WEBVTT
00:00:01.721 -->
00:00:11.976Hey guys, welcome to Code and Counsel, another series in all things legal tech, ai and every other domain.
00:00:11.976 -->
00:00:18.609So today we've had a very eminent founding team.
00:00:18.609 -->
00:00:26.471Wish to welcome Rathinamurthy and also Ajay, who are the founding team at Crio.
00:00:26.471 -->
00:00:36.582Crio is an excellent place for training engineers as well and as a company.
00:00:36.582 -->
00:00:49.808We also interacted with Crio and very recently the team sold Crio and now they're exploring their next set of opportunities.
00:00:49.808 -->
00:00:58.396So when they have time for all of these things as well, I thought I'd request Ajay and Ratna to come and talk to our viewers as well.
00:00:58.396 -->
00:01:04.266So welcome on board both of you, thanks, thank you on board with the field, thank you.
00:01:04.266 -->
00:01:09.527So basically just asking, how did Crio start?
00:01:09.527 -->
00:01:12.563I mean, how did Crio come about?
00:01:12.563 -->
00:01:16.984And just talk us through your journey of how you did all of these things as well.
00:01:18.099 -->
00:01:27.731Yeah, I mean I was the flipkart before starting Crio, so I mean being part of the Flipkart journey from 2011.
00:01:27.731 -->
00:01:34.787I think, going to the peak of scaling Flipkart and solving some hard problems.
00:01:34.787 -->
00:01:41.031I think that kind of personally gave me the confidence that building a company is definitely an exciting option.
00:01:41.031 -->
00:01:46.600I think wanting to do a startup and probably have a larger impact was the first thing.
00:01:46.600 -->
00:01:49.525Then I had a friend, Sridhar.
00:01:49.525 -->
00:01:54.447Sridhar and Ajay were classmates in USC, so he moved back to India.
00:01:54.480 -->
00:01:54.965At the time.
00:01:54.965 -->
00:01:56.843He also was exploring opportunities.
00:01:56.843 -->
00:02:04.506Then we kind of got together I think one common space where I think both of us were keen to spend time was education.
00:02:04.506 -->
00:02:15.945I think that's how it started and we started doing some early experiments and we found, even though there's a lot of opportunities for people to learn and how people learn has not fundamentally changed.
00:02:15.945 -->
00:02:26.810So I think one of the initial thought and I think it's kind of stuck with us at Crio was how can people learn the way they learn at work?
00:02:26.810 -->
00:02:34.247I think a lot of us actually graduated, joined the first job, first internship, I think that's where we had a lot of learning.
00:02:35.441 -->
00:02:41.145And probably we're fortunate to have a good company, a good manager where we had opportunity to learn, but not everyone gets it.
00:02:41.145 -->
00:02:48.350I think the fundamental thesis behind Crio was can we actually create a place where someone can learn the way they learn at work?
00:02:48.350 -->
00:02:55.013I think learn by doing and what we call experiential learning was the starting point.
00:02:56.099 -->
00:02:58.567Then we got some of the friends together.
00:02:58.727 -->
00:03:04.342Ajay joined us, a couple of more people joined us as a founding team.
00:03:04.342 -->
00:03:07.129I think that's how we started in 2018.
00:03:07.129 -->
00:03:14.331I think in the beginning we started operations, but also after that, so much has changed, isn't it?
00:03:14.331 -->
00:03:21.586In terms of how learning is now more application-focused, is how we feel as well.
00:03:21.586 -->
00:03:30.068So one of the things that even at Poco that comes up, customers will come and say can you teach us how to do some of these things?
00:03:30.068 -->
00:03:34.127Otherwise, the flip side of this is you'll get too many support requests.
00:03:34.127 -->
00:03:35.503You know you will get.
00:03:35.503 -->
00:03:39.546You know you'll have to do knowledge bases and videos and things like that.
00:03:39.546 -->
00:03:45.389So one of the things that has also fundamentally changed is also how people want to learn.
00:03:45.389 -->
00:03:54.127What are the different things that they can use from an application standpoint application usage standpoint and things like that.
00:03:54.700 -->
00:03:58.371I think that still hasn't really caught up.
00:03:58.371 -->
00:04:14.527But at Enterprises this is one of the other things that I have seen in the sense that you know earlier, you could go to places like Rio Udemy, learn a few languages get trained in some of these things as well.
00:04:15.159 -->
00:04:25.218Now, if you go to enterprise learning, it's more about you know how can I do this from a sustainability point of view, from an ESD standpoint, and so on.
00:04:25.218 -->
00:04:38.095So some changes that are happening there as well, but what you guys did is pretty phenomenal as well, and Ajay Ajay has got even more stellar career as well.
00:04:38.095 -->
00:04:40.661You know, for all of these things, ajay.
00:04:40.982 -->
00:04:40.982.
00:04:41.283 -->
00:04:44.711I wouldn't say that, but good opportunities at places that I can.
00:04:44.711 -->
00:04:59.189I think I finished my master's in the US and got an opportunity to work at a startup called Unheadport Systems which did a lot of email and web security and compliance access control in enterprises.
00:04:59.189 -->
00:05:04.672I was part of their data analytics team back when doing machine learning.
00:05:04.672 -->
00:05:06.425Machine learning was not that common.
00:05:06.425 -->
00:05:09.788So, 2008 is what I'm talking about, right?
00:05:09.788 -->
00:05:12.348So that company did well.
00:05:12.348 -->
00:05:13.485It was acquired by Cisco.
00:05:14.399 -->
00:05:21.589Got to see how the data space was evolving so rapidly and how machine learning would actually make an impact.
00:05:21.589 -->
00:05:23.608Like spam, filtering was new back then.
00:05:23.608 -->
00:05:28.959People used to get a lot of spam if you had an email account back in 2008.
00:05:29.440 -->
00:05:33.012Some guy who came up with Razor some marvelous.
00:05:33.012 -->
00:05:37.127Sorry to interrupt, it's a Bayesian spam filter.
00:05:38.860 -->
00:05:44.507I think Polygram also wrote some of those initial spam filtering algorithms, vipul's Razor.
00:05:45.168 -->
00:05:52.677He was quite active on something called CIX in Bangalore a long time back.
00:05:52.677 -->
00:05:59.071I remember that he had written something which impacted something like this.
00:05:59.071 -->
00:06:00.514But sorry, I keep interrupting you.
00:06:03.461 -->
00:06:03.661Yeah.
00:06:03.661 -->
00:06:10.687So I was there for a few years, got to explore other opportunities within Cisco, go to the data centers and other things.
00:06:10.687 -->
00:06:25.269Then, from the new opportunity, I joined a startup called Nest which was doing home automation, like intelligent home products, started by accepting people, you know again, extremely good experience there.
00:06:25.269 -->
00:06:29.771Hardware we were building our own hardware and software to go with it.
00:06:29.771 -->
00:06:34.165So hardware, software, services all under the same roof, very much like the Apple culture.
00:06:34.165 -->
00:06:40.726In fact, 90% of the company was XApple, so that was a very good.
00:06:40.726 -->
00:06:48.803And this was acquired by Google and I got to see how a startup, how do they go about building their products?
00:06:48.803 -->
00:06:54.545How do you escape to the first N number of users and from there how do you go larger.
00:06:54.545 -->
00:06:56.531That was a very, very good experience.
00:06:56.572 -->
00:07:05.225And then going through an acquisition and what happens post acquisition, because I was at Google for about four to five years after the acquisition as well, so I got a really good idea about this.
00:07:05.225 -->
00:07:18.970I decided to come back to India after that, before the kids grew up, and that's when Sridhar, my classmate from Masters, he was calling me and every time I visited.
00:07:18.970 -->
00:07:19.271India.
00:07:19.271 -->
00:07:26.781I would meet with him and he'd say he'd come a few years before me and he was evaluating startups in that door.
00:07:27.242 -->
00:07:36.050He had many prototypes and many small experiments he had done and he had done many experiments on cryo, what eventually became cryo.
00:07:36.050 -->
00:07:47.910Going to a college running some coding camps there to see how we can actually teach these college graduates how to pick up coding in a more project-based way give them a project?
00:07:48.459 -->
00:07:50.228In the course of building the project, can we teach?
00:07:50.247 -->
00:07:53.086them how to do coding, even if they didn't know much at the start.
00:07:53.086 -->
00:07:58.069So look at the early platform implementations of Crio as well.
00:07:58.069 -->
00:08:04.526Shared a lot of ideas, but then eventually said, okay, fine, you know, this is very interesting, let's get together and start.
00:08:04.526 -->
00:08:05.449So's.
00:08:05.490 -->
00:08:20.007when I joined Crio as part of the founding team, so it wouldn't be wrong to say that if you get Ajay in your team, he's been through three acquisitions now 100% of acquisitions.
00:08:23.439 -->
00:08:24.502So something like this.
00:08:24.502 -->
00:08:26.769So that's when the standard record comes in.
00:08:26.769 -->
00:08:37.037I think very few people have actually gone through.
00:08:37.037 -->
00:08:42.985I can't think of anyone other than Ajay who has actually been through three equations, things like this.
00:08:42.985 -->
00:08:49.788Maybe in the US there may be one or two founders who have been through something like this, but notable companies as well.
00:08:49.788 -->
00:09:01.988But also you know it's interesting that you mentioned this the way things have evolved or are evolving, even with the next generation coming to the workforce and things like that.
00:09:01.988 -->
00:09:06.990These guys have already been trained on application programming.
00:09:06.990 -->
00:09:10.389They might know a few languages as well.
00:09:10.389 -->
00:09:20.907Nowadays, if you are in ICSC or some of those, you start from standard 7 on Python programming and before that on something else.
00:09:20.907 -->
00:09:36.125I remember we were trained on BASIC and basically how to start the machine and shut it down correct Apart from a few other things, maybe a little, but that was nothing compared to what the kids are as well.
00:09:36.184 -->
00:10:02.330So I think in the next couple of years, if you ask me, and also with the advent of AI, correct, and the way that it is used, the way applications are designed, how products come into the market, how companies are built, are fundamentally going to change as well, and these are some things that you've been thinking about as well.
00:10:02.330 -->
00:10:04.287So what do you think, ratna?
00:10:04.287 -->
00:10:05.541Do you think this will change?
00:10:05.541 -->
00:10:07.186Uh, how, how will?
00:10:07.186 -->
00:10:08.811How will products get built?
00:10:08.811 -->
00:10:11.378How will the air impacts on these things?
00:10:11.378 -->
00:10:17.962You've been, you've built companies, you've you've also been with companies, so you've seen the journey as well, right.
00:10:17.982 -->
00:10:19.326so how do you think the next?
00:10:19.326 -->
00:10:23.182I would say four to five years, I think even that is a stretch sometimes, correct?
00:10:23.182 -->
00:10:25.431I think we have been dabbling with it.
00:10:25.431 -->
00:10:42.448I think what I could sense is I could definitely sense a huge paradigm shift in the way software is getting built, I think with all the foundational elements and with kind of fine-tuning is available on top of it for every domain.
00:10:44.019 -->
00:10:49.966I think the way the software building is going to happen is definitely going to get reimagined.
00:10:49.966 -->
00:10:56.880So I think suddenly you have a machine which has the world knowledge and you can instruct the machine, you can train the machine right.
00:10:56.880 -->
00:11:03.947I think that makes it so much powerful for the builders to build much, much intelligent software I think otherwise.
00:11:04.027 -->
00:11:04.870I, I think you.
00:11:04.889 -->
00:11:08.388Just look at the software like pre-a era right, I think.
00:11:08.408 -->
00:11:15.149So far it's mostly, you have a database, you have interfaces you can store, you can compute, you can just probably present the data.
00:11:15.590 -->
00:11:23.388I think there's nothing and probably a little bit of here and there, machine learning models, mostly leveraged like advertising, probably in search.
00:11:23.950 -->
00:11:25.273It's a very limited application.
00:11:26.620 -->
00:11:33.509I think the way I see it is probably in the next few years I think a lot of applications will be definitely reimagined.
00:11:33.509 -->
00:11:40.628I think the way to think about solving a problem itself I see it's fundamentally changing.
00:11:41.740 -->
00:11:44.989Suddenly you have a lot of power Suddenly.
00:11:45.009 -->
00:11:52.485The interactions can be multimodal People can talk and machines can understand and other way around as well.
00:11:52.485 -->
00:12:02.027Suddenly, the interfaces are changing and you have the models which are so powerful, which can process, which can actually create stuff.
00:12:02.027 -->
00:12:07.287So I think what was possible with software versus what is going to be possible is fundamentally changing.
00:12:07.287 -->
00:12:18.145I think I see a possibility of a lot of applications getting reimagined and getting access to so much powerful applications right from consumer applications across the enterprise.
00:12:18.346 -->
00:12:19.610So that's why I see this evolving.
00:12:21.801 -->
00:12:22.020Probably.
00:12:22.020 -->
00:12:25.129It's a matter of probably a few more years, I think consumer app.
00:12:25.188 -->
00:12:26.421I think you would have seen.
00:12:26.883 -->
00:12:30.520Apple's WTC demo yesterday, right?
00:12:30.520 -->
00:12:33.648I think now they just started putting it in all the Apple devices, right?
00:12:33.648 -->
00:12:36.341I think it can get as common as that, right?
00:12:36.341 -->
00:12:42.863So, when it comes to every device, I think they're putting on-device AI capabilities where it gets into a lot of your personal data.
00:12:42.863 -->
00:12:48.870It's personalizing for you right, so I think it's going to be very, very commonly adopted.
00:12:49.240 -->
00:12:55.389I think that you know, and this podcast for viewers is being recorded in June of 2024.
00:12:55.389 -->
00:13:07.033And there's been a recent Apple WDC event which showcased a lot of AI integrations in Apple and with a focus on privacy.
00:13:07.033 -->
00:13:26.168I think, Ajay, you've been at Nest so you know some of the privacy issues that come up when you use AI and automation technologies in the consumer space as well, and Apple also made quite a few uh noises around.
00:13:26.328 -->
00:13:35.024Uh, how they, how they've safeguarding privacy, yeah, how they have safeguarding privacy here, which and if it, you know it basically is what ping their own llm?
00:13:35.024 -->
00:13:47.014If it doesn't, then it's going to go to the cloud and figure out which of us is there and then, you know, do a call to chat GPT to see if there's an answer for all of these.
00:13:47.014 -->
00:13:53.774Ideally, I'm sure that computing will reach a stage where it's on-device.
00:13:53.774 -->
00:14:02.365You know, on-device LLMs are going to become much more easier to use and you can load any LLM that you want.
00:14:02.365 -->
00:14:04.345Essentially, at that point in time.
00:14:04.345 -->
00:14:14.251But I think, from a privacy standpoint, and if you look at AI and how it's impacting different things, how do you, you know, what are your thoughts on this?
00:14:14.251 -->
00:14:14.793See, there's two aspects.
00:14:14.793 -->
00:14:16.918One is regulation and then there's privacy things.
00:14:16.918 -->
00:14:18.000What are your thoughts on this?
00:14:18.000 -->
00:14:18.421There's two aspects.
00:14:18.441 -->
00:14:19.703One is regulation and then there's privacy.
00:14:19.703 -->
00:14:21.409I'm going to separate that.
00:14:21.409 -->
00:14:36.933When it comes to privacy, if people see good value from sharing their private information, if it's making a marked difference in the way they experience software or the value that they can get out of the software, they're more willing to give up some private information.
00:14:36.933 -->
00:14:47.216For example, with Nest we knew when people were entering their houses because the thermostats that we were selling and heating and cooling is very expensive.
00:14:47.216 -->
00:14:48.587It can become very expensive.
00:14:48.587 -->
00:14:54.673If your heater is running all day or AC is running all day, it becomes very expensive to run.
00:14:55.600 -->
00:15:05.177If you don't run it at all, then coming into the house the house is uncomfortable, it's very cold or extremely hot, and so the Nest thermostat knew when exactly you entered a house.
00:15:05.177 -->
00:15:08.105Now, that can be extremely private information.
00:15:08.105 -->
00:15:10.610That's very private information.
00:15:10.929 -->
00:15:18.400Now, you could also have nest devices in your bedrooms, in your corridors and passages in the house.
00:15:18.400 -->
00:15:31.120So, depending on whether you were in a particular room, we could think of localizing the heating or cooling to a specific area of the house.
00:15:31.120 -->
00:15:35.134So with this data, you know where in the house a person actually is.
00:15:35.134 -->
00:15:42.626But people are willing to give up this information to nest because they see a tremendous value in what is coming up.
00:15:43.067 -->
00:16:00.192I think similar things will play out with the AI wave as well, if you can show really good value and you clearly state this is the information, this is your private information I need in order to deliver this value to you, then it's a question of informed consent.
00:16:00.192 -->
00:16:03.253So informed consent is fine.
00:16:03.253 -->
00:16:15.130As long as people know exactly what they're signing up for, what they're giving up, what data they're giving up in exchange for a certain tangible value to their own lives, then I think it is fine.
00:16:16.445 -->
00:16:21.880Regulation is different, but that is something you're an expert at, I think you know especially.
00:16:21.880 -->
00:16:37.897I face this, especially because we have products at Poco that use extensive AI, I mean which uses AI extensively to ensure accuracy of plans and things like that.
00:16:37.897 -->
00:16:40.152There are two ways to look at this.
00:16:40.152 -->
00:16:53.225We are not focused, for instance, we are not always building consumer-facing products, and what happens there is people ask saying how does this impact you?
00:16:53.764 -->
00:16:56.658It actually doesn't impact them ask saying you know how does this impact you?
00:16:56.658 -->
00:16:57.642It actually doesn't impact them.
00:16:57.642 -->
00:17:03.148Sometimes if you use, if you do a direct version of chat GPT, you enter the name of the company and you start.
00:17:03.148 -->
00:17:05.174As an example, you will be drafting an agreement.
00:17:05.174 -->
00:17:11.388This is you know, this agreement is entered into so-and-so company.
00:17:11.388 -->
00:17:14.156You give the name of the company with this registered office.
00:17:14.156 -->
00:17:22.893Now, today, you know, if you use some of the latest models, it'll because the data which is already there in the element automatically fill out that list.
00:17:24.345 -->
00:17:27.012People get people think how did it know about all of these things?
00:17:27.012 -->
00:17:33.746It is, it is available publicly, so people don't realize.
00:17:33.746 -->
00:17:35.428They think it's an invasion of privacy.
00:17:35.428 -->
00:17:46.961So if you ask me, even if you look at the trajectory of like consumer apps that are there on, say, google Play Store, or even, you know, say, for ChatGPT correct, that particular application.
00:17:46.961 -->
00:17:55.098The others are more in terms of it's like a solution looking for a customer rather than the other way around, correct.
00:17:55.525 -->
00:17:59.308But if you look at the enterprise space, there are so many.
00:17:59.308 -->
00:18:08.253You're dealing with millions of contracts, thousands of people working, people coming in, people leaving the company, people joining in.
00:18:08.253 -->
00:18:10.772So there are quite a few things.
00:18:10.772 -->
00:18:22.373It's an ecosystem of its own, where some of these things can be custom deployed, and it's much more appreciated because these corporations tend to be gargantuan in size.
00:18:22.373 -->
00:18:24.967It's not necessary that it's so.
00:18:24.967 -->
00:18:39.594You would think that even Gen AI, for that matter, the upcoming Gen AI, and even the current versions of what's available are more probably suited in the enterprise space and in the consumer space, even if you look at Apple.
00:18:39.594 -->
00:18:40.390It's like.
00:18:40.390 -->
00:18:41.683Summarize this document.
00:18:41.683 -->
00:18:48.871How many times have we asked in daily life to summarize an email.
00:18:48.911 -->
00:18:49.834I think pretty early.
00:18:49.834 -->
00:18:50.556It's pretty early.
00:18:50.556 -->
00:18:57.208It's putting out something to figure out what works, yeah, so it's like a solution looking for a problem correct, rather than the other way around.
00:18:57.569 -->
00:19:00.153So that's how the important part is on privacy.
00:19:00.285 -->
00:19:07.751So I would think that the privacy debate is misplaced to a large extent because of what we just discussed as well.
00:19:07.751 -->
00:19:19.829Yes, there are privacy concerns, but if you're looking at you know, say, an in-house sort of like a you know deployment where you really need the data, imagine you're a company with, say, 500 subsidiaries.
00:19:19.829 -->
00:19:25.116You want a line item on some annual report from 2006.
00:19:25.116 -->
00:19:32.633Now go and figure out first, find the report, open it, get the approvers to get it to do all of this.
00:19:32.633 -->
00:19:33.756It's not possible.
00:19:34.019 -->
00:19:43.294It's a couple of days in an enterprise, but you really ask something with this program to give accurate results, you'll be able to get something out.
00:19:43.294 -->
00:19:51.991This is an example where an enterprise at a level gives up some amount of its data, but it's running internally as well.
00:19:51.991 -->
00:20:02.673And as long as people need to realize that when they're working in corporations, it's the corporation's privacy that we should be concerned about, not necessarily like an individual privacy.
00:20:02.673 -->
00:20:13.951I feel that sometimes in the AI debate, these two things get intermixed and people don't have a complete idea about how some of these things actually run and pan out and things like that.
00:20:13.951 -->
00:20:44.405But you know, you also just moving on to the next sort of topic on this related, I think, as we run companies, we've also found that certain functional areas say and we discussed about this as well say, maybe legal or finance or tax, for that matter, if you're doing and these become very complex, very, very soon correct and it's not necessarily given that every company has the best advice possible, right, you know?
00:20:44.907 -->
00:20:59.255uh, it may also not be the case where a lawyer runs a company where he might know, have some answers on legal and compliance, or maybe a chartered accountant runs a company where he may have the domain knowledge and accounting, but overall.
00:20:59.255 -->
00:21:02.172If you look at the ecosystem, it's not just limited to India.
00:21:02.172 -->
00:21:02.532As well.
00:21:02.532 -->
00:21:06.813As you expand globally, you will find a myriad set of actors.
00:21:08.148 -->
00:21:11.838I was in Australia recently, where all the accounting firms are led by Indians.
00:21:11.838 -->
00:21:15.067I'm like, okay, I know where the baseline is so sorry.
00:21:15.147 -->
00:21:21.915I know where the baseline is so sorry, but you all faced some set of issues.
00:21:21.915 -->
00:21:25.796We'll not be able to get homogenized responses.
00:21:25.796 -->
00:21:28.593The quality varies.
00:21:28.593 -->
00:21:38.550The persons that you're interacting with are not long-term players, and then they may not be.
00:21:38.550 -->
00:21:39.534You know earlier.
00:21:39.534 -->
00:21:44.675You would have firms that would actually be the repositories of institutional history.
00:21:44.675 -->
00:21:48.653Institutional knowledge, institutional context generally get missed.
00:21:49.035 -->
00:21:49.757Get missed right.
00:21:50.204 -->
00:21:54.763Now, you know so far, but in the age of AI some of these things can be overcome.
00:21:54.784 -->
00:21:57.492Yeah, so because it's no longer a case.
00:21:57.492 -->
00:22:18.432And how do you feel if I offered you a legal stack that did end-to-end for you I would wake up at 3 am in the morning and you could work on it and not complain and where you'd be able to do so many things with it correct, right from drafting or through workflows.
00:22:18.432 -->
00:22:23.736Whatever the case may be correct, some of these things did not exist, or existed in bits and pieces earlier.
00:22:23.736 -->
00:22:36.354What is your thought as a founder correct, saying, you know, would these things have impacted the way that I ran my company correct or I interacted with my customers?
00:22:36.354 -->
00:22:38.452What do you think about all this?
00:22:38.904 -->
00:22:41.634Generally for example, until I started the company.
00:22:41.634 -->
00:22:43.972In large companies, legal is a black box.
00:22:43.972 -->
00:22:46.753Whenever I have something, I'll go to my legal counterpart.
00:22:46.753 -->
00:22:47.247I'll ask them.
00:22:47.247 -->
00:22:48.190They'll give an answer.
00:22:48.190 -->
00:22:52.653It's mostly sorted After we started running our own company.
00:22:52.653 -->
00:22:54.226I was looking at mostly the corporate functions.
00:22:54.226 -->
00:22:54.986I think there's a lot of own company.
00:22:54.986 -->
00:22:57.048I think I was looking at mostly the corporate functions, right?
00:22:57.048 -->
00:22:58.588So I think there's a lot of complexity.
00:22:58.608 -->
00:22:59.868First of all, you don't know what you don't know.
00:23:00.549 -->
00:23:05.452Right, and I think we had to rely on some trusted advisors.
00:23:05.613 -->
00:23:12.277Right, and we also realized, beyond a point, it's not just if you go to advisors without right base, they also cannot help you.
00:23:12.277 -->
00:23:21.403If you don't put right processes for example, if I just don't keep all my contracts in one place I mean there's due diligence there's no way I'm basically going to help you.
00:23:21.403 -->
00:23:25.067So I think we went through almost including the one for the M&A.
00:23:25.067 -->
00:23:36.467We went through three DDs right, so dealt with three kind of legal forms, dd on us, right.
00:23:36.467 -->
00:23:37.717So I think there's a lot of learnings there in terms of right.
00:23:37.717 -->
00:23:42.324From initially we were not very sure about, like, what is the importance of these contracts, employment contracts, keeping all the vendor contracts?
00:23:42.324 -->
00:23:52.131I think there are cases, for example, one particular vendor contract had one particular class which we kind of ignored, for example, especially when you're going through M&A.
00:23:52.988 -->
00:23:58.011One contract would say if you are going through change of control, you have to inform me and you have to get my consent.
00:23:59.105 -->
00:24:02.634He would have never imagined that he would sign the contract.
00:24:02.805 -->
00:24:05.851He just went there and signed the contract, but that came in the way.
00:24:05.851 -->
00:24:08.593So I think definitely it's very, very critical.
00:24:08.593 -->
00:24:19.210I think what we realize is putting the right processes and having the right repositories, having the right processes to review and sign whatever we're signing.
00:24:19.852 -->
00:24:27.044All of this is very critical, and that's my learning Just having a cluster at one place won't help you to put the right process in place?
00:24:28.207 -->
00:24:28.907No, absolutely.
00:24:28.907 -->
00:24:34.196Also, Ajay, I'm sure you're part of these DDS and I would not sort of like.
00:24:34.196 -->
00:24:39.575I mean, DDS tend to be quite pressurizing.
00:24:39.575 -->
00:24:42.393Especially on the founders and founding team correct.
00:24:42.393 -->
00:24:47.307It tends to be I think we've you know, I can speak from experience correct.
00:24:47.307 -->
00:24:53.069It never pans out as you planned it always goes over time over budget.
00:24:53.394 -->
00:24:57.315It'll be other issues which will come up, but you know, but these are some things.
00:24:57.315 -->
00:25:10.584In fact, when we envisaged our M&A product, it was something like this saying we spend it's an average dividends you may have, even if you're a small-sized company, you may have almost a thousand thousand-five hundred documents.
00:25:10.584 -->
00:25:13.864Different things Compliance contracts, a mental contract.
00:25:14.045 -->
00:25:17.275Something or the other will be there your tax filings.
00:25:17.275 -->
00:25:20.071Mca filings.
00:25:20.113 -->
00:25:21.134So many things are there.
00:25:21.134 -->
00:25:28.887Wouldn't it be so easy that if you dumped everything in one place at least you would be able to get the risks?
00:25:28.968 -->
00:25:32.435out immediately, instead of having on average DD takes what?
00:25:32.476 -->
00:25:39.2786 weeks, 4 weeks, 6 weeks and during yeah, mostly four to five weeks, yeah, and during that time, investor interest may also win.
00:25:39.278 -->
00:25:41.921Yeah, and that happens all the time.
00:25:41.921 -->
00:25:50.432Yeah, by the end of the interest at the beginning will not be the same At the end there will be people who will say, oh, I want to come in, I want to pull out.
00:25:50.432 -->
00:26:02.135So actually the deal making should be as a you, as I've done in my career as a lawyer more than 250 deals and each deal is unique in its own way.
00:26:02.155 -->
00:26:16.071Correct, yes, your needs scaling, but the basis I would say about 90-95% of it is based on whatever diligence that you've taken and that should be seamless, that should be in a matter of minutes, it should not be a couple of weeks.
00:26:16.071 -->
00:26:27.509And if you think about it from a law firm's perspective as well, it's not always that you're able to get a good, knowledgeable resource to be able to lead a DD.
00:26:27.509 -->
00:26:30.433Sometimes, yes, you have a client appointment, you have a partner.
00:26:30.433 -->
00:26:35.615You hope that the partner is hands-on or you hope that the associates there are really hands-on.
00:26:35.615 -->
00:26:40.003So if that doesn't happen, there can be some process or misses and things like this.
00:26:40.003 -->
00:26:41.970Right, this happens over the length.
00:26:41.970 -->
00:26:47.905So I would think that you know, and that's how things are changing as well, correct?
00:26:47.905 -->
00:26:53.814Uh, I, you know this is, you know, um this is on the software side correct.
00:26:54.295 -->
00:26:57.338But even then, you know you're moving into an area.
00:26:57.338 -->
00:27:19.041Maybe this is something where you know, even if you look at the enterprise space or you look at the M&A space or you look at some things like these, you're also seeing a trend where hardware is becoming more relevant, correct, and you know it's depending on whom, depending on who your customers are.
00:27:19.222 -->
00:27:26.457some customers, by the very nature, also require that their data is coming from local servers.
00:27:26.457 -->
00:27:27.885It's a resurgence, right?
00:27:27.885 -->
00:27:29.529It's a reverse of the cloud.
00:27:29.529 -->
00:27:48.790And also, you know, on an AI side, unless, also on an AI side, unless you have, say, 20,000, 30,000 users, where the cost of infrastructure is also factored in in all of these things you have so many employees, users, so somewhere the cost is factored out.
00:27:48.790 -->
00:27:59.413But even if you're a smaller company, say up to 1,000 employees, people have a tendency of saying, okay, I, I want to use a cloud.
00:27:59.413 -->
00:28:09.617Cloud gets expensive after a period of time and there's a resurgence where people want to use local servers or use AI on local and things like that.
00:28:09.617 -->
00:28:11.021Any thoughts on this?
00:28:11.021 -->
00:28:14.733It will come up more and more, I think.
00:28:17.219 -->
00:28:26.298Most companies already realize that their data is gold and sharing their data with any external person people on competitors is a huge risk.
00:28:26.298 -->
00:28:43.105So local storage and then LLMs that can not be very high powered, fully functional, but they don't mind taking a little bit of a performance hit in the LLM capability in exchange for keeping things local.
00:28:43.105 -->
00:28:44.565Actually, you don performance hit in the LLM capability in exchange for keeping things local.
00:28:45.109 -->
00:28:47.618Actually, you don't need the full LLM capability as well.
00:28:49.510 -->
00:28:50.896So those things will come up.
00:28:50.896 -->
00:28:51.218I think.
00:28:53.852 -->
00:28:58.717I think it's a trend of the future, some of these privacy issues come up.
00:28:58.717 -->
00:29:09.153But just moving on to all of these things as well, on the M&A front, correct, just asking?
00:29:09.153 -->
00:29:10.490So you went to an M&A?
00:29:10.531 -->
00:29:11.615where you sold the company.
00:29:11.615 -->
00:29:12.218What was?
00:29:12.269 -->
00:29:19.094your thought process, correct, as a founder, as a founding team, saying why did you want to sell the company?
00:29:19.094 -->
00:29:20.992What are the things?
00:29:20.992 -->
00:29:25.989And you already talked about how some of these things could have helped you, especially on the functional side.
00:29:25.989 -->
00:29:31.759Just getting the DD room up and running itself becomes a huge task yourself, correct?
00:29:31.759 -->
00:29:33.383So, but what are your thoughts?
00:29:33.383 -->
00:29:45.103I think we have raised two rounds of funding before this, right, I think when you're looking at and we at the point in time the company were mostly cash flow positive, uh, scale.
00:29:46.105 -->
00:29:47.548I think there are two options in front of us.
00:29:47.548 -->
00:29:53.809One is going to financial investors and raise more money, and you have to scale or look for some strategic investors, right.
00:29:53.809 -->
00:29:55.634Both options are in front of us, right.
00:29:55.634 -->
00:30:05.561And the financial investors side I mean, uh, at that point in time in the market, I think there are not much interests which are delaying and we have some interest from strategic investors.
00:30:05.561 -->
00:30:06.805There are two, three options, right.
00:30:06.805 -->
00:30:07.875So why not?
00:30:07.875 -->
00:30:19.262But in the particular case, I think we kind of entered with partners like Skillink, where we had common investors and, from a business perspective also, what we are doing was actually completely complementary to what Skillink was doing.
00:30:20.029 -->
00:30:22.252It is not like for some they already had the capabilities.
00:30:22.252 -->
00:30:29.343They were mostly into engineering education, mechanical, civil, those areas.
00:30:29.804 -->
00:30:30.724We thought it made sense.
00:30:30.724 -->
00:30:36.054We had to come forward to the team and it made sense for us.
00:30:36.054 -->
00:30:37.916That's why we decided to go ahead with that.
00:30:37.938 -->
00:30:46.076The process took close to nine months yeah, right, I think for the time he spoke to them first.
00:30:46.135 -->
00:31:03.989Uh, till the time enough uh the money hit the bank and the transactions happened, I think almost a year, right, I think that's something which a lot of us kind of told it still has to hit the bank A lot of us kind of told, still has to hit the bank once in some cases.
00:31:03.989 -->
00:31:10.744Yeah, I think it was an experience on its own Right.
00:31:10.744 -->
00:31:15.214I think, yeah, ajay, you know you've been through acquisitions in the US, acquisitions locally here as well.
00:31:15.214 -->
00:31:17.838Did you see any difference in the way things were run?
00:31:18.784 -->
00:31:22.551I think in the US I was not part of the founding or anything.
00:31:22.551 -->
00:31:25.974I was just an employee there looking from a distance.
00:31:26.329 -->
00:31:31.910So my contribution towards those acquisitions or anything, is very non-existent really.
00:31:32.771 -->
00:31:39.977But one thing you do notice is even from the time that a potential acquisition or a merger is announced, it is disclosed.
00:31:39.977 -->
00:31:46.059There's a lot of things in the way you operate internally also that you have to be very careful about.
00:31:46.059 -->
00:31:51.615You have to log everything, make sure it's being a paper trail of anything that you want to do.
00:31:51.615 -->
00:32:03.761Any new contracts you're entering into as part of generally running the business which we had to do in Rio as well purchase new business licenses in order to meet the demands of our scaling database.
00:32:03.761 -->
00:32:11.336So now these have to also go into the new diligence Until last month.
00:32:11.355 -->
00:32:19.016It used to be your capital or operational expenses so much going up here Some standstill clauses.
00:32:20.920 -->
00:32:37.603So these things I mean the, the in many part I was not involved at all because the rest of the operations and other things is what I was focused on of the of the business itself.
00:32:37.603 -->
00:32:38.751But yeah, you have to.
00:32:38.751 -->
00:32:44.417You have to think, okay, we need, like we need X more licenses, we need to buy a new tool and things like that.
00:32:44.417 -->
00:33:03.695We have to change our terms and conditions, you know, because we want to introduce a new product or a new service that requires an update in our, for example, enrollment policies, placement policies and several other things.
00:33:03.695 -->
00:33:09.757Right, so you have to be very careful about how you do this, because then those lawyers are like huh, now this has changed.
00:33:09.936 -->
00:33:18.619This has to go to someone else, but this is interesting and I want to spend a few more minutes, if that's okay with you, on this M&A itself.
00:33:18.619 -->
00:33:26.557If you look through a typical M&AA, you would start off by signing a term sheet and hopefully an NDA.
00:33:26.557 -->
00:33:50.153From a company's perspective, reviewing the term sheet itself goes through several rounds you have to go through and you have advisors to tell you if this is something that would be okay to sign or not okay, or what are the additional things as well.
00:33:50.153 -->
00:33:55.522But you know we've reached a stage on the technology front where all of this can be run through AI.
00:33:55.522 -->
00:34:00.766You can have legal risks on transactions immediately.
00:34:00.766 -->
00:34:08.757You know highlighted there you can go through a term sheet, identify risks for yourself as a founder plus as a company.
00:34:09.280 -->
00:34:22.351I think that's one aspect where so that can be done fairly quickly, because term sheet itself will take about, say what say, four to six weeks, correct of going back and forth to be able to even reach a signing stage as well.
00:34:22.351 -->
00:34:28.853And signing is not, is not, sometimes it's paper, sometimes it's a, it's all paper for us.
00:34:28.853 -->
00:34:34.456Actually, we just went to got the stamp paper, stamp paper, but you think about it.
00:34:34.456 -->
00:34:44.340You have to go and get the stamp paper and then on top of that you have to sort of like sign it physically, create a counterpart or counterparts, depending on how many parties there are.
00:34:44.340 -->
00:34:45.592Go and sign it.
00:34:45.592 -->
00:34:46.818That takes a couple of days.
00:34:47.269 -->
00:34:50.880Resigning is one, I think, getting it from all the angel investors serial investors oh yeah.
00:34:53.293 -->
00:34:54.115I think, beyond the point.
00:34:54.115 -->
00:34:56.471Whenever I ping them, it's still they'll go.
00:34:56.471 -->
00:34:57.614Man, this guy has come back again.
00:34:57.614 -->
00:35:04.777Someone sent me a signing contract that we signed in India the other day.
00:35:04.777 -->
00:35:06.643I was like you want to sign.
00:35:06.663 -->
00:35:14.538There are like some 20 things right which you have to sort of like do, which may not be applicable in places like India, saying, what do I care about?
00:35:14.538 -->
00:35:24.512Like a US legislation on electronic signature to be able to do this and the parties are Indian and technically you have to use a digital signature to be able to do all of these things.
00:35:24.574 -->
00:35:29.684Not that a electronic signature is invalid but, at least a meeting of minds is a contract.
00:35:29.684 -->
00:35:38.097But why would you use a DocuSign system where you had actually sort of like drag drop, drag drop and I was explaining to you as well correct the longer the document?
00:35:38.097 -->
00:35:41.235It takes one day to just get DocuSign on board running?
00:35:41.235 -->
00:35:56.679I mean, you should have systems on remote signing, and that's how we built out our systems as well, in the sense that you should be able to automatically sort of like ping people you know saying if you're not signed, please sign this.
00:35:56.679 -->
00:35:58.974You know it should be available on your phone.
00:35:58.974 -->
00:36:00.641It should be accessible.
00:36:00.641 -->
00:36:01.532You should be able to sign on your phone.
00:36:01.532 -->
00:36:03.096You don't have to necessarily sign on your phone and should be accessible.
00:36:03.115 -->
00:36:05.079You should be able to sign on your phone if you don't know how to necessarily sign on your phone.
00:36:05.099 -->
00:36:06.322We had only 6-7 people on the cap table.
00:36:06.322 -->
00:36:11.438I've seen friends actually who had some 20-30 people on the cap table, I think, to just get simple things done.
00:36:11.438 -->
00:36:25.855It's like impossible to just get everyone signed and you know you would have to procure stamp paper from places like Delhi, you know, which is not always feasible.
00:36:25.855 -->
00:36:29.927And then you know there's always lost in translation.
00:36:29.927 -->
00:36:35.230Sometimes there's a weekend that comes in and you're always if the FedEx doesn't come, then it's on Monday morning.
00:36:35.329 -->
00:36:37.056You'll be like did they really sign or not?
00:36:38.603 -->
00:36:40.550so these are some of the things that actually come up.
00:36:40.550 -->
00:36:50.438So in an M&A transaction, as an, so this is just on one document, which is your term sheet, then you actually move on to the diligence.
00:36:50.438 -->
00:36:52.936Diligence is like you know there's.
00:36:52.936 -->
00:37:00.277You need a space where they keep calling for documents, more documents, all documents.
00:37:00.277 -->
00:37:02.695Even before you were born, you know all of those things.
00:37:02.695 -->
00:37:03.523Sometimes it would be like I'm just you know, but you know how documents.
00:37:03.523 -->
00:37:04.001Even before you were born, you know all of those things.
00:37:04.001 -->
00:37:06.840Sometimes it would be like I'm just you know, but you know how it is.
00:37:09.384 -->
00:37:11.197It's not just about ticking on the boxes.
00:37:11.197 -->
00:37:20.389There's legal, there's financial, there's compliance, there's product and somebody wants to one day get into your source code, your repos, all of those things.
00:37:20.389 -->
00:37:21.233So that's another.
00:37:21.233 -->
00:37:24.297There are a whole bunch of risks to be taken care of as well.
00:37:24.297 -->
00:37:48.474But say, from a practitioner standpoint, you're giving all of them, they're evaluating this and you expect a report that is not necessarily sort of like available to you, because you'll never have access to that until the very last thing, until they decide to put all of the risks in the uh final documentation correct so I mean it is until then.
00:37:48.514 -->
00:37:53.552It's like a it's very rare for an investor group to be able to sort of like share.
00:37:53.552 -->
00:38:13.817I don't think we got a report, I think we just got some of those risks at the spot of our contracts, final contracts, yeah, and and the the most difficult thing there also Ajay is unless you have someone who has the domain knowledge, like you know, or okay, in the education space, this is what's going to happen, correct?
00:38:13.936 -->
00:38:14.773or in this space.
00:38:15.195 -->
00:38:16.137Now people will.
00:38:16.137 -->
00:38:25.530For instance, you know we've dealt with customers or dealt with companies that did enterprise, that only had enterprise customers.
00:38:25.530 -->
00:38:28.663You know financial guys would say what's the ARR?
00:38:28.663 -->
00:38:46.266You know, in an enterprise setting it really is not about the ARR, it's more about the TCV, because there are so many other things that get added on the ARR is one tiny portion of license costing, but then you have support you have screening you have so many other things that come up as well.
00:38:46.670 -->
00:38:58.385So this is where there's a mismatch and I felt also, even as a founder, you need to understand what risks might come up.
00:38:58.385 -->
00:39:05.690It's always helpful it's not always that once you give a report that you go back to your teams and say correct this, so you'll have.
00:39:05.690 -->
00:39:14.101If you had that information beforehand, it would be really helpful and you could do some course correction if necessary during that six, eight week period as well.
00:39:15.070 -->
00:39:24.021And then also, if you did the course correction, your violations would always tend to go up, in the sense that you would not have to bargain on a few things as well.
00:39:24.021 -->
00:39:27.179That's how I look at it.
00:39:27.179 -->
00:39:28.494But today that's available.
00:39:28.494 -->
00:39:30.016Earlier it was not available.
00:39:30.016 -->
00:39:32.338It's how I look at things.
00:39:32.338 -->
00:39:43.277Correct, you would have to either hire your own team to do a parallel diligence, which would be an additional thing, and that is not always feasible from a money perspective, isn't it and also the cost involved in this.
00:39:43.277 -->
00:39:54.038Sometimes, when you do M&As, you not only have to pay the other blogger's fees, but you also have to foot your own blogger's fees as well, and that adds up quite a bit as well.
00:39:54.038 -->
00:39:58.201So that is another aspect, and that takes eight weeks.
00:39:58.201 -->
00:40:00.717So you've got six plus eight weeks.
00:40:00.717 -->
00:40:02.753Now You're already into the 14th week, and then you have documentation.
00:40:02.753 -->
00:40:02.858So you've got six plus eight weeks.
00:40:02.858 -->
00:40:05.170Now you're already into the 14th week, and then you have documentation.
00:40:05.170 -->
00:40:08.581People will say I want to start documentation parallelly.
00:40:08.581 -->
00:40:10.701What do you think Ajay has?
00:40:10.820 -->
00:40:12.090that ever happened in your life?
00:40:12.090 -->
00:40:13.635Never happened.
00:40:13.635 -->
00:40:17.019There's always my request.
00:40:17.369 -->
00:40:18.172It never happened.
00:40:18.172 -->
00:40:31.460Because investors have this thing saying or an instruction to their lawyers, usually saying don't start until we give the go-ahead, because the deal is never done until the money hits the bank kind of stuff correct.
00:40:31.460 -->
00:40:40.559And in these circumstances, especially in a DD, whether it is an acquisition or whether it is something else, it is not done until it is whatever it is.
00:40:40.559 -->
00:40:46.737So, assuming you went through signing, now you have this documentation, isn't it one of the most stressful things ever?
00:40:46.978 -->
00:40:49.277Yeah, First of all, you don't know a lot of stuff, right, yeah?
00:40:49.789 -->
00:40:54.476And then it just goes on and goes on, getting both parties and both lawyers to keep talking, right?
00:40:54.476 -->
00:40:58.353So I think that's one of the most stressful things I would say Just tell me about.
00:40:58.353 -->
00:41:05.739As a non-legal person, I'm half a lawyer.
00:41:05.739 -->
00:41:07.695After the period of my emails, I start looking like a lawyer.
00:41:07.695 -->
00:41:26.838He's the hell-estated, as mentioned, and I'm sure this is probably the time when you also got introduced to redlining strikeouts, all of those things who said what, and then you had difficulties figuring out who put what where correct.
00:41:26.838 -->
00:41:27.380You never did.
00:41:27.380 -->
00:41:31.329And also, your email inbox is full of legal documents, correct?
00:41:31.329 -->
00:41:31.692Yeah?
00:41:31.692 -->
00:41:34.409I think, first of all, all the lawyers don't use Gmail, right?
00:41:34.409 -->
00:41:40.282They use Outlook, right, and they don't understand the fact that the emails get actually get into the same thread, right?
00:41:40.282 -->
00:41:46.842And then they just keep sending the same email subject line and then the different, different email talk.
00:41:46.842 -->
00:41:47.505There are so many versions, right?
00:41:47.525 -->
00:41:55.335you don't even know how to manage that and most of the time whenever I go back to them, I actually convert into google doc and send it to them and they complain about it.
00:41:55.335 -->
00:42:10.168And secondly, you know they there is no standardized version and I feel sometimes that you know I'm being a lawyer correct one of one of the things that you know which is actually right is to maintain that you know.
00:42:10.168 -->
00:42:12.338Sometimes you can make out a document that you've drafted.
00:42:12.338 -->
00:42:15.632You know because you know exactly the language that you follow.
00:42:15.632 -->
00:42:16.693Sometimes I have.
00:42:16.693 -->
00:42:17.175You know.
00:42:17.175 -->
00:42:25.536There was a case once where you know, initially, like many years back, I did the first inbound swap transaction, okay, in money.
00:42:25.536 -->
00:42:28.237There was no document, literally wrote it from scratch.
00:42:28.237 -->
00:42:34.237Then, a couple of years later, a friend of mine asked me saying I'm doing, I'm like I'm looking at, I've seen this language before.
00:42:35.231 -->
00:42:38.277Then you realize your own language has been ripped off so many times.
00:42:39.173 -->
00:42:46.836Right, so you know, but apart from but you know, apart from, like a legal pride, you know, one of the things that really struck me was the lack of standardized documentation.
00:42:46.836 -->
00:42:47.358Yeah, yeah.
00:42:47.818 -->
00:42:48.099You know.
00:42:48.199 -->
00:43:04.045Wouldn't it be so much easier where you know the same classes I think we just keep discussing across multiple yeah and have we been at everybody's throat and becomes like a you know, it's like a need type personality all along.
00:43:04.045 -->
00:43:06.297We had three tier one law firms.
00:43:06.297 -->
00:43:11.494I had one, the company who acquired us had one and their lead investor had one.
00:43:11.494 -->
00:43:17.699Oh, wow, okay, right, in fact, one firm actually representing two of us, the different partners.
00:43:17.699 -->
00:43:21.536So Chinese was between them.
00:43:21.556 -->
00:43:29.028Yeah, it's too complicated and you hope that there's no conflict at the end of the road, and so these are some of the things that really come out.
00:43:29.028 -->
00:43:35.110But if you have standardized documentation, hopefully you know you will be able to get through it 90% of the time.
00:43:35.110 -->
00:43:38.719I really like the way, for example, if you get the same document of YC right.
00:43:38.760 -->
00:43:41.815Yeah, when you get a seed money of 100K, 200.
00:43:41.856 -->
00:43:42.938What documentation do you want?
00:43:42.938 -->
00:43:43.579Some simple doc.
00:43:43.579 -->
00:43:44.722Everyone just agrees and goes.
00:43:44.722 -->
00:43:50.936But I think in fact when we had initial convertible note also, I had a lawyer because we didn't.
00:43:51.512 -->
00:43:58.530I think India the safe is not safe is not recognized, we have to bring in the money within 6 months so we have to get in a lawyer.
00:43:59.114 -->
00:44:09.195I think lack of standardized documentation is definitely so much unnecessary pain and I've seen some angel funds call it something else.
00:44:09.556 -->
00:44:13.317They'll say a note when it's technically not allowed or something.
00:44:13.317 -->
00:44:28.880I've found dubious documentation saying you would actually be subscribing to a CCD or a CCPS and in reality they call it like a grant or something else which is like totally, you know, which is totally misleading, correct.
00:44:28.880 -->
00:44:41.635So lack of, I mean, even if you're doing in many situations like standard documentation, standard board resolution, standard cap table, big management becomes so difficult, correct?
00:44:41.635 -->
00:44:42.505You know you shouldn't.
00:44:42.505 -->
00:44:43.110Your automated documentation should have the capability.
00:44:43.110 -->
00:44:43.027It.
00:44:43.027 -->
00:44:43.454Cap table management becomes so difficult.
00:44:43.454 -->
00:44:49.139Your automated documentation should have the capability to automatically maintain the cap table.
00:44:49.139 -->
00:44:51.556Why shouldn't it?
00:44:51.556 -->
00:45:00.474Why would you still want to use equals bracket, open something in Excel and try and compute?
00:45:00.474 -->
00:45:08.114These are some things which I feel and we also implemented correct, especially in our products.
00:45:08.114 -->
00:45:15.157Make sure that things are sort of also streamlined, hopefully, you know, and when we did action deals correct.
00:45:15.157 -->
00:45:21.094Also, you need, like a common one of the things is you mentioned, email.
00:45:21.094 -->
00:45:24.981You know you have hundreds of emails in the inbox, correct?
00:45:25.510 -->
00:45:26.373And you don't know which one is.
00:45:26.373 -->
00:45:27.918The latest one, correct.
00:45:28.351 -->
00:45:33.050So why can't you, in this age where we can do online collaborations?
00:45:33.612 -->
00:45:35.498why can't you have an online collaboration?
00:45:36.210 -->
00:45:41.952Call everybody on the same page, mark it out, ask whatever comments you want, have teams, so on.
00:45:41.952 -->
00:45:42.773It's so simple.
00:45:42.773 -->
00:45:45.614But the thing is, I think until this guy gives the comments, the other guy will wait.
00:45:45.614 -->
00:45:55.014Yeah, yeah, then you take two days to give a comment and then in the middle there'll be some holidays and there's Deepavali and there is Christmas and you know, therefore, it goes for a class, right?
00:45:55.094 -->
00:46:04.835I mean, this is something which and I've been in situations where you know, where lawyers have changed firms in the middle of deals yeah, this goes on.
00:46:04.835 -->
00:46:05.577It's been nine months.
00:46:05.577 -->
00:46:06.797Of course, you will change.
00:46:06.797 -->
00:46:10.663So I mean, you know, life events happen, so it would be.
00:46:10.663 -->
00:46:11.704It's one of those things.
00:46:11.704 -->
00:46:15.407But, there should be like one touch collaboration.
00:46:15.427 -->
00:46:17.289What is so difficult in something like this?
00:46:17.369 -->
00:46:33.530right To do some of these and that kind of, and you will only tend to realize this because I mean, the more complexity that you bring in, the more inefficiency you're trying to hide for all of these things.
00:46:33.530 -->
00:46:35.179And I think that's one thing.
00:46:35.179 -->
00:46:40.003If you're able to do an online collab on standardized documentation, finish it off.
00:46:40.003 -->
00:46:57.605Maybe it takes a couple of days, I think, versus getting on calls where you know that you'll be like you have no idea whether the deal is going to go through or not, and people are getting angry all the time and screaming and all of these things that happen, all of these things.
00:46:57.605 -->
00:47:01.967And after that the documentation has to be finalized.
00:47:01.967 -->
00:47:13.916Assuming, in a document set, or what's called a document bible, you've got about, say, minimum, 30-35 documents, how do you sign this?
00:47:13.916 -->
00:47:16.481Assuming you've got about, say, minimum, say, 30, 35 documents, yeah, how do you sign this?
00:47:16.501 -->
00:47:20.050Assuming you've got 10 parties, would you take 10 sets of 100 page printouts?
00:47:20.050 -->
00:47:28.748You know burning bunny trees, yeah, burning, burning cartridges on on printers.
00:47:28.748 -->
00:47:33.880You know burning enough money on courier charges back and forth.
00:47:33.880 -->
00:47:35.458What would you do?
00:47:35.458 -->
00:47:36.864What's the simpler way?
00:47:36.864 -->
00:47:40.242No, I think that's how it happens.
00:47:40.242 -->
00:47:43.599Fortunately, we have this done-so's of the world.
00:47:43.599 -->
00:47:45.626Keep sending across the city and send it couriers.
00:47:45.626 -->
00:47:48.643It's definitely a logistical nightmare.
00:47:49.143 -->
00:47:49.385It is.
00:47:49.385 -->
00:48:04.980And if you really looked at how deals are done you know so nobody has done like the carbon analysis or carbon emission analysis or deals or legalities being done it would be off the scale, right, and things like that.
00:48:05.081 -->
00:48:08.295I mean, you mentioned sending Dunzo, I think how many times you know?
00:48:08.295 -->
00:48:12.182At least DUNZO probably makes a lot of money through these things.
00:48:12.182 -->
00:48:19.206Yeah, or even Sleekly for that, or Porter, but yeah, now it's Porter and then it's something else.
00:48:19.206 -->
00:48:20.610But you know, all credit to them.
00:48:20.610 -->
00:48:21.974At least some of these things get done.
00:48:21.974 -->
00:48:23.766We need a carbon tax on the energy.
00:48:24.295 -->
00:48:25.677You need to think about it.
00:48:25.677 -->
00:48:27.481No, but, wouldn't you?
00:48:27.481 -->
00:48:31.349Actually, you know, if you were a software, it would be nice if we just had a software.
00:48:32.277 -->
00:48:33.523Someone can just look at what it is.
00:48:33.523 -->
00:48:36.936Look at the critical parts, I think.
00:48:36.936 -->
00:48:38.543See, I'm doing it for nine months full time.
00:48:38.543 -->
00:48:42.302Imagine, let's say, an investor who has one person's stake in it.
00:48:42.302 -->
00:48:44.367He has to still sign Sign all of these.
00:48:44.367 -->
00:48:45.188What is he going to review?
00:48:45.188 -->
00:48:46.478What is the summary for?
00:48:46.518 -->
00:48:46.599him.
00:48:46.599 -->
00:48:48.396What is his story about?
00:48:49.237 -->
00:48:50.518But you missed the best part.
00:48:50.518 -->
00:48:52.440Yeah, you know, you know people in India.
00:48:52.440 -->
00:48:55.882If you send it to 10 parties, has the other person signed?
00:48:55.882 -->
00:49:00.407Has he signed?
00:49:00.407 -->
00:49:03.010I'm not going to sign until this person has signed, isn't it?
00:49:03.090 -->
00:49:06.297I think we in fact have an option for some of the investors.
00:49:06.317 -->
00:49:06.998Do this or that.
00:49:06.998 -->
00:49:09.445Okay, what is the other person doing?
00:49:09.445 -->
00:49:13.036It's also interesting, yeah, so it's one of those.
00:49:13.036 -->
00:49:18.619I think, especially if you've got a lot of angels right, this, this becomes like a little difficult to manage?
00:49:18.659 -->
00:49:24.539certainly we have some people who we know us right, it becomes uh much easier otherwise I can imagine.
00:49:24.539 -->
00:49:28.594Yeah otherwise for our own calls correct, saying oh, this did not happen.
00:49:28.594 -->
00:49:36.862You know it becomes very stressful, but i'm'm just saying, look, we went through like how typically like you would use automation and AI for something like this, not for anything.
00:49:37.242 -->
00:49:45.583We spent nine months, assuming you had, say, nine weeks, or if you could complete the whole thing in, say, a week or two weeks.
00:49:45.583 -->
00:49:46.425Why nine weeks?
00:49:46.425 -->
00:49:47.996You know, wouldn't?
00:49:47.996 -->
00:49:48.677I mean?
00:49:48.677 -->
00:49:50.722I would think that it would be a lot more easier.
00:49:50.722 -->
00:49:52.847It can happen in 9 weeks.
00:49:52.847 -->
00:49:53.768Significant value.
00:49:54.056 -->
00:49:55.161Significant value, isn't it?
00:49:55.556 -->
00:50:01.956I think, ajay missed the part, even though I was mostly running these operations along with another founding member.
00:50:01.956 -->
00:50:11.143But I think the critical part is people like Ajay, who is leading the team, have to keep the entire team motivated.
00:50:11.143 -->
00:50:13.961This information has not gone out.
00:50:13.961 -->
00:50:17.400I think the more time it takes, it's more painful.
00:50:17.400 -->
00:50:21.641You're running a business, you're managing cash flow, you're keeping the people together.
00:50:21.641 -->
00:50:23.159It's very, very critical.
00:50:23.159 -->
00:50:24.918So this time plays a critical role.
00:50:24.918 -->
00:50:28.083The more it extends, it becomes even more tougher for us to keep things together.
00:50:28.275 -->
00:50:46.858Absolutely I think Ajay, and every company has one or two people like that especially in the operations sector Because as a founder, I can imagine there'd be days where you don't even want to take calls or you don't want to do anything, you don't want to talk to anybody.
00:50:46.878 -->
00:50:51.307You just want to be left alone.
00:50:51.307 -->
00:50:58.702But I think you know all credit to the teams that actually put things or hold things together, isn't it?
00:50:58.702 -->
00:51:16.545That's why the time, if the time can be kind of compressed by even let's say one third of it there's so much value and also now you know one of the things that especially in doing this, whether you do it in the US or India, you'll have some obscure FEMA.
00:51:16.545 -->
00:51:21.003Especially we have foreign investors, yeah, yeah, and they'll have some obscure FEMA.
00:51:21.175 -->
00:51:22.425I think that made it very complicated for us.
00:51:22.594 -->
00:51:24.463Actually, the deal became very complicated just because of that.
00:51:24.463 -->
00:51:38.527And if you look at, you know the legal is one portion and this is another thing which we also discussed previously before the podcast as well, saying the banking and the regulations and things like that make it very, very complicated.
00:51:38.527 -->
00:51:50.603There are KYC procedures, so you get into some structures to avoid it or probably to make it less complicated, but the whole deal becomes very complicated because of that and it is nothing less complicated, but the whole deal becomes very complicated because of that and it is nothing.
00:51:50.603 -->
00:51:52.838I mean it was not as complicated as before.
00:51:52.838 -->
00:51:58.978But just to imagine, if you're getting money from outside India, you know and it comes to your account, you have to.
00:51:58.978 -->
00:52:03.342Basically you'll get a notification from a bank saying fill up all of these things.
00:52:03.342 -->
00:52:26.864You have to do a CETRS or the EQUA, you have to do firms filings on RBI, you have to do all of these things as well, which adds from personal experience In some cases it's added like six or simply receiving some FDI now will result in six or seven months of delays because of all of the compliance issues that come in.
00:52:27.416 -->
00:52:34.382Then the local bank has to, even before releasing monies, it'll go back and check why they need to use a SWIFT.
00:52:34.382 -->
00:52:36.822Correct, swift has not been.
00:52:36.822 -->
00:52:39.420I mean, none of the international banks actually use SWIFT.
00:52:39.420 -->
00:52:45.360Think about it right In the sense that to communicate messages right, but Indian banks do that.
00:52:45.360 -->
00:52:50.898Is a team's call from an authorized person going to change anything?
00:52:50.898 -->
00:52:52.543Why does it only have to come through Swift?
00:52:52.543 -->
00:52:54.320Think about some of these things.
00:52:54.320 -->
00:53:00.648There are some procedural inefficiencies which you have no idea because there's no.
00:53:00.648 -->
00:53:03.824I mean you could have a decentralized protocol to do all of these things.
00:53:03.824 -->
00:53:16.867If you have a central bank that does a CBDC, how difficult is it for it to implement a level one protocol you know on blockchain, to be able to sort of like exchange messages with another bank?
00:53:17.514 -->
00:53:26.603You know, so these are some of the things that actually come you know external but which you cannot control.
00:53:27.215 -->
00:53:41.083But do you see, now, with AI and automation correct and also all of these things coming up and there's an actual need because if you're looking at becoming like the third largest or second largest economy in the next five to six years.
00:53:41.485 -->
00:53:49.474We're talking of India in this case now you are going to face all these structural issues which the economy and I mean nobody actually planned for, the economy, and I mean nobody actually planned for.
00:53:49.474 -->
00:53:59.661These are all 1970 type issues where you had to still send letters and not where, and SWIFT itself as a protocol probably is 50-60 years old now.
00:53:59.661 -->
00:54:02.432So what do you think, ajay?
00:54:03.284 -->
00:54:03.733It's inevitable.
00:54:03.733 -->
00:54:06.184I think the people are ready.
00:54:06.184 -->
00:54:11.438The question is, when will government and regulations and other things follow?
00:54:11.438 -->
00:54:23.943It has happened in the case of, for example, ubi very convenient government backed initiative making it very convenient to do money kind of transactions.
00:54:23.943 -->
00:54:29.025Now formalization of legal contracts and other things.
00:54:30.409 -->
00:54:37.065First of all, it requires simplification, if you have a very complicated set of regulations and rules that you can't follow.
00:54:37.666 -->
00:54:39.920formalizing those things is bizarre.
00:54:39.920 -->
00:54:45.407So formalization will hopefully start or lead to simplification.
00:54:45.407 -->
00:54:57.643They have to happen in parallel and AI is there, you know it's there to make enforcement of these regulations much easier at scale, country scale.
00:54:57.643 -->
00:55:02.934You can do country scale enforcement of regulation which previously required manual work and other things right.
00:55:02.934 -->
00:55:06.005So I think, all three will have to go in parallel.
00:55:06.005 -->
00:55:12.733You have to simplify, you have to formalize, and then AI comes in as an enabler and an enforcer of all of these things.
00:55:13.275 -->
00:55:31.168I think one of the things also, for instance, in the talk of payments and say even cross-border payments as an example and an aside to all of these things if you are operating as a business in India, or any individual for that matter, correct, this gets very complicated very, very soon.
00:55:31.168 -->
00:55:38.452Imagine you know you're a business where you receive, say, get paid, say 50 invoices a month.
00:55:38.452 -->
00:55:42.340I'm just throwing a number from customers and then what happens?
00:55:42.641 -->
00:56:14.275every 50, every, there are 50 FIRCs, correct, foreign Invoice Remittance Certificates for every remittance that your bank account receives, but before the FIRC is issued, but you can't have a foreign account and get it transferred once you can, you can potentially, but still you know you need a foreign entity for that, you need a foreign entity for that and you know the legal way of doing things at least is to make sure, at least in India, that you have a subsidiary of a company and you need to have a track record for that ODI scheme.
00:56:14.275 -->
00:56:17.003And it gets complicated very, very, very fast.
00:56:17.003 -->
00:56:24.322Or you can just choose to ignore but then face at some point the income tax department knocking on your door.
00:56:24.503 -->
00:56:27.567So these are two options that are available.
00:56:28.135 -->
00:56:35.557But imagine, even on the, you have a system now which is if you have AI, for that matter, you think about it.
00:56:35.557 -->
00:56:41.701You know there's a patent recognition system which basically says that, yes, now what happens manually?
00:56:41.701 -->
00:56:45.144You receive a remittance in your bank account, as an example.
00:56:45.144 -->
00:56:52.056Now the banker will send you one thing please categorize the transaction Only.
00:56:52.056 -->
00:56:53.248Then I'm going to sort of release this.
00:56:53.269 -->
00:57:03.340Now, this, you can have AI, sort of like, look at all of these things and do this, but the banker still, for instance, I've not reached this.
00:57:03.340 -->
00:57:07.983They need some proof saying that, oh, we have said that this is a legitimate remittance.
00:57:07.983 -->
00:57:11.181And I mean, if I'm using a bank for all of this, it's great.
00:57:11.181 -->
00:57:15.545But you know there are other means where you can do all of these things.
00:57:15.545 -->
00:57:22.385I won't get into the details, but you can do all of these things.
00:57:22.385 -->
00:57:27.057You can automate so many of these things as well, and you can.
00:57:27.057 -->
00:57:29.445You can a lot of opportunities in all of these things.
00:57:29.445 -->
00:57:38.048You know you are, you are founders who have spent time in training and all of these things.
00:57:38.918 -->
00:57:46.027Do all these things interest you, because there is a lot of opportunity as well to overcome all these things.
00:57:46.027 -->
00:57:49.005Some structural changes as the economy takes shape.
00:57:49.005 -->
00:57:50.853What do you guys think?
00:57:50.853 -->
00:57:53.416I think definitely there's a lot of opportunity, right.
00:57:53.416 -->
00:57:59.184I think in the overall, I think this is like we have got some raw digitization done.
00:57:59.184 -->
00:58:08.753We have not really optimized a lot of those processes, right, and I think definitely there's a lot of opportunity across, not just in banking, right, but overall across the services.
00:58:08.753 -->
00:58:14.945As an example, but sorry, uh, but why banking comes up, is any scale business?
00:58:14.945 -->
00:58:18.818The first hurdles that you hit start hitting is, the bank isn't.
00:58:18.858 -->
00:58:21.606It is this something I mean regardless of.
00:58:21.606 -->
00:58:23.492I mean we as founders.
00:58:23.492 -->
00:58:25.476You know, you know the first hurdle.
00:58:25.476 -->
00:58:26.880You can talk to any founder.
00:58:26.880 -->
00:58:27.983This would be the first hurdle.
00:58:27.983 -->
00:58:29.146That is not correct.
00:58:29.146 -->
00:58:31.119So I think that's right.
00:58:31.119 -->
00:58:34.295You're just having a credit card, corporate credit card, with them and then managing that is a pain.
00:58:35.137 -->
00:58:45.347Oh yeah, totally you know, you'll have to hit certain turnover levels to be even accessing corporate credit cards, but that's still tied to your Aadhaar number, your personal guarantee.
00:58:45.347 -->
00:58:50.746So I'm like there's no point in having a corporate credit card because that is once again tied to something else.
00:58:50.746 -->
00:58:58.644So I think small things like bill discounting make it a big thing.
00:58:59.795 -->
00:59:01.543Very recently, even government schemes.
00:59:01.543 -->
00:59:13.001One of the banks told us that as an MSME you can get lines of credit, uncollateralized lines of credit, which never happens in India.
00:59:13.001 -->
00:59:17.396But you know that's apparently the government guarantees all of these things as well.
00:59:17.396 -->
00:59:19.275You go and ask the bank for all this.
00:59:19.275 -->
00:59:21.777They say you must not have any loan anywhere else.
00:59:21.777 -->
00:59:22.681Correct, like?
00:59:22.681 -->
00:59:29.416Show me any credit report of any person in India which does, who has no loan somewhere, whatever it is right.
00:59:29.838 -->
00:59:31.202So the banks themselves laughed.
00:59:31.454 -->
00:59:38.342But you know, at the end of the day, I think this is one of the significant places where, you know, some of these things can change as well.
00:59:38.342 -->
00:59:42.003But you know, we've gone through all of these things.
00:59:42.003 -->
00:59:46.219But, in your next.
00:59:46.219 -->
00:59:50.900I wish you the very best on your journey as well on the next things that you're planning.
00:59:50.900 -->
01:00:00.960Very exciting stuff, Correct, and hopefully all these things that we discussed today also make it a reality in some ways, All of these- things as well?
01:00:00.981 -->
01:00:01.643Certainly hope so.
01:00:01.643 -->
01:00:15.726I think if you just lower the bar and make it easier for people to start businesses, Many people just get spooked by how complicated even the initial setup of the business is, the contracts receiving money.
01:00:15.726 -->
01:00:17.753You go to investors with an idea.
01:00:18.757 -->
01:00:25.893First question if they like your idea and they want to give you money first question they'll ask you are you incorporated Outside India?
01:00:25.893 -->
01:00:34.898A lot of questions like that, and then of course, they prefer outside India because they have their system set up there.
01:00:35.280 -->
01:00:42.003But now you see many large companies coming back to India like rehoming in some sense, like Groww is coming back.
01:00:42.003 -->
01:00:54.380We had a huge amount to move their taxes Because it's more meaningful to go IPO in the Indian market rather than the US market because of the market gap that they have.
01:00:54.380 -->
01:00:55.842So these things will come.
01:00:55.842 -->
01:01:04.592I mean, if you just make it easy for people to be compliant, start something, be compliant and run a healthy, the hygiene part of the business.
01:01:08.262 -->
01:01:10.527I think it will encourage a lot more people.
01:01:10.527 -->
01:01:18.704I totally agree, I think, that portion it gets complicated really fast Very, very early in the life of the company as well.
01:01:18.704 -->
01:01:25.039I think spot on on some of these things as well about how some of these things can be streamlined.
01:01:25.039 -->
01:01:31.465Also, I think there is especially on opening outside India there are.
01:01:31.465 -->
01:01:33.155I think there is especially on sort of like opening outside India there are.
01:01:33.155 -->
01:01:37.726I mean from a legal and also as a founder's perspective.
01:01:37.726 -->
01:01:50.605Many founders don't realize how complicated things can get for them, and it's not just from a tax perspective but from a compliance and a reporting perspective.
01:01:50.605 -->
01:01:56.659And what happens one day if you cannot travel to that country, then what happens?
01:01:56.659 -->
01:01:59.295What do you do with all of these things?
01:01:59.295 -->
01:02:15.360And also you also lose some home market advantage which is always maybe you have, at least you have for e-procurement.
01:02:15.360 -->
01:02:16.581Things are moving towards that.
01:02:16.581 -->
01:02:24.425You know there are some advantages as an economy grows as well, but I think it's not but I see a lot of these things as well.
01:02:24.445 -->
01:02:28.067If you see my Instagram feed and it's profiled you as a founder.
01:02:28.067 -->
01:02:33.210Not only this open in Dubai, open in America, open here, open there.
01:02:33.210 -->
01:02:36.192Correct, this is exactly how it is going to be correct.
01:02:36.192 -->
01:02:38.413I would you know.
01:02:38.413 -->
01:02:45.143I would think that there's a lot more nuances than just deciding to sort of like set up shop there.
01:02:45.143 -->
01:02:59.710And probably the decision I don't know, I mean just I mean, would you, and probably that decision I don't know, ratna, I mean, would you wait for a company to grow to a certain stage before you obviously take the decision to sort of internationalize the company, or would you do it on day one?
01:03:00.210 -->
01:03:01.652Narendra.
01:03:04.614 -->
01:03:11.028See, I think for us the market was India, so it was a straightforward decision for us, I think, the founders who are actually having their primary market outside India, I think.
01:03:11.028 -->
01:03:15.460My understanding is, I think the decision has to be taken a lot in the beginning itself.
01:03:15.460 -->
01:03:16.693Right, because it makes a difference.
01:03:16.693 -->
01:03:19.420Let's say, if you want to get customers, right from day one from the US.
01:03:19.936 -->
01:03:23.235It makes sense for you to have a company in the US right but.
01:03:23.235 -->
01:03:27.322I think, for us, it was straightforward because we have primary businesses in India.
01:03:28.797 -->
01:03:36.422So an aside and a joke on that, I asked someone very senior in NASSCOM very early on, saying so what should we do?
01:03:36.422 -->
01:03:38.601Should we look at this or a foreign structure?
01:03:38.601 -->
01:03:39.739What are your thoughts on this?
01:03:39.739 -->
01:03:41.661He said please stick to home ground.
01:03:41.661 -->
01:03:52.561If you go to the US correct you will have your neighbor and everyone else competing against you on the same product as well, so against you on the same product as mine.
01:03:52.561 -->
01:03:57.014So I think he was like first achieve some PMF before you even bother all of these things.
01:03:57.014 -->
01:03:59.893But I mean well meaning even NASCOM things like this.
01:04:00.318 -->
01:04:04.121So just saying but thanks so much on this, thank you.
01:04:06.025 -->
01:04:06.974And pleasure having you guys on board.
01:04:06.974 -->
01:04:10.007Thank you for spending time with us.
01:04:10.007 -->
01:04:10.387Thank you.
