Full Transcript

Utsav Somani: Alright, listeners. It's a Friday show. Today, we've got an action packed lineup for you, all different industries. And I'm starting off with Sridhar, who's the co-founder and chairman of PurpleTalk. Sridhar, welcome to TON.

Sridhar Muppidi (PurpleTalk): Oh, thank you, sir.

Utsav Somani: Ready to get the Friday special started with us?

Sridhar Muppidi (PurpleTalk): Definitely. Looking forward to it.

Utsav Somani: Alright. So you started two thousand eight, May, PurpleTalk. Like, you've done many things along the way. You've launched India's first game on the App Store. You built an ad network as well. You did quick commerce also, and now you're running a multiple suite of products under one company called PurpleTalk. So do you wanna do a quick two minute recap of what this journey is, what are the learnings, have been?

Sridhar Muppidi (PurpleTalk): Definitely. I started my career with video gaming way back in '99. You know, starting my first company during the .com boom. Sold it for quite a bit of money back then. Then, I got into telecom because that was, you know, we were working on the original SIP protocol and others when we backed in 2000. So was in telecom building those solutions for almost, like, five, six years. And then when Apple App Store was about to be announced, you know, we were already playing around with the Apple SDKs and, you know, others in the jailbroken phone. So me and bunch of other developers I was working with, we just decided to start PurpleTalk. So and then wanted to do something simple. So games made sense. So we built we we were the first company in India to build a game. We also launched a small ad network back then, which, you know, which at that time did quite well. You know, and then, you know, along the way, we started helping folks we know. So, we kind of became a services firm or a consulting firm, building products for other people. So we spun off a group called [x]cube LABS as part of that. I personally started focusing more on video games, and we spun off a group called YesGnome. We didn't really do much in quick commerce. There was an opportunity in quick commerce, but, we decided to focus on the mom and pop, Kirana stores and try to help them be more efficient. So we launched this company called Nukkad Shops, which is like a commerce solution, takes care of all your technology needs of a small mom and pop store. Right? And, we for past, few years, we've been heavily invested on, trying to see where the opportunities are, you know, because we we all enter this, you know, this beautiful new world of AI and, you know, what it can do for us. So, yeah, we're trying to see how we can stay relevant, I guess.

Utsav Somani: And, I mean, let's talk about the two new products, Mina and Ello. What are they? So

Sridhar Muppidi (PurpleTalk): there's something new happening right now, which is basically most of future or software consumption is gonna be from AI agents. So almost every SaaS product, every business which is out there, where humans enter data, humans retrieve data, you know, make decisions based on that, most of that is gonna be outsourced over to agents. Right? The data entry part, data retrieval part, all you are doing is you're just talking to your your companion, your AI companion, which could be your ChatGPT, your Claude, your whatever, and they are supposed to get you all that information. So, we have this hypothesis that there are gonna be a lot of tools which would help you both in data entry as well as retrieval. And the biggest part is voice. Right? Because that you know, most of our, you know, human interaction, we are used to Right. Talking to people. Right? We we want something. We just talk.

Utsav Somani: Right? But, I mean, just diverting a little bit. Like, Naval recently posted a tweet where he said that chat is gonna be your primary interface. Do you think voice or chat?

Sridhar Muppidi (PurpleTalk): Which one's gonna bring between I think it's interchangeable. Right? So, most of like, we have, like, five, six hundred programmers. Right? Significant number of them just talk voice, versus chat. Right? I mean, of course, in a larger room when people are there, you would rather type. But if you are alone, you are only talking because you there there there's no way, you can type as fast as the way you can think and speak. So significant number of people who are working remote as or, you know, on all by themselves are gonna use voice. If you are in a crowded place, maybe you would type, but that's not but when what he means most probably by that comment is you would interact. You won't press buttons. You won't scroll through reports that anymore. You would just ask, you know, either through text or voice. Right. So that's what he meant, but not, you know. So so we we looked at that area, and we said, hey. That's something we wanted to invest on. There's a lot of problems we can solve there because the reality of voice is, you know, there's a bunch of technologies you use to make it work because we're using so many different technologies. Emotion gets lost, voice gets lost, you know, a lot of things. So once we figure that out, how do we make voice agents more efficient? So that's where Ello comes in. It's a beautiful technology we built with. A lot of people use it as a companion tech in their apps and, you know, products more than for telecalling and others.

Utsav Somani: And we'll come to Meena later, but if you were to get more technical about Ello, like, I think there's a lot of noise in the voice AI space. Right? I mean, there's, of course, the global giants like, ElevenLabs, then we've got Sarvamm with India, then we've got many others who are doing, like, different verticals. It sort of become verticalized that somebody's going out to BFSI industry very deep and Mhmm. Very use case heavy. Where do you see how did you build about, your data, which is, I'm guessing, a big mode for you, and you've done 12 languages. And how did you think about how your technology is way ahead or at least on par in terms of benchmarks with other players out there?

Sridhar Muppidi (PurpleTalk): Two, so there are two areas people seem to be focusing on when when it comes to voice. The low hanging fruit tends to be, can you automate your call center work? Right? Your your telecalling or marketing work, the inbound support work. And voice technology is amazing for that. Right? Yes. There are some problems to solve, but that that's pretty amazing. That's not what we are focused on. What we are focused on is, automating conversation with your software. Right? Can can your software, become a companion for you? Can can it become a real assistant for you? So, think of it, you know, if you have an app, where you spent a lot of time bringing that user to the app and you're onboarding them, you would rather have a voice companion along the way helping them through onboarding versus expecting them to figure everything out. So, what Ello does is it can tell you exactly where the friction points are in the onboarding. It is obviously gonna help the user get to the endpoint of a purchase or endpoint of whatever, but it will also say, in the screen, they seem to be struggling or these are the kind of questions they are asking. Right? So it's gonna give you a lot of data which you don't have access to right now. As a product developer, you get that data. The second thing is it's also solving the customer problem of they don't trust you enough. Right? They have this question and it can answer it right in the by the way, we have these these things we follow, or you can come to our office and meet us. Right? Things like that. So we help with that. Right? So, companion technology is our focus with Ello.

Utsav Somani: And you also said that, I mean, you're spending over $1,000,000 on tokens also. And we're being heavily subsidized by venture dollars and just in race of, acceleration for adoption. But, what do you think? Where does this world go when we are all talking about AI as a technology, AI as something that delivers substance over hype?

Sridhar Muppidi (PurpleTalk): So I think, that argument is gone. Right? I mean, the high part I mean, you know, we're all AI player.

Utsav Somani: It's here to stay.

Sridhar Muppidi (PurpleTalk): Obviously. We we can't live without AI. We all know that part now. So, most of our, you know, I think significant part of what we spend is, you know, is on, tokens to build things more than, you know yes. Some of the products are, you know, the companion tech and all the other things. But to build things and we while it's not the right word to use, we we are almost forcing out, you know, developers to token max as much as possible because we believe smaller models eventually are gonna be powerful enough, you know, maybe in a year down the line, eighteen months down the line, where token costs again are gonna drop. Right? So, you know, so what we can do right now, the your biggest models, what they can do now in one year down the line, the smaller models at a fractional price can do the same thing. Right? So because of that, we don't want to be in a position or regret the fact that we haven't spent that extra million now to learn and be ahead of the curve. Right? So that's how we look at it while we know it's you know, there is probably half of half of the tokens we seem to be just wasting, we would rather waste and learn versus miss out that opportunity. So that that's the thinking right now.

Utsav Somani: 1,000,000 is a lot to pay for education. So where is that 1,000,000 going, like, approximately?

Sridhar Muppidi (PurpleTalk): So about about 500,000 is going into, you know, Claude, OpenAI and Google and others, you know, using their tools. I think other half, goes into the models, we use in some of our software.

Utsav Somani: And in terms of voice, like, I think latency and all of these things matter because over a telecom and data pipes, voice does tend to travel slightly slower than, say, other packet. So, I mean, how do you define I mean, or how do you worry about latency while building voice products? Because there is a lot that's happening where people have tried to reduce the latency, do some, I mean, magic on device versus, using the cloud. So how does that actually work when you, do a companion app?

Sridhar Muppidi (PurpleTalk): So, when you are doing, sort of, like, software based voice versus, telephone based voice where you you don't see anything. Latency does matter. Like, human ear detects latency if it is more than four hundred milliseconds. So, you know, that's a bigger problem. But thanks to face, you know, WhatsApp and others, people are used to about a second delay now. Right? So right now, technologies are able to do one to 1.2 seconds, delay and unless unless you need frontier level intelligence. Right? So most of the phone conversations don't need frontier level in intelligence unless you're, like, asking analyze this for me, then that's when you are gonna go to, you know, you use bigger models. And that's when it's gonna add, like, a couple of seconds more delay. But otherwise or if you're using tools to fetch the information. So where people tend to use fillers. Right? Give me a second or things like that. But as long as users know they're talking to AI, they're willing to wait. You know? If if you are deceiving them trying to say that, you know, this is human, then yeah. I mean, obviously, they would find flaws with it. If they feel like they're talking to AI, they're okay then, you know, the delay is fine because you are talking to a machine at the end of the day. So I don't think that that should be anybody's priority right now in terms of trying to fix the priority should be how do we make your voice agents as capable as human agents if you are using them in support or others where the empathy, the conversation, knowing, the pain point, what exactly need to and collecting a lot of data. Because at the end of the day, what humans cannot do, which voice agents can do, is be consistent. Right? So over long term, you know, AI will do a better job. It's just that we we just need to keep iterating on it.

Utsav Somani: And what about the meeting assistant that you've just launched?

Sridhar Muppidi (PurpleTalk): So it's a meeting teammate. Okay. So traditional meeting assistants are note takers. Right? If they they take notes and they're done. Right? Mina is a teammate which can participate in the meeting. It can open a canvas and, you you you can set the agenda. You can see what people are talking. So instead of you getting a summary after the meeting, you are seeing what's going on in the meeting on the screen. Right? Action items, who's been assigned what. You can change it. By the way, I didn't mean Mahesh. I meant Raghu should be doing this task. Right? Different canvases based on what can a meeting like

Utsav Somani: an active agent, and do they have access to company data as well?

Sridhar Muppidi (PurpleTalk): Exactly. So, it tracks what's going on with the meeting. Again, type of meetings, if it is a scrum meeting versus a sales meeting versus a generic meeting, based on that, Canvas is different. It has integrations with everything from Jira to Salesforce to everything. So, all the activity within the meeting obviously is dynamically going into this fetching that context is coming in. Unlike, passive assistance, Mina can participate in the meeting where you can basically say, hey, Mina. Summarize this. Or, hey, Mina. Set a meeting for us next week, and it does it instantly. Right? So it can talk to you and you, you know, you can ask it questions like, what did, you know, what did what date we did we commit? Or you can basically ask it to make a presentation or something like, can you show them our use case for this agriculture or whatever? It's gonna open that presentation and make the presentation for you if you want. It's also, has a executive assistant where it's talking to you offline on Slack and other places where, say, you know, in this case, you know, it could be nudging me saying, Sridhar, let him talk. You've been talking too long. Right? So, yeah. So it's, it it can be a coach. You know? So it's basically very, it it's a companion which is gonna help you do your job better.

Utsav Somani: And, I mean so I think the world of AI agents acting as coworkers is upon us because Slack is making a big push where they want their platform. And even Jack of, Twitter announced that I think he's building an open source platform where agents and humans can coexist as employees.

Sridhar Muppidi (PurpleTalk): It's called Buzz. Yeah.

Utsav Somani: It's called Buzz, I think. And, then even Slack, I think, recently announced that Salesforce, that they wanna call I mean, it a common platform where agents have their own existing workflows, and they can do a lot more of these things. So how far out that world, is from us right now where agents are actually doing, I mean, full departments worth of work?

Sridhar Muppidi (PurpleTalk): So, where we are, that's a good question. So this is like something I keep thinking for myself, which is basically, I've been internally or externally been talking about what we need is high agency. We will survive this upcoming whatever, you know. But now I don't believe that anymore. Agents are high agency themselves. So, you know, so what do you bring to the table? Yes. Taste judgment and all that. But, so the world is upon us. We've been, like, our Slack groups, have, you know, multiple agents, as well as humans. We we even have, humans reporting to agents, where, you know, the agent's job is to be a program manager, manages who's doing what. We have an agent which is a scrum master, maintains Jira as well as thing. We have an agent which primarily is a tester. So, the program manager agent, you know, basically gives task to humans, you know, submits PRs, assign task to test the agent to see, hey. This has been submitted. Let give me a report if, you know, if you find any bugs. So we are already in that world. We are already doing that.

Utsav Somani: But do you trust it enough? I think the hallucinations and all that error correction, I think, will, I mean, improve significantly as we go along. A long

Sridhar Muppidi (PurpleTalk): There are multiple ways to build it. Right? So the way we've done is almost all of our agents have a supervisor on top of it. Right? So while the agents which are talking to us are your open claws, Hermes, and those guys, right, in your Slack Human

Utsav Somani: and the local.

Sridhar Muppidi (PurpleTalk): And not human. The the supervisor results an agent. In the one Well, one them is your clawed agents and the others, which are basically every hour checking what they are doing in this different projects and if there are course correction. And they talk to the human offline to see, hey. This is what I found. It it is showing a tunneled direction. It does not giving enough time to others. You know, things like that, and it course corrects, and and creates a a sort of, like, a a checklist on how to do its job better, and you're always compacting the memory. You know, just Excel learning will

Utsav Somani: require the loops, I guess. Yeah.

Sridhar Muppidi (PurpleTalk): Yeah. It's it's not it's no way perfect, but we all see enough value in it that we don't see going back to just working with humans, anymore.

Utsav Somani: And, Sridhar, maybe as a final closing question, I wanna get some operator insights from you because you started as a product company, then you went to services, and now you're building products again. What do you have to learn and unlearn, two things that you can share with us as you go along from a service to product mindset?

Sridhar Muppidi (PurpleTalk): That's a good question. Speed, I think. Right? At the end of the day, GTM is not your problem in a services company. GTM is everything in a product company. Right? So, we always had great product instincts, but, unless you can figure out growth, you know you know, you you could have a great product, which will become stale very soon, if nobody uses it. Right? So, the it's not learning. It's just, something we always knew. It's just that we need to work on it. Right? Getting better at, growth. Awesome. A services company, that's what they need to know. Right? It's, closing in a customer one customer to get a million dollar business is different from, you know, I need,

Utsav Somani: you know customers to close a million maybe.

Sridhar Muppidi (PurpleTalk): Yeah. That that's that's a different ballgame. Right? It's,

Utsav Somani: So using those customers, I think, would mean I mean, most of them would be self serve or product. You can, of course, scale the gross margins, of course, significantly, go up also. I mean, many things. Yeah.

Sridhar Muppidi (PurpleTalk): It tends to be a problem, you know, for engineers. Right? As an engineer, you're always thinking, how do I do automation? How do I you know, I think the focus should be, you know, talk to your customers as much as possible, help them as much as possible, be in the loop. Eventually, you can automate. Right? Or automate cannot be the first, solution because then then you're detached from the customer. So the solution is always talk to the customer. The the first 100 customers, you should be talking to every one of them. And, eventually, yes, you know, once you once you help the product market fit, you know, you could do whatever you want.

Utsav Somani: Alright, Sridhar. Thank you so much for dialing in from Hyderabad. Wishing you the very best. Thank you. Alright, listeners. We're moving on to our second guest. We've got Arth from Tirex Chargers. Arth, welcome to the show.

Arth Patel (Tirex Chargers): Yeah. Hi, Artur. Thank you for having me. Hi, everyone.

Utsav Somani: Our pleasure. Our pleasure. Thanks for being a part of offline. I think we'll do an EV charger one zero one first. What is an AC charger? What is a DC charger?

Arth Patel (Tirex Chargers): So AC is something that typically you install where you have a lot of time, like your house, your office, mall, maybe. So anywhere you spend more than, let's say, two hours or three hours, that's where you put an AC charger. DC is something that's fast. Your highways, your petrol stations, or, like, highway food courts, where you have to spend thirty minutes, twenty minutes, or in order to charge heavy vehicles, which is like your trucks, buses, your construction equipment, where your batteries get much larger than your cars, that's where you put a DC charger.

Utsav Somani: Alright. And, what is Tirex doing to improve this, industry?

Arth Patel (Tirex Chargers): So, we are a seven year old company. Today, we are number one, I would say, when it comes to heavy vehicle charging in India. Be it buses, trucks, or construction equipment, We have more than 40% market share. So out of the 15,000 electric buses in India, oddly, we charge more than 5,000 on a daily basis. Delhi, Bangalore, everywhere, Mumbai, Ahmedabad. We are primarily a DC company, but we started our AC division also two years back when we acquired a UK based company who was doing some good work in their field. So today, we supply to MG and VinFast who are our prime customers. So if you happen to have an MG Windsor electric car or a VinFast, EV, there is a, I would say, very, very high chance that the charger you have at your house is made by me and my company.

Utsav Somani: Wow. And that, of course, is tremendous scale. What is blocking the or what is a big hurdle right now in adoption of EVs in India? Is it the charging network by any chance?

Arth Patel (Tirex Chargers): No. So there are two things. One is definitely options. Right? Today, most of the consumers in India are buying cars which are sub 10 or 12 lakhs. If you look at sub 10 or 12 lakh category, there is hardly one option of electric vehicle available, which is just, I think, Tata Punch. Yeah. Above 12, yes, there are many options from almost all the automotive OEMs. And above, I would say thirty, thirty five, there are multiple options. So the more we go in the larger size of market, which is more price sensitive, the options are less. So that is why people are not willing to shift. Second, charging, I would say, is a mindset, for I've driven an EV for the last eight years now. I've never yeah. Eight years. Yeah. And I've never charged at a public station. I always charge at my house. Because on a daily basis, right, out of, let's say, three hundred sixty five days a year, out, let's say, three fifty days, you are driving to your office, to your home, or to your friend's place, or to your

Utsav Somani: to your car.

Arth Patel (Tirex Chargers): Right. That's less than 100 kilometers a day. Any way in the market can do that on a daily basis. So charging is a secondary concern. More is it has to do with options.

Utsav Somani: I mean, it can benefit, like, the pollution in Delhi at least a lot by, I mean, solving that. But when does the maths really work out? I mean, just a fun fact. Maybe you can share with us sort of maths you've done on, EV. How many kilometers do you need to drive? How many years do you need to own for it to actually make sense over a combustion engine car?

Arth Patel (Tirex Chargers): So, on a typical, I would say, unit economics basis, an average EV will cost you some 1 to 2 rupees per kilometer. While a CNG car is 3 to 4 and your typical petrol diesel vehicle, is 8 to 15 depending on the average your vehicle gives. So EV on a unit economics basis is any day cheaper than all of the options. And now that the options for example, if you buy an Exxon, right, an Nexon ICE versus a Nexon EV, there is hardly 80,000 or 1 lakh rupee difference between the two. So that can be covered in less than a year's time if you're on if you're an average commuter driving 50 kilometers a day. So that concern that, you know, EV is more expensive used to be there two years back or three years back when there was a difference of 4 lakhs or 5 lakhs. But now with the, you know, new battery technology coming in, the more volumes that the OEMs have got, the cost is almost the same.

Utsav Somani: Interesting. And how many, generally, EV cars exist or bikes exist for number of charges there is? How is that ratio looking right now?

Arth Patel (Tirex Chargers): Oh, surprisingly, in India, that ratio is very well. If you look at the global scale

Utsav Somani: for

Arth Patel (Tirex Chargers): a four wheeler so we typically work in four wheeler and above. Four wheeler trucks, buses, construction, etcetera. I can say for these segments, two wheeler and three wheeler is not our target area. But for four wheeler and above, globally, one is two twenty or one is two thirty is an acceptable ratio ratio. So for one charger, you have 20 or 30 cars. India also falls in that similar band. That's why I said that charging is not the problem. It's just the number of options and the mindset.

Utsav Somani: And I'm guessing the industry will, of course, there is a white space, then they'll, of course, go into Mahindra is doing great work also and Mahindra

Arth Patel (Tirex Chargers): has been amazing work. Mahindra is doing amazing work on both fronts. So, yes, they have many options which are very beautiful. They perform very well. They're also building their own charging network. So we are supplying to Mahindra the DC chargers for their own networks. So Mahindra did not rely upon the public charging. They said we'll create our own for our customers. They can be used by other people also, but at least the person who's buying a Mahindra car is happy that, okay, my OEM has made their own charging network. I can anytime go there and charge.

Utsav Somani: And the prime minister also announced the PM E-DRIVE where they want, like, over 6,000, public charging points. Yes. What is the point on that?

Arth Patel (Tirex Chargers): So government obviously has to play its part because, in any new industry, you know, they have to give a push to get it rolling and then private players come in. So government PM E-DRIVE covers a lot of things. Charging infrastructure is actually a very small part of it. Most of the PM E-DRIVE is focused on electrifying your buses, your trucks, your ambulances, and these things. Today, from a pollution perspective, you see, buses and trucks contribute more pollution than the four wheelers. So government is very heavily focused on electrifying the local, let's say, Ahmedabad or Delhi Municipal Corporation Transports, the intercity between, let's say, Delhi, Jaipur, Delhi, Chandigarh, these sort of routes. So their focus is very heavy. They are also now incentivizing lot of corporates to go electric. For example, we recently supplied a few chargers to a cement-making company. So they had a fleet of 100 electric trucks, and they have bought in 50. So they've converted 50% of their fleet to electric. They are seeing a value in terms of their operation, plus they are getting an incentive. But if you are a large corporate owning your own trucks and whatever buses for your tech for transport and all, if you go EV, we'll give you some incentive. So there is a big focus by government on these areas.

Utsav Somani: Interesting. And your corporate structure is also very interesting. You started this company very young. You sold majority to Gulf Oil, which is part of Hinduja Group listed. Then you've got, somebody from Whirlpool India as, who's the non executive chairman on your board.

Arth Patel (Tirex Chargers): Yeah.

Utsav Somani: So, I mean, why this move to become part of a larger corporate? Does this open door? Does this give you distribution? What are the advantages? How does life change after this?

Arth Patel (Tirex Chargers): So life changed significantly, I would say, after. So three years back, we had an opportunity to be part of Hinduja Group to Gulf Oil, obviously, and we took that chance. It definitely opened a lot of doors, brought in a lot of investment for us to do a new plant, new R&D capacity, and all to bring to bring products, I would say, not just for India, but for the globe. I come from a family background. I have business family background, so I never had the mindset to go for a PE or VC. That doesn't fit in my I'm a Gujarati by nature. So at the end of the day, profit is something that has to be made in a business. I cannot run a loss making business. That doesn't fit in my head. So due to that fact, a strategic investor like Gulf and especially Hinduja Group who is so large, and they are also very much focused on this segment as a whole. They have Ashok Leyland who's making electric trucks and buses. They have Hinduja Renewables who's making battery energy storage. So for them, it was a natural extension to be part of the whole mobility value chain. And for us, it was a chance to be part of a larger group where we chose to have the control of the company in many ways also, and it opened the doors to a global platform. Hence, we went ahead with that opportunity.

Utsav Somani: And are you looking at export as well?

Arth Patel (Tirex Chargers): Yes. Definitely. So the group is global. We are present in more than 180 countries. So export is definitely on the cards. Not immediately, but maybe one year down the line, we will start looking at exporting DC and AC charges to Southeast Asia, Europe, the Middle East, other parts of the world, for sure.

Utsav Somani: And talk to us a little bit about your R&D because, I mean, this is a super technical field. And also, is making India a priority for you right now, or are you importing most of

Arth Patel (Tirex Chargers): So yes. So, so we are actually one of the few companies who have a 100% control of our product in an industry where a lot of people are either assemblers or traders. We have invested very heavily into having the control of both the power electronics as well as the software side of it. The goal is not to I would say, I it's wrong to put it in this way. The goal is to localize. Yes. But the larger goal is to have control. Because when we have control of a product because we are a product company. Unless we have control over how our product performs in the field, how reliable it is, then and only then we can have a value proposition ready for our clients. So localization, because it's too early, the industry is too small. Though we are focusing on it, we are strong believers of Make in India. So almost, I would say, 50% of our product is made in India. We are trying to increase that value to a larger extent to up to 80% the next one, one and a half year, but it comes with scale. Currently, the industry is not large enough for us to have a fully made in India charger. But, yes, a fully controlled in India charger designed and developed by us that we have. For AC chargers where the volumes are very high, it's 100% made in India. So any MG and VinFast chargers, now also some of you other OEMs we are signing up, that those are 100% made in India. For DC, because the market is not very big as of now, the focus is more on control rather than localization.

Utsav Somani: What is the split for you in terms of DC? Everything you said, it's the majority.

Arth Patel (Tirex Chargers): In terms of revenue, 80% is DC. Only 20% is, AC. And who is

Utsav Somani: this the biggest customer for you? Like, what's the ideal customer profile for

Arth Patel (Tirex Chargers): For us, the ideal customer profiles are bus manufacturers, truck manufacturers, I would say large corporate fleets. There are lots of logistics large logistics companies. So these are our ideal, go to customers for DC chargers. For AC, it's always the car auto OEMs.

Utsav Somani: And what are your wishes going forward, in the next three to five years? What are some things that you hope improves in the EV ecosystem in India? I think one you mentioned that more affordability.

Arth Patel (Tirex Chargers): Yes. More affordability, definitely. I would say better battery technology. I have been going to China for quite some time now because EV electric mobility is very closely linked to that country. The options that are there available in China for the cost, are much better. So Indian OEMs have to really step up if they have to provide a value. If I'm buying a 20 lakh car in India, it has to be the same level of a 20 lakh car that is available in China. So the wish list would be, yes, work on battery technology, work on affordability.

Utsav Somani: And tell me about back I mean, now that you've mentioned battery technology, can you maybe give us a two minute primer on what is being used right now? What's there in China? What can be improved in India?

Arth Patel (Tirex Chargers): So currently, it's all, LFP batteries that are being used in the car, which is lithium phosphate, I believe. Mhmm. Globally, there are lots of different technologies going on. Lithium phosphate is lithium nitride is one. I think gallium is also being introduced. So there are different researches going on. Battery research is something that takes time. It requires a lot of funds. Fortunately for India, all of the large groups, the Reliance, the Tatas, and the Adanis have entered the space. So I would say in the few years' time, at least five, six years, we will see some, monumental growth or I'll say some good breakthrough, from an Indian company doing great in the battery space.

Utsav Somani: Alright, Arth. This was super informative. Thank you so much for coming on our show.

Arth Patel (Tirex Chargers): Thank you. Thank you, sir.

Utsav Somani: Alright, listeners. Moving on to our third guest today, Mayank Jani. Mayank, welcome to the show.

Mayank Jani (Nanta Tech): Hi, sir. Good evening. How are you?

Utsav Somani: Good to see you. Good to see you. Dialing in straight from Ahmedabad.

Mayank Jani (Nanta Tech): Yes.

Utsav Somani: 2050 India, the 2047 Viksit Bharat pavilion.

Mayank Jani (Nanta Tech): Ah, okay. Okay. You see in the background. Yes. Yes. So that's what we are seeing in the future. Zero five zero.

Utsav Somani: You went to the BSE SME exchange last year, and you're doing I mean, a couple of businesses. EV, sorry. AV is one of them. And, the robotics aspect is much more newer and I think much more exciting also, for you. So describe the group to us. What is that?

Mayank Jani (Nanta Tech): Yeah. Yeah. We started because of the background, we started with AV in 2018. But then, we moved to robotics. We started robotics in 2021. 2021 to 2023, we just done the R&D on the robotics. We just understand the software part, hardware part, how the integration has been happen. So in 2023, we started selling the robotics. So the as a group, right now, we have a two major verticals. One we call is AV, which is our old vertical, which, we are doing particularly meeting room, boardroom, conference room kind of environment for digital transformation. And, robotics, what we call is a physical AI. So just not a moving machine, but, entire ecosystem, whenever the physical is required, it is a second vertical part we are focusing.

Utsav Somani: I mean, that's fascinating. And what are the use cases of robots? Like, are they heavy industrial robots? Are they moving simple things? Like, what scale should we visualize them with?

Mayank Jani (Nanta Tech): Okay. So, it's it's actually everywhere. So, heavy industrial robots are there, since quite some time right now, by some of the Japanese and Chinese companies. But what really happening right now, it is a small piece. It's like hospitality sector. The entire hotel can be run without even a single human, you know, with, use of the robots.

Utsav Somani: Is that happening, frankly, in India?

Mayank Jani (Nanta Tech): It's not Indian right now, but, yeah. Lots of groups are moving towards that. They've been achieved 25, 30% kind of automation as of now. So I think next three, four years, we are seeing lots of hospitality chains running without, any human, most of the time. In Europe, it is actually happening, at the some of the places. So that is the one. So hospitality is everywhere, so it's even high in residence, hospitals, hotels, all the industries. And second is the warehousing, which is a much more used case where the, dark factory concept where even on the factory floor, you don't need a human to run the application. Yeah. Yeah. But in

Utsav Somani: the hospitals, we're gonna like, how do you this thing because the dexterity required or that would be humanoid level, dexterity. Right? Which I think people are still researching on.

Mayank Jani (Nanta Tech): Yeah. So that is that's still, research is going on. So that's why I said, but, for example, you are ordering something from your room and from the reception, when robot come through the lift, reach out to your room, and call on your PBX that, you know, and so and so. Open the door. You open the door. It will open, its space, and you take out the whatever you order, and then it goes back. So that level of automation is already reached. Hand is still externalities, again, you rightly said that it's still, under the innovation. I think in next two to three years, we have that kind of expertise also into the physical AI. But once that reach, I think, then the, the the sky is the limit.

Utsav Somani: And how capable are these robots? Like, what kind of data, what kind of models are being used, and what kind of training is required, and how fast can you deploy? Say, suppose I have a hospital of 50 beds, then how fast can something like be, this be deployed?

Mayank Jani (Nanta Tech): So for a particular hospital or hotel kind of environment, the already test has been done. So already, the training has been done on the physical models. So the difference between physical AI and, which is right now software AI is that on physical AI, the training model is not not there. You have to train the models yourself. So, for example, if you are working on one hotel, you have to train the models, how it works with the hotels. And plus, you have to do the lots of integration of the software has, all the hotel and hospitality industry had their own, standards, the home software. So HRMS software, visitors software, and everything. You have to integrate everything and then give it to the customers. So a typical our, case study so the three to six months we take, to reach out to 50% kind of automation right now, which is right now the limit, to the hospitality sector. I think within next two years, if you go phase wise, two years, you can have the 100% automation, the particular hotel kind of scenario.

Utsav Somani: And pulling up a letter of intent that you signed, and filed with the stock market exchange, I think, Connplex Cinemas.

Mayank Jani (Nanta Tech): Yeah.

Utsav Somani: You're supplying 100 Allbotix AT10 marketing and delivery robots on a very interesting I mean, SaaS, but, you know, robots as a service model pay, which is 40,000 per robot per month. So how does this work? Why do cinemas need robots?

Mayank Jani (Nanta Tech): So now, what we are seeing major change that is happening that is is a RaaS model. So that was one of the case where, what they are doing is there is a big LED in front of the robots, then there are the plates on the behind of the robots. So from the LED, they can run the advertise for the customers for the maybe trailers and everything, and they can earn the, from that, advertising. So that is earning from them. And on the next side, they can put all the marketing material, whatever they want to put, maybe popcorn or chips or anything. So this this one of the concept. They want to digitize the, waiting area, which was there on a complex. Plus, also, they want to generate a revenue. So if they are considering the ROI within three, four, two months, three, four months, they are able to get the revenue back, what they are spending on the robots. And but, overall, as a RaaS model, it is now becoming very popular. A lot of the FMS company, which is a facility management service company, they want to have a robot as a human. You know, they are asking that, you know, I want robot for a two year job because they have

Utsav Somani: staffing agency but for robots.

Mayank Jani (Nanta Tech): Yeah. So now lots of FMS agency now converting to try to enter into the robotics. So they're asking for a RaaS model for three, five years kind of model from us, and we are talking to lots of big companies in India, outside India also, where they are going to get robot as a employee, not as a asset.

Utsav Somani: Okay. So is may I mean, talk to us a little bit about your R&D, the team, and how much of this is built in India, assembled in India, or how much of this is just imported and packaged for India?

Mayank Jani (Nanta Tech): Okay. So, in robotics, as Riley said, a lots of training needs to be done. So there are the three parts right now where, Chinese are having, most of the market sell with actuator, controller, and their batteries. What we have done here internally is we have spin off our R&D, and we have given the name TR India Private Limited, which is only focusing on the R&D part of the physical AI.

Utsav Somani: Okay. So collecting data, including models.

Mayank Jani (Nanta Tech): Everything. Everything related to physical AI, R&D, whatever in the software also that we are doing in TRN. So there, we are taking orders from the robotics company itself also. They want to do something, R&D into their own model. So we are providing we have the engineers, we have the machineries, we have the, lab setup and everything over there. So these three parts is right now still, across the globe. 99.9% people are still dependent on China. And we are trying to get three three parts, in next two years to get in India also. So we are hoping that in next two years, we are going to have a 100% Made-in-India robots. Right now, apart from these three parts, which is around 40% of the entire physical, robot, 60% is Made in India. All the operating systems, techs, software techs, AI techs are our in built in house. So and that we are that's how we are doing, for the R&D part. Because R&D is the most important thing right now in the physical AI.

Utsav Somani: In robotics and, I mean, especially physical AI, I think is becoming a big big, opportunity. So if you were to give, like, a cheat code to founders or potential founders listening in, like, what are two white spaces that they can focus on?

Mayank Jani (Nanta Tech): So that in physical AI, the two white spaces, which is right now not heavily focused, which is, again, smaller, models, which is not like a full humanoid kind of model because humanoid will take time and it will have their own problems when in terms of the manageability and everything also. Responder like, for, so one of the project maybe for working for, old age people. Well, they are alone at the home, where you have a robot, which is not maybe a full stride robot. It's moving around the house, and it can take care of the patient. It knows the, you know, them, and it can take care of them. Because right now in India also, that is, going to be a problem where parents are living alone and their children are outside India. So where if they have something autonomous which can help them, on a daily task. So that is one of the medical sector. Physically, it's, again, a very big white space here. Lots of people have not concentrated, but there is it's going to be a very big market in the future.

Utsav Somani: Interesting. And then, in July, in our automation expo in Bombay, you announced AT-Base where half robot same stays the same, like the brain, the capacity, and, I mean, all the other things, like processing battery, everything remains the same. You customize the stuff on top. So it gives you, like, a modular robotics platform. Talk to us about that.

Mayank Jani (Nanta Tech): Yeah. So the AT-Base is the base platform where all the autonomy has been given, and we are very proudly said that it was 85% Made in India. Only couple of item only we bring from the China. Otherwise, everything was from India, including the hardware and soft text. Now on that base model, you can create anything. For example, if you want autonomy on a outdoor platform, you want you have a farm and you want to do the patrolling of your farmhouse. There, you just put a camera, mic, system, and sensor and software around it, and you can use it. So this is the base platform that we have created. We are in talk with lots of companies. They want to create their own model for their own requirement, where we are going to give the base to them along with the software and IP, and they're going to build their own robot and they can use into their own environment. So that's the base level that we already launched. We are, in process of launching much more models in next, six months. But this is the base and the basic model that we are launching globally. Even UAE also, we are in talking with lots of robotics company in Dubai to provide this base to them.

Utsav Somani: Amazing. So you're going international. What does the revenue for Nanta Tech look like? All of this is public info, whatever you can share, of course.

Mayank Jani (Nanta Tech): Yeah. So, last year, we covered up

Utsav Somani: AV and robotics. I'm guessing robotics is a much newer bet for the firm.

Mayank Jani (Nanta Tech): Yeah. Yeah. So robotics, we have, almost, doubling every year since last three years. And this year, looking at the work orders and looking at the funnel, I think we are going to do more than double than the last year, what we have done. And AMB is going to grow, conservatively, 40% year phase where we are getting the repeated order from the same customer most of the time. So, yeah, we are very excited about the physical AI. This is India. Dubai also, we are expecting a good number of orders coming, from the government and private both the sectors. And Malaysia, we have a sign of distributors, so we are hoping a good business, from that distributor for Southeast Asia market.

Utsav Somani: And what's next for you?

Mayank Jani (Nanta Tech): So next is, again, we are going Europe very soon. We are, we are about to finalize one partner over there. So we want to be a company which has a base software base and hardware base along with the Ready. And, globally, we can tie up with the companies, and we can provide end models to them. So that is the idea, in the next coming years.

Utsav Somani: I think maybe as a side question to end, because you listed and I think maybe one of the first few founders we've had on the show who's listed on the BSE stock exchange, the BSE SME exchange, what was that process like? Why the decision to go to SME exchange? Why not wait for the main board? Like, what were the decision making parameters for you when you decided, to go for the BSE SME exchange listing? And, also, how hard was it? Because during, I think, the last two, three years, a lot of founders have explored that as a liquidity option as well.

Mayank Jani (Nanta Tech): Yes. That was hard. So, it was not easy. But, as we were very heavily focusing into the robotics, we are requiring a heavy fund to do the R&D. And from the banking and all sector, there was limited fund that is available, for the R&D purpose. They provide the fund for the business purpose. For R&D purpose, there is no one provide. So we are, fortunate enough to, meet some of the good investor who are very optimistic into the robotics sector. And, with them, we have decided to go for a a BSE SME first and next, as long as we able to complete the three years, we will be immediately moving to the main board. So very

Utsav Somani: And you need to be profitable and hit a certain scale, I think, or, is there some qualification?

Mayank Jani (Nanta Tech): There are there are direct criterias to go to reach out over there, but I think we are easily able to fulfill that criteria in two years only. And three years, it's a wait time that we have to wait, so we will be going in three years for the main board.

Utsav Somani: Man, this is fascinating. Thank you so much for coming on our show. Wishing you all the best.

Mayank Jani (Nanta Tech): Thank you so much, sir. Thank you for having me.

Utsav Somani: Yes. Alright, listeners. We're moving on to our final segment for today. We've got Shezan from Fixxly. Shezan, welcome to the show.

Shezan Bhojani (Fixxly): Hey. Hi, Utsav. Thanks for having me.

Utsav Somani: Did I get the pronunciation right?

Shezan Bhojani (Fixxly): Yeah. You did. For somebody for a change, you got it right.

Utsav Somani: Awesome. So even before an invoice was issued, you raised 5,500,000 from three of the best investors, Lightspeed, Accel, Fireside. How did you tell them that thirty minute, construction materials is, an industry waiting to happen?

Shezan Bhojani (Fixxly): Mhmm. Well, I think the for anyone who's been, building in India, I think the problem is kind of obvious. I mean, if you build a house, I mean, just or had to repair a shower in your house even. You know, just think about, you know, if you had to repair a head for a shower or the last time an electrician came to your house, they come to your house and, you know, they say, okay. Problem major. Or even if the guy comes, you know, he comes and he said, I'll come back and he will come back tomorrow with the park. Right? So obviously, there is a lot of time that's lost for him and for you. So the problem was obvious that, hey, man. This there are a lot of businesses that are being built on quick commerce, but there are very few that where quick delivery is actually solving a problem. And I think here, the good thing is that Quick actually solves a problem. Right? You don't have to, lose time, labor, etcetera, etcetera. And this happens in millions of site across India every day, no matter large or small. Right? When you can talk to a builder building million square feet or you can talk to, you know, the repair and placement, it happens every day. So it's frequent. It has a large AOV unlike in other other big companies, AOV because it's it's generally expensive. So I think people got the problem statement. Yeah.

Utsav Somani: No. No. I mean, that's a crazy lineup of investors. I believe you're eleven days away from launch.

Shezan Bhojani (Fixxly): You're right. I we are about a little less than eleven. Yeah. About there thereabouts. Yes. Eleven days from

Utsav Somani: And you're going in Bangalore. So I think the problem statement is fairly obvious. Thirty minutes, you need x y z, like, a plywood or something like that to come to your place, even a shower, head. But, I mean, the catalog I mean, so you've, of course, built a DesignCafe, which was a home interiors firm. And, now you're doing this. You've understood the problem, but the solution must be, like, I mean, so much out there. Right? Because the catalog is so heavy. And your cofounders from Zepto. So Zepto, smaller items, quicker, like, delivery possibility because you can package all of this and have multiple task force because the use case frequency is also higher. But in this, like, I mean, a plywood or something much bigger, it can vary. Like, there are multiple variations. The SKU category is I mean, the possibilities are endless. How do you think about these as constraints or as opportunities?

Shezan Bhojani (Fixxly): So I think you you like, if any problem, you go first principles. Right? And you say, okay, listen. What do people need most frequently? Right? And in in and the and what can I and and good business are built, focusing on the the core problem, not the fringe problem? Right? So if you look at any construction site, there's just a few things that you need every day. A core. Right? Whether it's let me call it this paint, whether it's private, etcetera, etcetera. That's a limited basket. So one of the guardrail of the design of the business was anything that can be bought without touching or seeing. Like, if you can close your eyes and buy something, that's what he looks at. Right? You don't because you know, I mean, paint, white paint I mean, you don't need to, like, touch or dip your finger in the paint, you know, white paint of every color screw or whatever it is. So the consumer needs to, know what he's buying before buying it. So that's that means we don't have decor or any of those things which, you know, require the large. So when you narrow that down, you have actually a fairly limited SKU count in every dark store because there's only so many essentials in building material that you can start. Right? So that's, I think, one way to look at it. The other thing that is sure, if you either have small delivery vehicles or larger vehicles. Right? The advantage of large delivery vehicles is the AOV goes up. So if somebody's already 10 ships of plywood, that's, you know, $30,000. I'm very happy sending a bigger vehicle to that. Right? For for other quick call to get $30,000, they probably have to do, what, 20 30 degrees or more than that. So, like, 200

Utsav Somani: control logistics in this? Like, other fleet will be on?

Shezan Bhojani (Fixxly): No. I mean, dedicated fleet from three people, obviously. But yeah.

Utsav Somani: Okay. And how do you think about returns? Because if somebody doesn't like it, then you have the return, thing to also worry about. Right?

Shezan Bhojani (Fixxly): So I think, again, the good thing is, the category that we are kind of building, like, semi color paint. I mean, I mean, you like it. You know what it is. If it's damaged

Utsav Somani: Paint, maybe also see. Right? I mean, I remember my when my wife was choosing a, paint. Like, I mean, there are, like, hundreds of just pinks. So how do you, like, I mean, know?

Shezan Bhojani (Fixxly): Yeah. So do you want

Utsav Somani: like, I mean, type of paint or the same I think the paint code can vary between different manufacturers also. That's what I learned.

Shezan Bhojani (Fixxly): That's true. But the thing is when you're punching in an order, you already know what paint you want. So you're punching in so for example, Asian Paints has 3,000 SKUs of 3,000 colors. Right? So you punch in that exact color code that gets, so we have those 3,000 color codes on my app. If you're picking that Asian Paints color code, the exact same thing will come. So, like for like, you wanna get the exact same thing. In case there's a damage, we, of course, have reverse logistics where you pick up etcetera. There is a return and replacement, but the the number of returns in this, my again, it's a guess, will be far lesser because of the kind of SKUs we are keeping. Right? So if you look at if I were to look at my, analyze what would be the most selling SKUs and where will color paint sit? It'll probably be the bottom 1% of auto value or number of orders. Right? Most of my orders are going to be the regular items that are just being bought every day. Right? So I think unless there's a damage, you're not going to have a return or replacement. Right? Unlike in fashion or something where it's only a fail or you don't like what you're buying, you're in India.

Utsav Somani: Or you order multiple sizes of the same thing. But here, how how big are your dark stores compared to say, like, a Blinkit or a Zepto?

Shezan Bhojani (Fixxly): Our dark stores are similar size similar size.

Utsav Somani: Oh, wow. I think one advantage that you've not mentioned is, something that I discovered when, like, I think we're doing our place. Man, the same thing in GK can be, I mean, 40% higher where you go, like, to Gurgaon, I think it's 30% cheaper. Because the commissions are so notoriously high in this that you don't know I mean, there's no concept of MRP in this. Or MRP here, but, like, the number of discounts that you can get from because there is no interior designer, in the loop or no contractor in the loop or the contractor is passing some discount along, to you as a customer. Is that, something that you think will be a big benefit on why people will come to a trusted platform?

Shezan Bhojani (Fixxly): I I think that's you're right. That's a very big one for anybody. Right? I think in general, unorganized markets have an capacity to develop. Right? And any organized player wins because they cannot win transparency overnight. So I think for us, whether you buy, you know, one, you know, liter of paint or you buy 100 liters of paint, the price per liter is the same. I'm not gonna charge you different, because you're somebody I know for a long time. Right? There will be loyalty and rewards which are transparent, but you don't have to guess, Kim, am I getting the right price? You're getting the same price, as the guy buying 100 liters of paint, because that's the price we select.

Utsav Somani: And supply chain credit, how do you think about those things? Because I was surprised that your dark store is a similar size as, say, a Blinkit because you might have to store many more SKUs. So maybe the quantity is lesser. But restocking them and having connects with the end man I mean, the first manufacturer to bring them to you and all of those things, I think, must be a really complex thing to solve for.

Shezan Bhojani (Fixxly): It is. It is. I think, I always think that it should always solve an obvious problem, but, it should be complex because that's your moat. You know, and if you're able to do a good job on that, it becomes a very strong moat for the business.

Utsav Somani: Interesting. And, I mean, given that your team of course, the cofounding team also has experience from Zepto, what are a couple of things that you are to unlearn from that, head of grocery, quick commerce?

Shezan Bhojani (Fixxly): Yeah. I think, just, I would say surprised at how this supply chain works very differently to FMCG supply. Because I think the FMCG supply chain is a far more organized, far more used to distribution, etcetera, etcetera. Whereas the construction and building material ecosystem is still fairly stuck, you know, as I said, decades or so behind. A lot of capacity, a lot of dealers, a lot of, you know, stuff that it's just it doesn't function in the FMCG world today. Because if you look at in construction, even today, I think barring cement, steel, and paint, in every other category, only 50% is the organized market share. Right? So that's a large that's probably that's about $40–50 billion of unorganized market that's exist. Just in the unorganized brand. Right?

Utsav Somani: And you will stick to the 30-minute promise. I think that's a promise that you, this thing. And who are you building this for? Like, I mean, you mentioned the urban company example where somebody comes from urban company, but do you think contractors will be a big, customer for you as well? Because they might have five sites running at once versus, say, somebody who's just getting a shower repaired.

Shezan Bhojani (Fixxly): No. So I think we're very clear that we are building for the prosumer, which is the professional. Right? So the way we look at it is there are small contractors, big contractors, but essentially, we will do this every day of their life. And we are clearly building for the prosumer.

Utsav Somani: Alright. And India's, I mean, notoriously a do it for me market versus do it, DIY. Right?

Shezan Bhojani (Fixxly): Exactly. Exactly. Exactly. Yeah. So

Utsav Somani: Alright. Wishing you all the best. Won't keep you too long. I'm guessing there's a lot to do before the launch. So wishing you and the team at Fixxly all the best.

Shezan Bhojani (Fixxly): Thanks a lot, also. Thanks for having me.

Utsav Somani: Cheers, Shezan.

Shezan Bhojani (Fixxly): Cheers.

Utsav Somani: Alright, listeners. That's it from us. Hope you had fun listening into four solid operators about entirely different industries. We will see you on Monday at 04:00 same time. Have a safe and fun weekend. Bye bye.