is what is the premise you are trying to solve and it could go really deep. And whether it's using an LLM or a small language model, we want to see, is it sustainable? I think we are more pragmatist rather than a short-term, hey, do they have a .AI? So I think that is the philosophy. On the second side, when we are working with other companies, the big companies or the smaller ones, our advice is always about the context of the business. It is around your business situation. Like, for example, if you're operating in China, the data has to be secured in that particular geography or in Germany having data residency. So it completely changes how AI can be deployed, how it can be used. So like you say, God is in the detail. So for us, it's not just a fad. It is about applicability, how it's going to be used, and how does it help our companies or people we advise to earn more revenue, cut down their bottom line, and get more customers. I think that's our mantra. Wonderful. And since you brought up the difference between advisory and investor, I'm going to build on that line of question and we'll start with you first. As an individual investor, if some people are sitting here and say, you know, I have a bunch of money, I want to use it properly. Are there any specific industries in your mind, and each one of you think through this, where AI solutions would be most obvious and therefore ROI will be highest? Can you recommend or suggest some industry sectors? Yeah. So I'll say this investment is always subject to market risk, please. Yes. There's an asterisk there. And there is a potential to lose all the money. Potential to lose all the money. The tectonic plate actually shifted on Friday. Till Friday, the whole world was saying it is all about the large language models. It's all about the big infrastructure, like what Salil was saying. And guess what? On Friday, everything was junked and everyone is like, hey, vertical SaaS is the future. So when people have such kind of dichotomies in their thinking, again, it is go back to the basics. And if you really look at the basics, the example Salil was giving is a beautiful example of traditional industries who are going to be disrupted. Think insurance. Think when you have a car insurance. Why Geico is successful is because it allows you to take the pictures and get an instantaneous review as compared to farmers. So I think when we think about these traditional industries, what we at Factorial are very excited about is these legacy industries of insurance, of healthcare, even though there is a lot of, I shouldn't use the word bureaucracy, I am in DC. There is a lot of oversight on those. Except there are impediments. Thank you. So I think those are the ones where we think is going to be, you as an investor have a long play. But if you have a longer period of investment, I would say transport, like what is going to happen with the whole, just think when the physical and the virtual world come together, there is a lot of inefficiencies in physics. So if you think about the world supply chain, we have $95 trillion worth of economics going around. And we all understood the supply chain difficulties a few years back. So I would give you two perspectives. If you are a short-term investor with a horizon of three to five years, highest perspective at this point of time today at 446, I would say it is the traditional industries of insurance, healthcare, or even SAS, which is not really fully AI. But if you have a long-term aperture, I would say things like fossil fuels, transportations, I think if you have an aperture of logistics, 10 years, I think those would be significantly, you will get a significant upside to your investments there. Thank you, Nitin. To the point answer, I really appreciate that. So Nikhil, you have to follow up. I'll try my best because I'm not a traditional investor per se, but I have a small Robinhood portfolio. I see a lot of amazing technologies that are coming out there, which are solving big, big solutions. If, I mean, I personally invest heavily on quantum because that is the future. It's going to break the standard way of the general computers. Right. So if it's a long-term investment, we just have to see what direction the technology is taking. Right. And what are the industries it's going to solve heavily? What is the major problem? I mean, supply chain is one of the biggest problems for sure. And the climate. And even cybersecurity. I mean, if quantum comes in, the cybersecurity, all protocols will just break just like that. Right. If the quantum computers are already existing, let's say. So these are like existential problems. I mean, imagine a cybersecurity infrastructure of a country destroyed by a quantum computer run by maybe some ex-country, let's say. So it is an existential problem in many ways. That's a long term. But again, when it comes to the quantum, I'm sure, as we're hosting this event, so it's not too far. So it's not that long term too, in my opinion. With what China did with DeepSeek, let's say, I don't think they're too behind on quantum too. So I won't be surprised in the next two to three years, they're going to show up and then say, hey, I'm going to break this. So just understanding that traditional trend that technology is moving to. So we started with very basic computers, and then phones, and then the next thing, then routers, Wi-Fi. So the technology is moving. So we just have to move with the technology, with the investments. Thank you. Salil? So, I mean, I would add federal government and state and local and education industry. So I would add that from a vertical. But then, you know, any solution that could go horizontal across the verticals is what you need. So cybersecurity, of course, any good AI companies that does offensive security in terms of, you know, launching red team attacks, doing penetration testing, right? Coming to the next level, doing red teams, or any company that is defensive security, right? Which has a next level of security options center, right? MXDR, or next level. So, which can go, you know, horizontal across the industries is something that I would suggest people to invest in. And that's going to be the future. So basically, among three of you, you've covered the whole gamut of investing, right? One is niche-selected vertical choices. Other is find and support economy structure, wherever it's growing. And what you are saying is end-to-end solution. And it could, cybersecurity is the main powerful example of that. Anything. Like, you know, any solution, I would say. But cyber, because... It cuts across... Yeah. If you see the industry data also, cyber companies are growing 200-300%, right? I mean, contrary to, you know, what we see on the other side. So, that's why. So, now I'm going to flip. I think one thing, probably just, I think all of these are absolutely right. At the end of the day, humans care about money. Most of us in the capitalist society. So, I think the biggest advantage in front of us is financial services. Whether you're getting loans, whether you're interacting with a bank. How that micro, macro, or at scale, those things are happening, that's going to change. And, of course, our friends in the crypto industry would love to tell us that there's a use case there. I've yet to find that. But, again, I'm sure there is something there. It's an interesting protocol. But you can't put the cart before the wheel at the end. I think if there are... And that is also something we'll quickly try to wrap this up, this section up. It's not only the problem what you're trying to solve. It's also who are the people who are trying to solve it. Because, at this point of time, everyone worth their salt is talking about a quantum computing company or an AI company. So, you have to make sure, as an investor, are they the domain experts? Have they some relevant experience? Have they solved problems like this at scale? So, the example we are giving about the Israeli company. The founder of that company was a commander in the unit 8200. Which is like the most legendary Israeli company on surveillance. So, again, I think that is one of the things we would say is that who is the team and what is their mission to solve is also a very important factor in these investments. Thank you. And also, just to add, right? So, you know, even if you look at the industries, banking, financial, insurance, JP Morgan shares. I mean, they have close to 3,000 AI engineers today, right? So, they are competing with specialist companies like us, you know, who are trying to go out in the market and sell. They are investing. I mean, you take Capital One, massive. I mean, probably they might be calling themselves an AI company, right? So, that is what's happening. So, industry is investing heavily. The government is investing heavily as well. And they are trying to solve their issues. But all that they are solving is repeated business processes. But that was probably solved by the industry maybe 10 years back by introducing robotics process automation. But that is not, you know, use of generative AI and large language models or small language models or, you know, because earlier you needed an H100 machine to run a large language model, right? Or otherwise you would have to keep, you know, paying tokenization to the, you know, to Google and, yeah, you know, everybody of the world, right? So, that's very, very important, right? To know that these people are also investing, like industry is also investing. So, how is it that you could bring in value? And how is it that you could bring in a return on investment to those models is the most important. I mean, there are hundreds of companies who come and pitch, right? Exactly. So, I'm being signaled that our time is up. Sorry. But both of you answered the lightning question I wanted to ask. So, let me give Nikhil the reverse. Two of you are being pitched and you look for characteristics. You gave a very good example of the background. And you gave a very good example of solution is being satisfactory. Nikhil, when you go and pitch to raise money, what are the two or three critical concepts do you pitch to investors? I mean, I think that's an apt question because traditionally I took the money, not gave the money. That's why I'm asking you to pitch again. So, I think, again, I don't think, but it's more towards like how deeply you're connected to the problem. What that problem means to you. Because it's not like tomorrow morning I wake up and I do not like my Gmail, I want to build my own Gmail, let's say. But it has to deeply bother someone at the core level where he had to leave all the other carrier and focus on the solution. I mean, that takes a lot of determination. So, when we pitch to the investor, we show why it is needed. Especially for the market that we're solving. Because the market that we're solving is traditionally overlooked market. In the sense like, these are very small businesses. Any new technology that comes into the market reaches them at the end of the cycle. Once the business goes to IPO and all this stuff, then maybe they give some discounts and these guys use it. But for us, we wanted to change that. So, the technology that we built, we have started with these small businesses. So, traditionally, who do not have technology or who are not familiar with using technology, we started there, giving them access to this technology. And making them look at the business in a different way, in a technology lens, in a way. Thank you. So, I would just add, if I have invested in his business, I look for what his customer acquisition rate is. How much money is he making and what is the burn rate and things like that. Many factors. I know we are running out of time. So, just one last thought. Anything on investment and AI? Investment and AI, I think, look at, of course, the model. The solution that it's offering. The industry it is focusing upon. If it is horizontal, very, very good. And then you can have industry-specific models. And then, ROI. What is the ROI? Fantastic. You already answered. Nitin? So, for me, personally, when we think at Factorio, it's about think big and the opportunity. I think, of course, you look at a lot of these short-term metrics on burn rate, other things. But it's also, if things were great, what is the outcome you can get to? I think that is the philosophy we try to follow. We are very influenced by Greylock. So, I think, you know, we try to use that model. Fantastic. If I can just add one thing. Again, I understand investor perspective that money is the outcome, let's say. But for us in a founder perspective, money should be consequence of the work that we're doing. Of the efforts that we're putting. So, we are traditionally focused on that, solving the problem. Our customer is everything for us in that sense. But, like I said, it has to be the consequence, at least for us as a founder. But if I run behind the money, then I'm losing the tech. I'm losing my customer in many ways. With that, let's give them a big hand.