Full Transcript

Dhruv Sharma: Hey there listeners, how is Monday coming along for you? There's a lot of interesting stuff happening in the world right now and before our first guest who incidentally works in the space domain joins us, we're going to cover a bunch of these interesting things that are going on.

Utsav Somani: Yeah, I mean, two very, very cool guests actually in all different industries. One enterprise, super deep in the world of SAP's and ERP's, and then we've got a space founder. We've had a bunch of space people come on the show as well and educate us from launch vehicles to different kinds of things, satellites and drones, and I think it's just cool what's happening across the world. So, I mean, but the coolest thing I think which is happening now is we're beating some records. So, NASA is sending people on Artemis 2 and I think this will be the mission that beats the record set by Apollo 17 or Apollo 13, I think, right? Do you have some information on this drone? I'm tuning into Netflix tonight at 10.30pm to see this live.

Dhruv Sharma: They're going to do a lunar flyby, right?

Utsav Somani: Yeah, lunar flyby, Artemis 2, that's what we're calling it. And it's a generational moment because first humans at the moon in 53 years. India's space program, Chandrayaan program, mapped the South Pole. Now NASA astronauts will see what are the potential future landing sites. And we might explore the lunar South Pole as early as 2028, so it's not just a hobby anymore. It's all becoming real infrastructure. So everything that we've discussed from Elon Musk putting satellites and data centers and compute in space to now this, so it's all happening.

Dhruv Sharma: Yeah, one step at a time. And this is Artemis 2. Artemis 1 went in orbit, I think, in November 22 and it took the same, you know, command module as this spaceflight, which is Orise. And the way it works in like, you know, space explorations, you've got to test everything step by step. And so I think I'll just pull up the scope of this particular mission, which is to test crew survivability. This is a manned spaceflight. Artemis 1 was unmanned. So that's number one. And then, you know, they validate all of their systems, their hardware systems, their software systems and ensure that they're able to even do some tests while they're up there and check out emergency, you know, their emergency operating procedures and also verify some of their subsystems. Yeah, I don't think the manned spaceflight itself is happening until 2028. I think they've just set the schedule a little bit differently for that. That's going to be Artemis 4. There will still be an Artemis 3 and they will attempt docking in low Earth orbit, but Artemis 3 is not going to be the one where they again put humans back on the moon's surface. That'll be Artemis 4 and it's scheduled for, I think, sometime in 2028.

Utsav Somani: Where are they all launching from?

Dhruv Sharma: I think, God, I'm not sure, but I think this site is somewhere in Texas, if I'm not mistaken, this particular site.

Utsav Somani: Okay. And they're carrying NASA astronauts and a Canadian astronaut, a 10-day mission around the moon. And this is the first crewed mission beyond low Earth orbit since Apollo 17 in 1972. And coincidentally, I happened to see this movie, watched this movie called Project Hail Mary, which is Ryan Gosling's new one that's come out. What's it about? I mean, there are these, I mean, sort of these bacteria or like these creatures which are eating away the Earth and they're eating away the sun actually first. And then because the Earth is going to like basically cool down to a temperature where human life becomes unsustainable. So they go on a mission to this distant galaxy to find, 11 light years away, to find why the same creature is not eating away the solar system there. And he meets an alien there, then ends up settling down on the planet of the alien and learns their language as well. It's a fascinating movie, two and a half hours, but I mean, got a great review, so I had to catch it in an IMAX, but made me wonder, like I went with some friends and I wanted to ask them that, do you think like space travel will be a reality in our lifetimes or in the next generations? Yeah.

Dhruv Sharma: In our lifetime, it will be.

Utsav Somani: I mean, do you think before we like leave this earth, unless Brian Johnson has his way, we might live slightly longer than a hundred years, but do you think we will be in space?

Dhruv Sharma: You know, so again, step by step, we already fly at about 35, 40,000 feet. And the next milestone would be to take a human space flight to about 65, 80,000 feet well within the stratosphere. And that's the work Blue Origin is doing as well, right? To take space tourists again, essentially into low earth orbit. But will we be able to do like interplanetary flights in our lifetime? Perhaps not. That is still a far way out, but going to space is, yeah.

Utsav Somani: Elon also changed his stance on that, by the way.

Dhruv Sharma: Going into deep, deep space is a little far-fetched for now, unless, you know, something dramatically.

Utsav Somani: It'll be fascinating. I know Virgin was selling tickets, right? Where you can like, space tourism was the big pitch there. Was that, I mean, what's happening with that project?

Dhruv Sharma: I'm not even keeping up with them. I don't know if they're still at it. Maybe they are.

Utsav Somani: It turned into a SPAC for Chamath. I don't think the SPAC did well. Yeah. All right. Let's talk about something closer to earth. What's happening on our Strait of Hormuz now?

Dhruv Sharma: I mean, there is this one story doing the rounds, which is, there is either a analyst or a team of analysts from Citrini Research who have gone behind enemy lines and done some research of their own. I think this is the boldest a Wall Street analyst has ever been in history and gone into harm's way with, I believe, armed with a wad of cash and a box of Cubans.

Utsav Somani: He was actually caught, by the way.

Dhruv Sharma: And I think he was interrogated for a while and then he bought his way out of that situation. And then he went and realized that there's a, I think that there's a phrase in the report also that there's a lot of situation monitoring slop going on right now. A lot of the people say, for instance, the Strait of Hormuz is mined, as in it has like underwater mines, maybe doesn't have underwater mines and many of the people who are unable to cross the Strait of Hormuz, it's just plain common sense. Nobody wants to die attempting something like this. And there's drones flying over and they can't really discern between a 40 year old like fisherman's dinghy and something that actually might be, you know, a valid target. But yeah, pretty cool information now.

Utsav Somani: This is what chasing the alpha looks like. That you're not relying on Chad Chibby and Claude, you're literally going to the front lines and just getting your data. Pretty fascinating. But talking about data, let's welcome our first guest and get some insights from him. He's capturing...

Dhruv Sharma: Who actually works in the maritime surveillance domain and without embedding like human analysts in the ground.

Utsav Somani: So once this constellation is up and running, maybe we get all of this data available like at an interval of 30 minutes to all of us available. Gaurav, welcome to the show.

Gaurav Seth, CEO & Co-founder of PierSight Space: Hey, Utsav. Thank you so much.

Utsav Somani: Probably made a mockery of the whole explanation. So why don't you come and explain what Pearsight does?

Gaurav Seth, CEO & Co-founder of PierSight Space: No, you did a good job, but I'll have a shot again. So hi everyone. I'm Gaurav Seth. I'm the CEO and founder of Pearsight. At Pearsight we're building a constellation of satellites, having this power to see through clouds image during nighttime and having a constellation ensures that we have a persistent revisit over all parts of the oceans.

Utsav Somani: And you're promising 30 minutes gap.

Gaurav Seth, CEO & Co-founder of PierSight Space: Yes, 30 minutes gap. So currently how it happens is that ships, they are asked to share their live location. For example, how I ask you to share your WhatsApp live location with me. But whenever you do something illegal, as a bad actor, you will try to turn it off. So right now that technology is as good as the on and off switch. What we are doing is augmenting that with the imaging system, which is a synthetic aperture radar. So it ensures that along with having this AIS signals, we also capture imagery over that area. So even if somebody turns off the AIS, we still have a visual signature. Kind of like a CCTV over the oceans.

Utsav Somani: We'll get into the technical side of things, but your story is fairly interesting as well. You spent nine years at ISRO and we briefly mentioned the Chandrayaan initiative as well. So you worked on the Chandrayaan-2 radar as well. One of the most prestigious jobs that any scientist in India can have and you left to start a startup. What was going through your mind?

Gaurav Seth, CEO & Co-founder of PierSight Space: So after Chandrayaan, I was working for a Venus mission. So Venus is full of sulfuric acid clouds and humanity hasn't seen the surface so far. The only way you can see it is through the same tech, which is called SAR. So we were building a mission for Venus, which was supposed to launch in 2023. But then COVID happened and we missed that launch slot. And the next time, Earth and Venus were coming close enough such that we could send a spacecraft for 2033. So that was a time I was figuring out, okay, so what to do for the next 10 years. And interestingly saw this maritime white space that this is also a huge area which needs to be solved where our tech can come up. And around the same time, India also opened up the private space sector for individuals like me to start a company, build these satellites. So I kind of thought that I am at the right place at the right time, have spent about 10 years in ISRO, seen the complete cycle of satellite development. And here is one unsolved problem. So that's how a few things came together. And PSI was born.

Dhruv Sharma: That's really cool. Gaurav, just to contextualize today, when you guys do maritime surveillance, is it within India's exclusive economic zone or is it in the international waters as well?

Gaurav Seth, CEO & Co-founder of PierSight Space: It's also in the international waters, Dhruv, because our satellites, they cover all over the earth. Like in 90 minutes, it takes a circle over the earth. So we are not restricted to looking only at one part of the oceans. We can see all over.

Dhruv Sharma: That's amazing. And how many such satellites do you have up in orbit right now?

Gaurav Seth, CEO & Co-founder of PierSight Space: So we launched one in-orbit demo in order to prove the tech, because we were building these satellites in a very different way, a small form factor, which is rapidly proliferable. Because you also have to ensure that if it takes a few billion dollars to build such a network, it may never see the light of the day. So we built something bespoke for maritime domain. To prove that, we launched our demo satellite called Varna on ISRO's POEM platform in December of 2024. End of this year, we'll launch our first commercial satellite of the constellation. And in the next four years, we'll proliferate it to make it to a 32-satellite constellation.

Utsav Somani: We were actually, before the show started, Dhruv and I were chatting about what launch vehicle were you using. So thanks for answering that. And you've launched one, and your timelines also vary in the sense that are pretty compressed compared to say ISRO, which is taking 18 to 24 months to build a normal satellite. And you've done that in under nine months. So what did you do differently? What were the learnings? And what was, I mean, how did you like rethink this whole timeline?

Gaurav Seth, CEO & Co-founder of PierSight Space: Yeah, so first major thing which we are doing differently is building application-specific SARs. So ISRO and you know, other old space agencies, or let's call them, you know, traditional space agencies, they were used to a time when you could send one satellite up in three to four years. So, you know, those were large satellites, the size of a school bus, about 2,500 kilos, doing multiple applications, maritime domain, agriculture, defense, mineral mapping, etc. What we are doing differently is we are picking a particular application, let's say maritime domain, where you only need to detect ships and oil rigs and activity happening over the oceans and building something bespoke for that. A good example can be, you know, in cameras, you see a DSLR, a GoPro, a dashboard camera, each one is optimized for its particular use case. Same thing we are trying to do with synthetic aperture radars. And hence our timelines are low, we can rapidly proliferate at a much lower cost.

Dhruv Sharma: I think our conversation would be so much richer, Gaurav, if you explain what a synthetic aperture radar is, and how it's different from, you know, other things that other things are recommended.

Gaurav Seth, CEO & Co-founder of PierSight Space: Yeah, yeah. It's looking for my phone. So you know, we all know our traditional camera, like if this camera, you shoot it up to space, what would happen is it'll take images, pretty good images, but seven out of 10 times it will see clouds in between. So you won't be able to image and in the remaining three times 50% would be night. And you know, this camera can only work if something is illuminating the ground, right? Like sun rays. So what a synthetic aperture radar does is it's a camera with a flashlight. And this flashlight happens to work in the microwave frequencies, because of which these microwave frequencies are not scattered by clouds, you can see through clouds. And because you're shining your own light, you can also image during nighttime. So synthetic aperture radar is a complicated name for a microwave camera. It's just like other cameras, but gives you a microwave image.

Dhruv Sharma: Does it emit like energy pulses?

Gaurav Seth, CEO & Co-founder of PierSight Space: Yes, it goes through, like, as I said, it emits microwave energy, which gets reflected back towards the sensor. And using this scattered back energy, you form images. And there's a bunch of signal processing, we perform on it to use the Doppler, which is generated, getting technical between the you know, platform and the target. And using that we are able to enhance the resolution to as good as an optical camera would provide it.

Dhruv Sharma: One quick question before we zoom out, does it does it actually have a physical antenna or is it only synthetic antennas?

Gaurav Seth, CEO & Co-founder of PierSight Space: No, so it does have a physical antenna. But how the name synthetic aperture comes is that, you know, in microwave wavelengths for a three meter antenna at 500 kilometer altitude, the resolution on ground you will get is about 10 kilometers, but essentially what we need is few meters. So what is done is that the resolution is synthesized over the entire period on which you know, the sensor moves above the target and that entire thing is integrated and using signal processing we enhance the resolution and hence the name synthetic aperture. So while it has physical aperture, but while forming the image, we use a large synthesized aperture which is created due to the satellite's motion.

Dhruv Sharma: Like a time lapse as opposed to a single shot image?

Gaurav Seth, CEO & Co-founder of PierSight Space: Time lapse is the wrong thing. Maybe like, you know, when you take a night image using a DSLR, you do a long integration. Consider a long integration, but just that the sensor by the virtue of the platform is moving against the target. So at each position you have a different Doppler signature of the target. As using this Doppler signature, we further enhance the resolution.

Utsav Somani: And who will be good customers for you in this business?

Gaurav Seth, CEO & Co-founder of PierSight Space: Yeah, so for maritime domain, because, you know, we are providing the complete insight of what is moving where. One major set of customers is hedge funds, people who want to have an arbitrage by having information of, you know, how commodities are flowing and where they are flowing. Next insurance companies, right, like we know several countries in the last few years due to war have been sanctioned, but still there's a huge parallel economy of, you know, few hundred billion dollars, which goes on due to ship to ship transfers, which happens over the oceans. So insurance companies are another big set of customers. Then of course, the obvious ones are coast guards, as well as, you know, the national EZ protection agencies, navies, right? So these are a few sets. And then also a lot of agencies which are against IUU phishing, IUU phishing stands for illegal, unregulated and unreported phishing. This happens in huge quantities, so much so that, you know, like 20,000 square kilometer of area in the Bay of Bengal has become unfishable. No fishes are there anymore due to the IUU phishing. So there are a lot of agencies such as the Global Fishing Watch.

Utsav Somani: Data like this, I mean, given that you're an Indian startup, is it restricted to be sold overseas, like in the sense that can you have all sorts of clients or will you need approval from the Indian Defense Institute or any sort of approvals required before signing on new clients like these?

Gaurav Seth, CEO & Co-founder of PierSight Space: No. So geospatial laws are stricter once you go towards lesser than one meter resolution. Then, you know, the more regulations come into place. But we are here detecting large vessels. So we are not, you know, very high resolution SAR camera. What we are doing differently is wide coverage, right? So hence, we will not fall under majority of, you know, regulatory hurdles. And also, as per the UN Outer Space Treaty, you can capture any part of the world from satellites as you want. Only when you go and sell that data is when the law of land applies. And usually, going by my previous point, this law of land is stricter when your resolution is less than one meter. So that is why we will not have any hurdle in selling this data.

Dhruv Sharma: Out of curiosity, what are the different sort of spatial resolutions or like what's the range of the size?

Gaurav Seth, CEO & Co-founder of PierSight Space: You can go as far down as 0.2 meters. But what happens is that the more resolution you want, the more amount of power in a SAR satellite you need to pump in. So that reduces your coverage. So there is always a tradeoff between resolution and the size of area you can cover in. And that's where, you know, our solution is very bespoke and tailor-made for the maritime domain, that you maximize your coverage while having the right resolution and sensitivity which is needed to detect ships.

Utsav Somani: And talk to us a little bit about this consortium that you are a part of. We've had the founder of Pixel come on the show as well. I think it's a 1200 crore consortium backed by the Indian government where you're building a constellation of satellites for them, I think.

Gaurav Seth, CEO & Co-founder of PierSight Space: Yes, yes. So that's another, you know, tailor-made need we are serving. So Indian government realized that every year about 1250 crores of data, different government departments buy and that too from foreign providers like Planet Labs, Maxar, etc. Why not get this served by an Indian-made constellation? So that's where they announced India's first public-private partnership, PPP model in the Earth Observation sector. And they collated all of the demands from these different departments. And using those demands, which are, you know, in plain English, they made specifications for about 12 satellites. And these satellites are synthetic aperture radars, high-resolution optical cameras, as well as hyperspectral satellites what Pixel is building. And that's where, you know, four startups got in together, Dhruva, Pearsight, Pixel, as well as Satyodh. Each of us are building a different class of satellites. And together, you know, we will be serving the Indian government through a new entity which is called Allied Orbits. And in the next four years, we'll be launching these 12 satellites, which will provide all of this data to the Indian users.

Utsav Somani: And this is on build-operate-transfer mode or do you run it on your, I mean, under your control?

Gaurav Seth, CEO & Co-founder of PierSight Space: We run it under our control. The InSpace, which is a regulatory body, provides all the structural, technical and policy support. Because, you know, there's a huge change which is happening here. These departments previously were procuring this data from ISRO, from foreign providers. And this consortium will bring it on to the government e-marketplace. So InSpace will be facilitating all of the, you know, technical policy support for that. But we will be operating these satellites under Allied Orbits.

Dhruv Sharma: Gaurav, does opening up the private sector to space and especially commercial applications free up, you know, what, and you've been at ISRO before, free up time and bandwidth for ISRO to do things that are, from a strategic standpoint, more important for us?

Gaurav Seth, CEO & Co-founder of PierSight Space: Yes. So see, if you think about ISRO's mandate, they have to do everything from building a space station to human spaceflight, launching missions to the moon, Venus, etc. So I think what they're trying to do here is very much like what NASA did previously, that, you know, don't build anything again. Once you've proven the technology, you pass it on to private players like us to proliferate and commercialize because, you know, we are designed better to do that commercialization. And ISRO in turn can go for the stars and much more challenging problems.

Utsav Somani: And you're expecting the full constellation by POSI to be up and running by 2028? Yes. Is there a different timeline?

Gaurav Seth, CEO & Co-founder of PierSight Space: 2029, to be precise, for all of the 32 satellites.

Utsav Somani: OK. And what does it look like in terms of where do you want the business to be when all of this is going up? What are you expecting in terms of contracts, ideal customers? How much will you be making by then?

Gaurav Seth, CEO & Co-founder of PierSight Space: Yeah. So first of all, let's think about how is it done today, right? Oceans are 70 percent of our Earth. So right now, a country like India spends close to about a billion dollars on monitoring its EZs. How do they do that? By sending helicopters, UAVs, aircrafts. Same with several other countries spend billions. But still we, you know, see attacks like Taj happening, a Taj attack happening, a lot of oil spills, etc. What I mean to say is the data is still not pervasive enough, right? As satellites, we have a natural advantage. A satellite moves at about 8 kilometers per second in space. And with the instantaneous view of about 100 kilometers, we can view 800 square kilometers every second. No amount of aircraft, drone can ever match us. So the vision here is to bring all of the surveillance to a control room so that a helicopter or a UAV can be used to catch a bad actor rather than survey that area. And, you know, we see a huge time.

Utsav Somani: Sorry to interrupt you, but data plus intelligence is what you're selling to them? Like, will the system or your system be able to identify that this is a threat and threat level and assign to it?

Gaurav Seth, CEO & Co-founder of PierSight Space: Yes, to some extent, we have also built a software platform called Matsya, which can detect oil spills, IUU fishing, as well as dark vessels. But we are also ready to provide this data to someone else who wants to build further intelligence on it and wants to do a GTM in that particular country. Right. So it's not that our data will only be consumed by us, but as you know, for runners in the sector, nobody else has seen maritime data getting flooded the way we will do it. So that's why we also need to ensure that a whole food chain is completed and we provide insights to the users. But we are also open for other providers to ingest our data and build further models and intelligence on top of it.

Dhruv Sharma: I was just thinking how cool it is that you guys actually, in a sense, you are capable of doing deep ocean exploration from space. Like it's such an oxymoron.

Utsav Somani: What is the depth you can go until?

Gaurav Seth, CEO & Co-founder of PierSight Space: So we cannot penetrate the surface because you know how the technology works is water has a very high dielectric constant. So we can just see till the surface. However, in a certain large submarines, when they are in the shallow subsurface, first 15 meters, the release wakes like how speedboats release. So we can detect them in the shallow subsurface, large things like a submarine. But the technology cannot see below water.

Utsav Somani: And why choose Ahmedabad over Hyderabad, Bangalore? I'm sure you had to answer this question a lot.

Gaurav Seth, CEO & Co-founder of PierSight Space: I love this. I love this question. So I grew up in a locality in Ahmedabad, which is called Satellite. And that's the easiest answer I provide. But the locality is called Satellite. Yes. The area where I grew up is called Satellite. It was named there because ISRO was started in Ahmedabad due to Dr. Vikram Sarabhai. And one of the major center of ISRO Space Application Center is located in Ahmedabad, because of which that locality is called Satellite. If you ever come to Ahmedabad and I grew up there and all of the SAR satellites which were built in India were built from Ahmedabad because that is the center which builds all of the payloads. The Bangalore Center builds the spacecraft. Thiruvananthapuram builds the rocket. So that's why we have a lot of SAR focused talent around Ahmedabad. And also the regulators, space regulator in space is based out of Ahmedabad. So something very good they did was that in 2021, after starting up of InSpace, they created a technical center which provides clean room, vibration, softwares. Basically, if you want to start a space company in 2018, you'll have to do a capex of $10 million even to build a 3U satellite. Thus, you know, in under $10 million of funding in the seed stage, we've raised total of $10 million so far. We've been able to launch an in-orbit demo satellite, build a product drone based SAR and also now we'll be able to launch our commercial satellite. So all of this wouldn't have been possible if we would have spent all of that capital on capex. So this was another huge advantage we got. And lastly, as I said, I am from Ahmedabad that definitely played a major role in that decision.

Utsav Somani: Have you seen that TV show called Rocket Boys?

Gaurav Seth, CEO & Co-founder of PierSight Space: Yes. Love it.

Utsav Somani: Fascinating. What a story. All right. And that's it for me. Dhruv, anything to wrap us up or?

Dhruv Sharma: Yeah, let's ask Gaurav one closing question. Gaurav, what is your message and Pearsight's message to India's next generation of space entrepreneurs and space scientists?

Gaurav Seth, CEO & Co-founder of PierSight Space: I mean, think big. There's nothing which is impossible. When I was starting this, you know, I was told there's no money or appetite for this kind of a business in India. But, you know, as time has passed, we are just seeing more and more interest. And the kind of interest we receive from talent to work into the space sector is unprecedented. Like, you know, we have one opening and we see a thousand people apply for it. Everyone wants to, you know, and such good talent because, you know, we see talents across the globe. Indian talent is amongst the top, I must say. So, yeah, think big, dare for it. Nothing is impossible.

Utsav Somani: Let's go. Let's go. Thank you so much, Gaurav, for coming on the show.

Utsav Somani: Really fun. All right, listeners, let's move on to our next guest, Reva from Dodge AI, who's building agents to solve this multi-billion dollar industry that's trying to remove issues from our ERPs. Reva, welcome to the show.

Rebhav Bharadwaj, CEO & Co-founder of Dodge AI: Hi. Hi, this is Rao. Hi, Dhruv.

Utsav Somani: Did I get the explanation right or did I do something wrong?

Rebhav Bharadwaj, CEO & Co-founder of Dodge AI: I think I was briefly right. Maybe I could chime in on it as well.

Dhruv Sharma: The only way to have made a mistake here would have been to say this is Reva from the Department of Government Efficiency, DODE.

Utsav Somani: Dude, we're investing a hundred billion dollars. Like, I mean, people actually spend a hundred billion dollars on SAP maintenance and like error correction. That is just fascinating.

Rebhav Bharadwaj, CEO & Co-founder of Dodge AI: Yeah, you're absolutely right. So just to sort of summarize on what we do, right. So as companies become bigger, they have to sort of maintain and implement these enterprise applications. So that would be SAP, Oracle, Kinaxis, Coupa. So different parts of the org for procurement. They have a software for like payroll. They have a software, right. So it turns out after you implement it, there's also upkeep. So that could be just solving simple incidents that come up day to day. I'm not able to log in. I'm not able to, you know, add a user. I'm not able to post this purchase order, right. And that could be some enhancements that come with it as well. So that in cases of like, hey, the business is evolving. We need to add another warehouse, another plant, right. So that this broadly the space comes under sort of ERP maintenance, traditionally done by IT services companies like TCS, Wipro, Accenture, right. Where their solution has been just throwing people at the problem. So they're aware they have gone and hired, mass hired from colleges in India and where they deploy people to these projects, right. So what we are really doing is saying, hey, now with agents, can you take a different approach where we are agents do 90% of the work and humans are just verifiers of the upkeep of the system. So that's to summarize, we are two parts. One is just a platform, which automates a lot of maintenance work. And second is our team of consultants who can also verify what the agent is doing. So we can think of us as a full stack IT service provider in that sense.

Utsav Somani: What's your personal motivation? Like, I mean, you were selling pharma software at 17, door to door, and now you're building AI agents for SAP. Like how did you land on this problem?

Rebhav Bharadwaj, CEO & Co-founder of Dodge AI: Yeah. So I think I was slightly tricked into it and slightly born into it as well. So I do come from like an IT services background as well and we used to run something very similar, but yeah, you're right. Me and my co-founder used to essentially sell AI AP agents very early around 22, 23. And one thing, and we had some customers like L'Oreal and one thing that we realized was, hey, a lot of these abstraction layers that we're actually trying to solve for is still far away where the fundamental systems that they're being built on, for example, SAP or Oracle Fusion Cloud, right? They have not even made the transition to cloud where a lot of these are still on-prem than to say that now we could go ahead and do agents on top of them. So from a motivation standpoint, it's really, we look at ourselves as like distributors of AGI, but can you bring, you know, except for the SF tech people, can you bring the rest of the world to a system which could actually have something like OpenClaw inside, right? So that's one way where we look at it as like bringing the systems up to speed, but AGI could be sort of transferred into it.

Dhruv Sharma: Let me ask you about this question. I've often thought that the SAP implementation and maintenance industry is maybe bigger than SAP itself. Is that true? Do you think that is, do you think that might be the case?

Rebhav Bharadwaj, CEO & Co-founder of Dodge AI: Yeah, that's actually true. So I think for every $1 spent on like an SAP license, so SAP user license is around $6 spent on implementation and upkeep, right? So maintenance. So that's probably again, because the way the industry has been defined is what could be a SAP solution or an ERP implementation for like, let's say QN or any other organization could be fundamentally different from a Google or fundamentally different from a General Motors, right? Because they're very different companies. So a lot of, it has become a requirements gathering slash customization problem over the years where there's a backbone software of something like SAP, but then there's actually a lot of people around it who are just trying to make it work for you particularly.

Dhruv Sharma: But let's try and understand what a DOJ agent is like in production. So, you know, from the time that let's say a ticket is raised to the time that that incident is resolved, what's the agent doing? When are humans intervening? Give us a sense of that entire journey, please.

Utsav Somani: While you were talking about it, a question came to my mind. Like, is it one of the most hated softwares? Like if SAP were to do an NPS score, like what do you think it would be?

Rebhav Bharadwaj, CEO & Co-founder of Dodge AI: Yeah, that wouldn't be super high. But it's sort of, is it designed to be, add a little bit of friction, right? Especially when you're in the large enterprises, each SAP software has three different instances to it. So that's sort of a dev, like a QA and a prod instance, right? And the reason why they have friction is every change on every particular thing that you do actually impacts the entire organization. So imagine you're sitting in Pepsi and you're putting a purchase order. You need to have those approval levels. You need to have those certain levels of cadence and governance to it, right? And, but going back to Dhruv's question, what we essentially do is every time a ticket is now getting locked, let's say a user has a trouble, maybe posting a purchase order, or they have a problem that they want to add new warehouse to the system, the traditional way that they've been doing it is going to ServiceNow or any of these ITSM softwares and logging a ticket saying, Hey, this is what I want. This is my description. This is screenshot of the problem. And traditionally after that, an outsourced partner like TCS picks it up and talks to the person who locked the ticket, right? So it's very similarly, we also connect to ServiceNow or Jira or any of these softwares from there, our agent picks it up, has a conversation with the user, tries to understand where the issue results from, and then the background while it's talking to the user, it goes and picks all these implementation artifacts that have been built over the years. So this sort of custom code that has been brought up and it sort of marries these two together to figure out what the issues and then go and actually resolve that issue using a browser agent.

Utsav Somani: And we read that SAP is also coming out with their own agents, Joule Studio, later this year. Will that, I mean, sort of bring some of that spend back to the SAP balance sheet?

Rebhav Bharadwaj, CEO & Co-founder of Dodge AI: Very interesting question. I think some parts of AI for the business user end is definitely something that Joule would do. For example, maybe a Joule procurement agent could help, like someone, you know, posting a PR, right? Where we think we come in as sort of the entire enterprise IT world. So oftentimes companies are not just SAP shops, right? They have multiple instances in place. So it's very likely that they have SAP with Databricks, Salesforce, a couple of other softwares intertwined together for working particularly for them. So my guess is there will still be a need for system integrators and sort of enterprise, horizontal enterprise IT companies. And things like Joule Studio help the business user with the co-pilot.

Dhruv Sharma: On average, how many systems do you find yourself weaving or stitching together in the enterprise?

Rebhav Bharadwaj, CEO & Co-founder of Dodge AI: That's an interesting question. So very often than not, they have, we have seen cases where they have had SAP as your main system, but for example, we are working with this one, Fortune 20. And they have the super legacy software for running their jobs. And then they have Dynamics 365 as their field service software. So every distributor of theirs use Microsoft FSM, and then they have Salesforce for their sales people. So you can sort of think any company which is decently scaled uses something like SAP or Oracle as their back office software, but different parts of the org still turn in different, different things. So an average you'll see five to six system of records per company.

Dhruv Sharma: Are you guys using the MCP approach for integration?

Rebhav Bharadwaj, CEO & Co-founder of Dodge AI: No. So this is coming back to my earlier point about SAP being super, super primitive. We have done it something more closer to the CLI. So we connect very similar to how something like an IDE would connect with SAP, right? So reason being SAP is not a cloud native software. So they do not have all these native integrations that something like a Jira or a Salesforce would have. A lot of it still has to be custom. So we have built up custom connectors to the CLI itself to work with SAP.

Utsav Somani: So recent milestone in your journey has been Accel Atoms in the Google Future Fund, the AI Future Fund. So how was that process for you?

Rebhav Bharadwaj, CEO & Co-founder of Dodge AI: Yeah, it's pretty fun. I think Accel and both Accel and the Google team, ever since we started working together have been super great. And in very fundamentally different ways as well, right? Well, Accel brings sort of the know-how and, you know, startup experience of seeing these journeys multiple times. I think Google brings you a sense of just knowing what's happening in AI. So we get to work very closely with the DeepMind team and understand, hey, are our internal benchmarks accurate? Can we do something about them to make them better, be ahead of the market in different ways, and also try to get the senses of using those models much more on a larger scale, right? Just from a data engineering point of view. So it's been super interesting so far to have both of them and have both of them in the conversation saying, hey, one of them is trying to think about scale and one of them is in the early stage journey.

Utsav Somani: And you're using a small language model approach.

Rebhav Bharadwaj, CEO & Co-founder of Dodge AI: Sort of an orchestration, yes. So you can think about it in three, like a three-layered sandwich. So the first layer is being sort of orchestrated on our side where that's specifically built for ABAP related tasks. So other than that, for more larger, more generic queries, we do use normal LLMs as well. So the first layer is custom built on a VPC for a particular company. So let's say we're working with Pepsi tomorrow, that's where the SLM comes in, where we'll have to work with their particular ABAP harnesses. Second would be a more generic approach and third would be just our harness on the company point of view.

Dhruv Sharma: You know, Riva, one of the things we like to do in the show is break down headlines and I don't have a specific headline, but I'll give you a sense of what we'll typically see, right? Like headline A will be X startup raises a hundred million dollars to disrupt the consulting industry. Headline B will be, you know, X ID services company lays off 10,000, 20,000 people. Headline C will be like, you know, like this consulting company records like their highest revenue quarter ever. And so it's all very conflicting and hard to make sense of maybe the truth is somewhere in the middle. What, like, what's your sense? What's your take both from within and outside the SF tech bubble?

Rebhav Bharadwaj, CEO & Co-founder of Dodge AI: Yeah, it's around 4 a.m. right now.

Utsav Somani: Are you going to bed or are you waking up right now? I will be going to bed. Living that founder life.

Rebhav Bharadwaj, CEO & Co-founder of Dodge AI: Yeah, sure. Only a question, Drew, I couldn't quite catch that. Were you mentioning something about the X services industry headline?

Dhruv Sharma: What is going on? We're hearing a little bit of everything, right? Startups raising up like large piles of cash to kill that industry, that industry, you know, at least giving forward guidance and record breaking quarters and profits, and then also mass layoffs, like everything is happening all at once. But what's it looking like for the ground? You're literally in the, in the weeds. Yes.

Utsav Somani: Yeah. Our listeners are typically founders, people interested in the startup ecosystem for them, I think it might be interesting to know what is truly happening on ground. We've seen all the headlines, like what Drew was pointing out on X, but I think what's truly happening on ground, like how are you seeing the AI ecosystem evolve, like give us the stuff behind the enthusiasm.

Rebhav Bharadwaj, CEO & Co-founder of Dodge AI: There's a mix of paranoia and like genuine sort of what people are calling EGIP illness, right? So some of the headlines are just driven by paranoia, a field where it's a bit too forward looking, where they see what the models are right now, the steady state may be in their personal PCs or simple tasks, and they're sort of extrapolating it within large jobs. And that becomes a lot of the, you know, layoffs and there's more forward looking and from the quarters and like findings point of view, those are again, backward looking, right? So there's a mix of paranoia and genuine enthusiasm, I feel, but the truth, I would say is there are only two possibilities. One is, hey, AGI is real and a lot of the work can be automated. Then a lot of this is sort of surmounted. For example, we had this one system integrator that we work with, partner with, who had recently tried out with our product and they were like super surprised on, hey, I don't think I need like 50% of my staff anymore. Well, that is true, but we were, being on the other side, we want to give them that hope, at the same time, we're also like, hey, I think you should do it a bit more responsibly, where while it might look like a lot of this could be automated, we still need, there's a time to evolve and put a harness around these things before they actually hit the market or giving you that 99, 100% reliability.

Utsav Somani: And as a final closing question, and let's go for something more, what do the next 12 months look like for you and Dodge, numbers, anything that's on your vision board? Like what are the next 12 months?

Rebhav Bharadwaj, CEO & Co-founder of Dodge AI: Yes, I think for us, it's probably just tracking down on two things for the next 12 months. One is, can you go into these sort of much larger enterprises, where you could say that, today you're working with something like a TechMind or a TCS, and you're paying them upwards of $30 million for these EMS contracts. And can we take a shovel of this and prove that in the next three years, we're able to bring that number down, the automation to human ratio and completely annul it, right? Let's say, August for the supply chain part of the organization has 100 consultants, which is the upkeep of an SAP system. Our goal is to say that, hey, can we crack down at least 10 of these brands where we were able to bring that number from 100 to like 15 or 20, and the rest 15 are just verifiers, right? And the second and more important part for us is also working with partners itself. We do think a lot of system integrators are much more regional. They have a relationship with their customers. So we also have a strong wholesale model, where we work with sort of the boutique firms of the world, where we go to market with them and say, hey, now it could be an AI native firm, where a lot of the work that was previously done by a back office can be done through the Dodge agents. But, you know, you can still hold the relationship with the customers and do the requirements gathering part. And all the background work done by the consultants are now being done by the agents. So a key part of a go-to-market is also trying to crack down 12 to 15 of these boutique firms.

Dhruv Sharma: Do you think someone else can build a Dodge for Workday, for Salesforce, or is it going to be you guys over a period of time?

Rebhav Bharadwaj, CEO & Co-founder of Dodge AI: I think it's very similar to how the Swiggy, the Zepto type markets, right? You could launch Swiggy in Delhi and Zepto in Bangalore, but over time, one of these guys are going to find success in one of these markets quicker and want to expand across. Just because of the nature of how close the solution is from a customer or the buyer persona point of view, I think it's very likely that the first person who's able to get that first 10 to 20 enterprises on board is very likely to cross sell than to say that, Hey, you need a completely different company for Salesforce, or you need a completely different company for, you know, Workday. But I do think there are companies in the panelists right now as well.

Utsav Somani: Awesome. Rewolf, thank you so much for coming on the show. Get some sleep. Awesome. See you guys. Cheers. All right, folks. Thanks for tuning in. We'll see you on Wednesday at four o'clock. Bye-bye.