Become a Creator today!Start creating today - Share your story with the world!
Start for free
00:00:00
00:00:01
From building Cars24 to creating an AI Native Hiring-Tech Platform: Avinav Nigam (TERN Group)
63 Plays11 hours ago
How did a founder who sold millions of cars and managed billions in real estate become the CEO competing with Palantir for government healthcare contracts?
In this episode, we explore Avinav Nigam's journey from Cars24 to building TERN Group, an AI-powered healthcare recruitment platform solving the global nursing shortage.
Avinav Nigam has built companies worth over $10 billion in enterprise value across three different industries - and his latest venture might be his most ambitious yet. After co-founding Cars24 (India's used car unicorn) and IMMO Capital (a European real estate investment platform), Avinav is now tackling the global healthcare crisis with TERN Group, an AI-native recruitment platform that's already generating ₹250 crores ARR in just 24 months.
What makes this story remarkable isn't just the impressive metrics - 40% month-on-month growth, 88% conversion rates, and competing with enterprise giants like Palantir for government contracts - but the deeply personal mission driving it. After witnessing his colleague's death due to healthcare system failures and seeing talented Indian nurses exploited by unethical recruitment agents, Avinav built a zero-cost platform that charges employers instead of workers.
In this candid conversation with host Akshay Datt, he reveals how AI is revolutionizing healthcare recruitment, why he believes "AI replaces humans who don't use AI," and his vision for becoming the future of human mobility across borders.
Key Highlights
👉How Avinav built three successful companies across cars, real estate, and healthcare - reaching $10B+ enterprise value
👉TERN Group's explosive growth: ₹250 crores ARR, 40% MoM growth, and 88% conversion rates in 24 months
👉Why charging employers instead of healthcare workers created a more profitable and ethical business model
👉 How AI-powered interviews assess 80 nursing specializations across 28 different attributes with human-level accuracy
👉 The global healthcare worker shortage crisis: 15 million workers needed by 2030 and how TERN is solving it
👉 Competing with Palantir and other enterprise giants for billion-dollar government healthcare contracts
#CrossBorderBusiness #AIFirstCompany #HealthcareWorkforce #TalentMobility #EthicalRecruitment #MarketplaceModel #RealEstateInvesting #StartupJourney #FutureOfWork #GlobalTalentShortage #B2GContracts #CompoundedCompanies #DigitalTransformation #InternationalExpansion
Disclaimer: The views expressed are those of the speaker, not necessarily the channel
Recommended
Transcript
Introduction to Cross-Border Business and Consumer Decisions
00:00:00
Speaker
Today, I hope to learn from you about two things. One is how to build cross-border businesses. And second is when a consumer is making a life-changing decision, how do you leverage that opportunity? What we're trying to build is a generational company, right? We're trying to build the future of how human mobility happens on
Avinav Nigam: Co-founder of Immo Capital and Entrepreneurial Journey
00:00:16
Speaker
the planet.
00:00:16
Speaker
Avinav Nigam is the co-founder of Immo Capital, which manages $2 billion dollars in real estate assets, and now the founder of Turn Group, an AI-native platform helping Indian talent find jobs overseas.
00:00:28
Speaker
We manage around a couple of billion dollars now in in assets across across Europe, working with some of the largest investors on the planet. Aren't you doing too much for a 24-month-old company?
00:00:51
Speaker
Avina, welcome to the Founder Thesis podcast. Today, I hope to learn from you about two things. One is how to build cross-border businesses. And second is ah when a consumer is making a life-changing decision, how do you leverage that opportunity?
00:01:07
Speaker
Why are you the right person to teach me this? No, thank you, Akshay. Super excited to be here, first of all, and and and really, really thank you for having me. um Look, I have, this is probably the third company I'm building now in this space, one in the car space to be able to buy a used car or car for An aspiring Indian mid-class person is is is a life-changing decision.
Challenges in Major Life Decisions: Cars, Houses, and Moving Countries
00:01:32
Speaker
It's a very hard, very difficult decision. It can often be a year, two years, five years in the making.
00:01:36
Speaker
ah Same thing about buying a house um tends to be an extremely complex, very challenging decision as well. And, you know, moving countries for a job. it's It's probably one of the most harrowing experience I've been through as well and learned through. So I've built companies in each of these three.
00:01:52
Speaker
um i've you know As a total enterprise value, you know my companies have you know grown over $10 billion. dollars There's a lot to hopefully, a little bit to learn from. For myself, it's always a new learning, but hopefully some things I can share.
00:02:06
Speaker
Amazing. Let's begin. Take me through the journey first. ah What was the first venture that you built and how did you end up over there? course. um I will bore you a little bit with background as well, I think, because it all connects back to why I build things and and and and what I've built.
00:02:25
Speaker
um you know as ah as ah as an as ah as ah As a confused Indian kid, um I grew up between the UK and India. my my My father was... I was also the head of Indian Railways, so head of HR, sorry, for the Indian Railways. so ah So I've seen large public sector workforce systems. I've sort of migrated nine times. I've sort of moved within the country. I've lived in Assam, I lived in Chennai, I've lived in Gujarat, lived in, you know, Delhi, in, you know, Bhavnagar, small towns, Vadodara, lived every everywhere.
00:02:54
Speaker
um I've lived in 90s Manchester, which was very, very rough, if you if you will. I've seen, you know, two cars being stolen in front of my eyes. from for for you know dad
Early Career: From Procter & Gamble to Entrepreneurial Ventures
00:03:03
Speaker
running after them. i've seen I've seen a lot of very interesting things, but um I then um went to IIT in Bombay um to do my engineering, studied chemical, and then took my first job, um as crazy as it sounds, my first job in in rural Japan.
00:03:20
Speaker
ah So I started with a company called Procter & Gamble, which is an American MNC, but they obviously had a big Japan business. um I went into a production plant with 300 Japanese people and one Indian guy.
00:03:32
Speaker
So that was my life. Japan is very hard for people who don't know the language, and I'm sure you didn't have Google Translate back then. There was nothing there. There was no AI, there was no life live translation, nothing.
00:03:45
Speaker
um so you know, it's one of the most challenging but most enriching experiences of my life, I think. ah to be able to really integrate and assimilate and learn the language. I speak Japanese fluently now.
00:03:56
Speaker
um For several years, it was useful, and then it was not useful, and it's useful again. I'll explain why as part of building turn, it's become useful again. So I spent several years in in Japan and then Singapore. I ran the Asia business for Procter & Gamble in in detergent, so Ariel, the brand.
00:04:12
Speaker
So ran that out of Singapore, then moved back to India to build companies. So this is sort of comes to the first one. So joined a friend of mine that were building out this company in India um in the car space. So essentially the model was, um it's it's it's it's a C2B model. So, know,
00:04:34
Speaker
The company is called Cars24. it's It's sort of been around the block, if you will. um The company has obviously done exceptionally well, continues to do well, really big believer in the mission.
00:04:46
Speaker
But it's a C2B model. Essentially, you've got a fragmented supply pool of people looking to sell their used cars. um It's a very difficult transaction because you know buying and selling a used car or a car in India, um you save up a lot of money. You're moving from scooters and to two to you know to to cars. And it's really a representative of the of the of the Indian economy and and the coming of of the middle class.
00:05:11
Speaker
um So they would sell it, but to be able to sell a car, you go to a dealer and there's a lot of complexity in the documentation. There's a lot of ah belief or distrust rather um that somebody is taking advantage of me for the price.
00:05:26
Speaker
There's a lot of, you know the duration tends to be very long as well in the RC transfer and so on. So there's a lot of these. Similar challenges you'll you'll notice across the three companies that are built. And then once we've bought the car from them, we sort sell it on to car dealers or car companies. So it's an instant transaction um for something that could be extremely stressful ah for a family to be doing.
00:05:45
Speaker
So the business obviously took off. massively. I was there until we were doing around 250 million revenues from zero. I remember the time sort of in a Gurgam petrol pump, we had the first first store and we would we would go up to people trying to fill their petrol and ask them if they wanted to sell their car and they would look at us like, are you stupid?
00:06:03
Speaker
like Or they would look at us and think, are you are you are you trying to steal my car from me? like So you know that is the moment where it all started and and today obviously across 600 locations,
00:06:14
Speaker
across the world, India, Australia, Southeast Asia, and Middle East. So it's a phenomenal journey, lots of learning ah for myself personally, um, as well. This is like a marketplace and and you have been building marketplaces since then. uh,
00:06:27
Speaker
ah Is this marketplace different from, say, a Flipkart, which is selling products versus this marketplace, which is first buying from consumers and the sale might be happening in bulk also. I don't know who the sale happens to, but does it have different characteristics?
00:06:46
Speaker
Oh, very different.
Marketplace Models and the Success of Cars24
00:06:47
Speaker
I'm glad you pointed that out. You've actually hit the nail on the head, right? So there's three broad differences, I would argue, versus a a typical retail marketplace or a if you will, even ah even ah a food marketplace and and and and labor marketplaces in in in hyperlocal.
00:07:05
Speaker
um So the the the key differences are, I think number one is is, as you mentioned, in our case, what we're doing is we were essentially, we realized that the biggest constraint in this category is supply. So the one company that controls the supply or the one player that's able to aggregate and consolidate the supply is the player that we bring the category, right? So essentially the supply is coming from a very fragmented universe of of sellers.
00:07:26
Speaker
So the first thing we had to change was the psychology from the for the customer saying, every brand is coming to me and saying, hey, buy this from me. Hey, buy that from me. Hey, use this of mine. But what we had to end up doing was essentially reverse the psychology for the customer to say, hey, why don't you sell us your car?
00:07:46
Speaker
um It's not just an odd transaction in the sense that a company is asking you to sell. It's also a very big life decision. So the second thing that tends to happen in these categories is what we call latent demand or latent supply.
00:08:01
Speaker
So there could be... 15 months before the actual transaction happening, the customer has started to think about selling or getting rid of their car. And the reason would be some issue pops up, the mechanic says you need to change, you know, this particular piece in the engine, or there's a little bit of a bump, or you know, it's crossed that four year mark where the price starts to drop, right? that's when you want to get rid of it. So there's lot of these factors that start to creep in or a life stage change happens, etc, e etc, right?
00:08:29
Speaker
um That's the second thing is the latest. So you need to almost tap into the consumer psychology almost 12 to 18 months before that actual transaction happens. So think of it a very long lead time into the final conversion. So the funnel is is is super in that way slow, um but also super important for you to be present in those moments.
00:08:47
Speaker
That's, I guess, the second difference in the characteristic. third difference is that The dynamic on the consumer side is very different from the dynamic on the C to B side. So we had a consumer side, we had a business side.
00:09:00
Speaker
um On the business side, what they were looking for, these are car dealers or fleet companies or rental companies, et cetera, et cetera, that we're looking to buy. For them, their biggest pain point was that I now have um a limited amount of space and limited limited amount of working capital, right?
00:09:17
Speaker
So if you are able to find a car for me, but I have a lot of buyers for those cars. If you are able to, on my phone, give me access to an infinite amount of inventory, an infinite amount of you know supply, suddenly i and can immediately sell a car to somebody who comes walks through the door because I am guaranteed on the back end that there will be supply ah to sell. Without create is it creating a crazy amount of working capital problems for me, which is the biggest issue in Indian SMB, right? So you're solving two very different problems. The third piece, which is probably similar to retail ah in some ways is is obviously inventory management tends to be one of the biggest issues
00:09:58
Speaker
ah success factors in this. So if you're holding stock as CAS24 for a long enough period of time, the asset is a depreciating asset, it's not an appreciating asset. So you also want to really, really, really ah tune down of the the model to be be minimal.
00:10:14
Speaker
Similarly, in retail, you also have shelf life of products, but I guess the reverse reason for it. um Here, the car's not necessarily there, but maybe it's dying. It's depreciating. So I guess there's a few differences, but a couple of similarities as well.
00:10:26
Speaker
um You know, transaction size is larger. There's a lot of Post transaction work as well in this category. Right. So once the transaction is done, there's the RC transfer, documentation, insurance, all of that needs to happen.
00:10:37
Speaker
um So I would argue a lot more diligence and a lot more documentation heavy as categories compared to, let's say, ah buying buying um ah buying a shoe on Flipkart. Right. So those are, I would say, the few differences versus the two two marketplaces.
00:10:52
Speaker
but Okay, interesting. um Correct me if I'm wrong. So, Spinney would be selling more to consumers. That's why, for example, they would invest in Sachin Tendulkar as a brand ambassador, not to get more supply, but to get more demand, get more buyers in.
00:11:07
Speaker
so So, their model is fundamentally different. Like Cars24 would be selling more to dealers, cab companies, and so on. Yeah, so I would say, um ah and and without sharing sort of exact data, I'm not privileged to share that much information, but it started, as I said, when I was there as a C2B model. um I think we very quickly expanded into a C2C model as well. So essentially, because the stock was available, because you had access to ah probably the most a constrained ah part of the sector, which was supply, you essentially then started to create the C2C brand as well. And that, I think, has also ah gone exceptionally well ah from that point on So in a way, I would say Spinny and Cars ah would be um in ah in a similar space. I do believe the one that will control the supply will control the category. And at this point, it seems like Cars24 is doing really well there.
00:11:56
Speaker
OK, interesting. So here, like, supply decides who wins in this kind of a... This is a characteristic of, I guess, all secondhand kind of marketplaces, right? Like, say, phones and so on as well.
00:12:11
Speaker
ah Though... none of the other spaces, uh, I don't think they are investing in selling to consumers that much. Like say cashify also, which is into second hand phones, they sell to dealers, uh,
00:12:27
Speaker
And of other such companies in the second and fourth market are more focused on supply, like supplies where they're fighting rather than building demand. Absolutely. So I think it really depends on how you see the sectors evolve. So I have the luxury of being able to see this globally. So I'm sitting in London, sort of in Dubai, in the US.
00:12:46
Speaker
and There's a few examples of, for example, you mentioned Cashify. There's a company called Back Market in Europe that has actually built out a ah supply first play in gadgets, largely phones and and laptops and and the likes and smart watches, they've sort of now pivoted or or rather they are now ah for the first, I would say four to five years of the journey, their entire focus was supply building and then selling to to to to resellers.
00:13:10
Speaker
um And now the majority of the focus is actually on on going into C2C as well. So selling back to the consumer as well. So I think it really depends on when you feel you've got enough of a flywheel around supply building and capacity, and you've created a brand of a trusted ah buyer.
00:13:27
Speaker
um That is when I think you actually have the halo that you can use towards the C2C side as well. I think it's it's a bit um challenging to build a two-sided um consumer brands for categories that have infrequent issues.
00:13:40
Speaker
sort of transaction cycles. um So I think you need to build one brand. The choice in this case usually tends to be supply first. Once that's built, you can use the halo, you can use the the similar marketing dollars to build ah the C2C sales, you know, the selling side as well ah back to the consumer um because the CAC obviously is is a bit higher than acquiring businesses that are doing reselling. So I think that's where I think it's a natural evolution. So I'm sure it'll happen as well um at the right point.
00:14:08
Speaker
Interesting. Very interesting. ah Are these businesses ah typically holding the inventory on their books? Or is it just passed through? So this is very interesting. So what you what you mentioned is very interesting. I think back in the day, what we essentially built, and I'm pretty sure it still continues with cars as well, um is this is this live auction model.
00:14:29
Speaker
um And what it does is, while you're sitting in the store, ah sipping chai, you know your car data is almost 140 plus data points that would set get captured ah digitally, um you know video, everything, you know the the onboard OBD from the car as well, everything would electric electrical would get captured and then published into an automated report.
00:14:51
Speaker
Now I think we're using a lot more AI as well to assess damage and understand real issues. um While you're sitting sipping chai, the the the the auction platform has already kicked off, right? So essentially the system already knows and is able to predict with a very high degree of certainty at what price that car is going to get sold at.
00:15:10
Speaker
And at the time i was part of Cars24, fortunately we were able to also have an auction board where the sale would have already gone through, right? So a pre-sale would have gone through. So you knew the first five people out of thousands of of of of dealers that wanted to buy the car.
00:15:27
Speaker
And I think that gave you a certain amount of assurance that the car was essentially sold before it was bought. So I think going down from you know ah several months of inventory, um you know that is is is true in this case for any traditional car dealer, down to ah days and almost being instant um I think that was really the way to win um the the, you know, because working capital is heavy, right? this is a This is a big amount of investment per transaction is quite large.
00:15:53
Speaker
So if if you can bring it down to an instantaneous transaction, the more you have the likelihood of winning.
00:16:00
Speaker
So there is also an element of some vehicles giving you better margins because you're able to bring them close to a new car, whereas other vehicles, you can never bring them close to a new car, so those will have poorer margins.
00:16:13
Speaker
Exactly, exactly. there's There's a bunch of factors, right? there There's like in in in in our real estate business also, there was like, we were tracking almost 250 variables um on on pricing for assets. in In cars as well, there's, as mentioned, over 140 variables.
00:16:29
Speaker
um I think there's a few big ones, right? So the brand tends to be important. There's certain residual value of brands um continues to be in India and globally, actually, there tends to be certain... models within that brand that do better.
00:16:41
Speaker
In certain years, as I mentioned, once it's beyond four years, ah the price drop tends to be more significant because that's where ah typically find people find issues propping up cropping up and ah the insurance sort of runs out and then people don't really renew and and so things sort of change. um beyond beyond that four year or seven year warranty.
00:16:59
Speaker
ah you that's That's the period you want to be mindful of. Then you know other factors tend to be the mileage. How much has it actually been driven? Has it been in any serious accident or not? Like, et cetera, et cetera, et cetera. So I would say those tend to be. And then I said, unfortunately, because consumers, consumers are very emotional buyers. The cosmetics do make a difference.
00:17:18
Speaker
um That's what I would argue people should look past and actually look at the the bones need to be good, right? If that makes sense. What did you do next? Amazing. So look, I um post the cars journey, which was phenomenal, and really enjoyed myself, really you know built out, I think, hopefully a brand that endures or ah for generations.
IMO Capital: Innovating in Housing Asset Classes in Europe
00:17:39
Speaker
um we we we For the birth of our first child, we moved back to the UK. So I set up my ah next company in a very similar C2B model. So essentially a fragmented supply in real estate. So housing, residential real estate.
00:17:55
Speaker
And we were working with institutions like ah private equity, family offices, insurance companies, pension funds, um basically response socially responsible investors that were looking to invest into European real estate at scale.
00:18:11
Speaker
and create a residential or commercial residential residential so the pain point there it's largely residential the pain point there was unlike um markets like India or Dubai where there's a lot of construction development happening a lot of the cities have grown organically in places like Germany Spain the UK where we were operating um and there's there's very little housing stock you have access to right as an investor um and and more so it's very attractive because this is one of the lowest beta which is risk um asset classes and one of the highest alpha right return asset classes so I think risk could risk adjusted returns this is probably at par with or close to equities across all asset groups which is why it's an extremely interesting investment but the challenge is um the US has has has created this
00:19:01
Speaker
ah asset class called sfr single family residential, but Europe had not. So essentially that's what we set out to build was to create a new asset class in housing, allowing large institutions to deploy capital into residential real estate at, you know, low interest,
00:19:19
Speaker
ah sorry, low return, but long term holds, which is what creates a very favorable economy, right? You're not trying to you know unlock ah crazy it returns like in Airbnb style um housing, right? So it's long term and and it's it's it's investing into the middle class. It's making sure people have access to better opportunity, a better standard of living and so on. That's what we are building. So again, a C2B model, very similar.
00:19:42
Speaker
Again, very similar characteristics of the category. um You know, it's one of the biggest decisions of your life is to buy or sell a house. um It is a complex documentation, legals.
00:19:55
Speaker
It's a long lead time, often even more than five, you know, three to five years sometimes in planning. um And there's a lot of trust deficit like the car sector because you don't trust brokers. Like, I mean, I don't trust brokers, right?
00:20:07
Speaker
um Personally, you know, I'm sure a lot of people would agree. um and So what we essentially built was, and you know, instantly ah able for consumers to sell their properties to us at good market pricing. We do a sort of evaluation, an instant offer.
00:20:22
Speaker
um And then they don't have to take care of any of the of the paperwork, the legals, the hassle, the payment, the documentation, everything was taken care of. um So that business is called IMO Capital, I-M-M-O.
00:20:34
Speaker
IMO largely means real estate in lots of European languages. As you remember, I grew up between the UK and India. So for me, this was sort of going back. My wife also grew up between Germany and India. So for us, it's always been one of the things that we we we knew we would do at the right and at the right point.
00:20:49
Speaker
But we obviously had large set up in India um as you know as as a smart entrepreneur would to get access to the highest quality talent you possibly can, which they it would the Indian ecosystem offers. So, built out the business over several years.
00:21:06
Speaker
Again, similar to cars, sort of a good outcome. ah We manage around a couple of billion dollars now in in assets across across Europe, um working with some of the largest investors on the planet, trillion dollar asset managers and so on.
00:21:20
Speaker
and And yeah, we raised a bit of venture, i think, did The last round was like 100 million Series B. We didn't really need more money than that because the business sort of is a very high margin business and an AUM business as well, similar to running a fund.
00:21:34
Speaker
um So, yeah, it's it's it's been it's been very, very interesting, I think, building it out over over over seven years. And... The business continues to do really well.
00:21:47
Speaker
They have scaled across Germany, Spain, and the UK. I'm more in an advisory capacity now as opposed to an operating capacity. But that's sort of another one close to my heart, which has which is done quite... The team to make us make me proud.
00:22:04
Speaker
How did you discover this? It's a fairly niche kind of a segment that there are private equity, retirement funds, pension funds, et etc.
00:22:14
Speaker
who want to hold assets and they can hold residential assets and they'll need a company that will help them acquire these assets. like they That's so different from what you were doing previously. How did you discover the space?
00:22:27
Speaker
Yeah, I mean, it's great question. ah Two things. One is look the pain point for, let's say, a Blackstone or a PGM or any of these players, teacher's insurance, that's one of the largest asset managers, right? The challenge for these guys is how do I um deploy 500 million in German real estate? It's just it's not possible today because I said, cities have grown organically. Any development new development opportunity is going to be in like you know tier four locations.
00:22:55
Speaker
um Any existing stock might come online, but be very distressed. right So not easy to easy to take care and and and fix. um and and they come in maybe once in five years. So it's very hard to invest programmatically or at will into these asset classes. The only way to do that is to go direct to consumer and buy one by one by one by one by one. So that's essentially what we built. Very similar to the cars sourcing machine, which I had sort of helped build. If you remember, you're able to build a direct to consumer brand.
00:23:25
Speaker
Hey, We can give you you know an incredible transaction, super smooth, super easy. Everything's taken care of. you know We'll come in, we'll do with the with the app, all the data in. Using the app, we're able to sort of predict the pricing as well, give you a confirmed offer.
00:23:39
Speaker
you know Your documents will be will be will be read automatically as well. Everything's sort of done. um So I think it was a very similar consumer experience, number one. So i think for us, that was... that was um the learning that I had from Cast24 really came in handy.
00:23:52
Speaker
um I think several of the investors that have backed Cast24 have also been a part of the journey at IMO. So there was that you know continuity as well. ah People believed in us, people believed in me and and and the team.
00:24:05
Speaker
um And, yeah, i mean, I ah had common friends with my co-founder um and, you know, we started jamming while I was still in India. um And we actually kick set up the company before I even left for the UK because we were so excited about the opportunity. We got our first client before I landed in London.
00:24:22
Speaker
um But it was just, you know, from there, it was just, as they say, it was amazing. it's it's been it's been a wild ride since that point. on But a lot of it came through similar learnings.
00:24:35
Speaker
Supply-constrained market, cannot access supply, very complex transaction, lots of pain point, probably the one of the largest categories you could play in, right? It's a trillion-dollar market in Europe. So um I think the TAM tech played a big part as well in in the choice for that particular um a category in the asset class.
00:24:53
Speaker
I'm assuming your friend would have had access to the PE and pension fund space to get buyers and generate demand. Exactly. Exactly. So they come from sort of Blackstone and and JPM and you know Goldman. So that's the background. So that's how the connection also worked. I was sort of able to build out the supply um and and the business and they were able to build out the capital, the demand side.
00:25:17
Speaker
You said that... ah The US market has an asset class called SFR, which Europe does not have. um What does it mean to have an asset class? Are you talking of that this is a regulated, recognized asset class? There is some regulatory authority or there is some credit rating agency or something like that. What what does that mean to have an asset class?
00:25:40
Speaker
Because SFR as an asset exists in both these markets, right? We're just saying that for an investor, that asset class exists in US but not in Europe. So that's what trying to understand.
00:25:51
Speaker
Of course, of course. No, look, I think it exists in different forms. um and So in in the US, you would have you'd have your REIT structures, which are your real estate investment trusts, which sort of consolidate them and sort of float in the markets and people can trade them.
00:26:04
Speaker
So that does exist. But specifically and uniquely, you've got a lot of large players. And Blackstone did this post the GFC crash, where there were thousands of of of properties that sort of went, defaulted on the subprime loans. And they went essentially...
00:26:22
Speaker
consolidated them, bought them at discounts and created the largest invitation homes, it's called, one of the largest landlords in the US, over 100,000 homes. um Back then, floated on the markets and that did exceptionally well.
00:26:36
Speaker
um And then post that, i think another 100 players have done a very similar approach. um The challenge with all of these plays, however, is that um you're waiting for a big market correction and you're essentially, I don't mean to say taking advantage of an unfortunate situation, right?
00:26:54
Speaker
Whereas we believed in something very different. That's why we are trying to build more of a socially responsible um ESG heavy play where, you know, the returns were were were were were sustainably good long-term for investors and the hold was both the returns of both capital and income.
00:27:11
Speaker
rental income. um And we can do it programmatically and like unlike the US where it's essentially happened happens basis some corrections in the market. um I think the more recent players have started to use a bit more technology. There's a few players that are doing it now in the US as well.
00:27:25
Speaker
But Europe does not have that, right? So as I mentioned, any large um institution um still finds SFR, unlike the US, in Europe they still find it new and risky even though it's the most, um if you will, most risk adjusted a best to return asset class out there because it's so new, right? So they don't know how to do it. And the difference obviously is Europe, you US you have a little bit more Standardization and in kind of the stock in the country sort of has you know standard regulations across the across the different different states as well. um there's There's a certain level of mobility existing, whereas in Europe you can have it's like each country is its own.
00:28:07
Speaker
regulatory ah challenge. So you also have to sort of be conscious of which ones you're going after. So those are some differences I would say between the US and that's why it's been hard for any player to, you know, private equity and otherwise to really build a large exposure across Europe. And we were uniquely able to provide that across different regions in not just one country.
00:28:26
Speaker
Okay. Interesting. So essentially when you say US has this as an asset class, it means there is lot of institutional capital already or a lot of assets under management already in this asset class.
00:28:37
Speaker
um I am not able to remember the name, but there was a competitor to Zillow which had the similar approach in the US of programmatically buying houses. Open door. Open door. Open door. Open door. Open door. And they shut down, I believe. Yes, they have. I remember reading something about that. I remember Zillow also had a programmatic buying of houses program in the US, which they shut down.
00:29:05
Speaker
yes So ah why did that not work? yeah Yeah. Yeah. Yeah. Why did that not work and why did it work for Emo? Yeah.
Technology and AI in Property Evaluation and Management
00:29:15
Speaker
So I think um the difference is um largely the open door model was um a C2C transaction business.
00:29:23
Speaker
So they were buying directly from consumers and then selling back to consumers. The challenge, as we discussed earlier, going back to the cars example is inventory. um So if you are buying from consumers and you don't have an instant sale already guaranteed on the other side, you're holding stock on something that you don't know what the real value of that asset is.
00:29:41
Speaker
um And in some cases, and I'm not pointing fingers at the company here in Opendoor, but in some cases, what you can do is in the mad rush for growth, you don't really have controls around the actual investment thesis.
00:29:53
Speaker
The way we were operating, I would say very differently is because we would set the criteria with professionals that knew how to invest in real estate that had certain targets for their funds, they knew what they were looking for, and we'd build the strategy with them.
00:30:06
Speaker
And then we'd go crazy within those bounds. So you can scale very fast, right? And we are deploying at one point, half a billion a year sort of freight, right? we're doing very, very fast acquisitions, um almost a billion actually actually at one point.
00:30:19
Speaker
but The thing is that always happened within the bounds of the thesis. So the way we had built it out of, view would scout the market for 8,000 properties, let's say across um the two cities, three cities at that particular point in time that the investor has ah target defined as, and we'd identify the top nine properties in that week.
00:30:39
Speaker
that actually were extremely superior in terms of returns, in terms of stability, in terms of, because we'd be looking at 200 pages of data and we'd be able to identify the risks within that property, whether the elevator has a problem, whether the roof has a problem, whether there's mold issues, we could identify that. So the ability for us to make a decision is significantly more confident because the asset is already sold as the moment it gets bought, if that makes sense.
00:31:08
Speaker
because it's rolling into structure a company structure. That makes all the difference. Right. Yes, that makes sense. I do believe the open door shut down because of inventory problem. The inventory they had lost value and they got wiped out because of that, whereas you did not hold inventory.
00:31:23
Speaker
You would start a search when you got a mandate from a buyer. So it was the other way around. Okay. Very interesting. Okay. um how do ah How do these ah screening and evaluation, how does that happen? You have some assessors who come in and are paid per assessment and that then gets converted into a machine-readable format. or like How does that happen?
00:31:49
Speaker
No. So we actually had people um that were largely um very good at um technology. They were very good at, of course, they had qualifications in real estate surveying. It's called surveying.
00:32:02
Speaker
um but you know people that did not have 40 years of experience, right these are people that had three to seven, eight years of experience that had done technically very strong on on on the on the actual asset assessment, but they were able to use the technology we were offering. They had they knew exactly what the information has to be like. We're collecting, as said, over 250 data points.
00:32:22
Speaker
you know, where's the fireplace in the property? How does the flow of the rooms work? um You know, using the app, they're able to detect, you know, the ceiling heights. um They're able to detect the size of the windows. ah That made a big difference. They were able to detect um the age and the probability of lead of lead and asbestos in the property.
00:32:42
Speaker
They were able to assess um using that, you know, what... Reconfiguration could be done to unlock value as I mentioned in the cars example as well. If you move the kitchen from here to here, it unlocks an additional room, etc. etc there's A lot of that literally happened with the data collection. Everything sort of went back to the app which is driven at that point by machine learning. Now we're sort of using much more AI um to read a lot of the data, both audio, sorry, with both video, ah static and sort of input.
00:33:10
Speaker
Once that data would go in, in real time, within seconds, we were able to actually create an underwrite. ah It's called an underwrite of a property that would take the the same amount of diligence that a Blackstone would do, for example, on $100 million dollar investment, we would be doing for a 100K investment.
00:33:30
Speaker
So it'd be a 20 page report underwrite with comparables, with sort of, you know, all the risks with a debt servicing ratio, everything included ah would be created instantly. And this is why from an investor point of view, they initially started with investing was crazy, we were investing 200, 300K deals, so two, three crore deals on WhatsApp.
00:33:49
Speaker
completely crazy how we started and then i sort of moved into discretionary buying where once they set the parameters they would only review it with us every quarter to understand how the portfolio expansion is going and then make adjustments that became discretionary but buying on behalf of ah large institutions because of how robust the underwriting and the AI ah sort of evaluation methods were.
00:34:10
Speaker
Okay, fascinating. um so You said that this is an AUM business. ah How is it an AUM business? You're buying and you're flipping properties, right? it's ah It's a trading business from what I understand.
00:34:24
Speaker
No, no. So we don't flip. So again, the beauty in this model is, um again, if I'm a if i'm um taking Blackstone example, not because we we love them, of course, but not because um there's no other examples out there, just for simplicity.
00:34:38
Speaker
If I'm a Blackstone, um now I have invested $350 million dollars into German real estate with IMO.
00:34:46
Speaker
I'm going to be handed 1,250 keys. how operate this.
00:34:53
Speaker
have way to that have no way to actually take that on do the renovation refurbishment and do the management with that from a rental point of view. I can't do it. It's impossible.
00:35:06
Speaker
It's a distributed portfolio, right? So that's obviously the one risk they felt existed. Now the beauty is this is actually not a risk because when they understand that they immediately got it is like, if you buy a building with a thousand or whatever, three buildings with a thousand units each, you know, with with a total thousand units.
00:35:25
Speaker
um The problem actually that you carry there is, there's this called concentration risk in real estate, right? Things could change, you know, the the way that the the nearest metro station or tube station is supposed to come up does not happen or permissions get delayed or the building permissions don't come through.
00:35:39
Speaker
or the labor ah sort of shortages start to appear, which is happening, um or yeah you know issues such as the tenancy is poor because you know that there's a new upcoming area. right So it's a very concentration risk.
00:35:52
Speaker
What we actually were able to offer was obviously a distributed, sort of diversified portfolio across the city. And this is, in many ways, a dream for an institution. right An institution like a large pension fund, like a private equity, they're looking for a diversified portfolio always. They're investing into the middle class of, let's say, ah Frankfurt in Germany.
00:36:14
Speaker
ah We were able to give them that, but they can't now take on that diversified port of portfolio and operate So we essentially took it on, did the renovation, um you're able to then, so again asset light, we didn't own the assets here, and then manage it on their behalf, which is where you would essentially get a percentage of the rental income, property management, and long-term gains um in the portfolio as well. So it's a pretty pretty insane, to be honest, business model, 70% plus gross margin, 75% gross margin business. So almost like a SaaS business, now but in in in the real world, if that makes sense.
00:36:47
Speaker
So essentially, like ah Blackstone would put in money into an SPV and you would be managing those assets. And therefore, this was an AUM business. You would be responsible for maintenance, rent collection, finding tenants, the whole nine yards around.
00:37:04
Speaker
like Blackstone is purely looking at it as a financial asset, nothing more. The operations is with you more. Exactly. Exactly. So they don't want to operate it. So we would essentially do the you know the renovation, upgradation,
00:37:17
Speaker
the renting out, the management of the documentation, the councils, all of that all of that you know painful operational work. But what it obviously gave us is the is the mo the defensibility that once you're in with that client, you know it's very hard for them to now find another um person that can do the acquisition and the operations. right So you either get one or the other. You'll get you know brokers or agents and even large players like JLL, CBRE that would be doing the the the brokering um of the assets.
00:37:46
Speaker
But then you have to find somebody else now to do the management of the assets. And given that it's distributed, even though it's beautiful from an investment thesis and investment management point of view, returns are pretty, pretty, pretty ah extremely good.
00:37:57
Speaker
Like we were generating in some assets, in some portfolios above 18 to 20% IRRs. which typically in in European real estate, I mean, if you get six or eight, you're lucky.
00:38:09
Speaker
um So it's very, very good in returns. um For an asset class, that is the most, if you will, risk averse, you know, low risk asset class as well. um And long-term holds like this is exactly what they're looking at. So it's predictable. Think of yourself as a pension fund. I would say even like some of our clients are like very large insurance players, like pension funds, like $1 trillion dollars under management ah for them.
00:38:31
Speaker
They don't, um you know, people people working there don't get fired ah because or they don't get promoted just because they make one percent extra return. They get fired if they lose the money. So for them, really investing into a stay safe, long term asset classes is a wet dream. right And that's what we were able to offer them.
00:38:50
Speaker
Okay, fascinating. And you would be paid ah ah like a margin from what money was being generated or was it a fixed fee? So how how was that structured?
00:39:02
Speaker
Of course, i love the I love the incisive questions here. We were... I will send you the P&L right after this. um No, we were um we were able to we were able to get a percentage of the transaction value. So ah like ah like a brokerage for making sure the acquisition happened.
00:39:20
Speaker
um We were able to generate a percentage of the ah mortgage, as the credit side of it, which is because the way this makes sense is... um at that point and and now slightly less so interest rates were pretty pretty bonkers right there was like 0.6 fixed for 10 years in Germany at 100 LTVs in fact 110 LTCs as well is what we we're getting so from a equity utilization point of view and the leverage is is just it's to the moon ah so lots of really really um uh you know you know huge upside for i would say for clients um so you would get the mortgage side as well then you would get a percentage of the actual renovation so the the asset management fee sorry the asset upgradation fees um then we would get a percentage of the rental income
00:40:09
Speaker
We would get a percentage on top of property management and sort of the actual reporting of it, the the portfolio management. um And in some cases, we will also get carry from a portfolio. And as I mentioned, that's sort of why it's almost like running a fund.
00:40:22
Speaker
um But we would have no cost because the only cost we actually had was our own staff. Everything else is actually paid for by the SPV, by the company. Wow.
00:40:33
Speaker
Fascinating. This is like a extremely cash cash flow, like healthy cash flow kind of a business. There are so many places where you're making margins.
00:40:47
Speaker
Why would you ever want to leave this? I mean, this is like... like yeah this ah this is this was This was the retirement plan, right? This was
Addressing Healthcare Workforce Shortages with Turn Group
00:40:55
Speaker
the plan. i was done. i was done, right? And that's what my wife is like.
00:40:58
Speaker
The reason she hates me, ah well, for many other reasons. One of the reasons she hates me is because, you know, I you know cannot sit still um in in buildings. as i The company's built now. We've got an insane machine. We've got like ah a pipeline of capital that's billions in in in in demand. We've got 150 million in the bank It's like, what are we doing? Let's just keep executing, let's retire.
00:41:20
Speaker
um So, but two things happened. So, one is, My colleague, this is 32-year-old lawyer, Oxford grad, extremely smart. She was like one of the one of the warmest people and one of the smartest people I know.
00:41:35
Speaker
um ah She ah got cancer in London. So she got leukemia, ah stage three, sort of went through um five months of chemo at one of the top hospitals in London.
00:41:48
Speaker
And and she she she she beat the cancer. So she was done. So we thought. um ah but This is 22, late 22, early 23. There's all these doctor strikes happening in the NHS, the nurse strikes and workforce issues and blah, blah, blah. um And she was asked to vacate her bed, hospital bed.
00:42:11
Speaker
Right. And she's asked to go into a social care facility where, you know, there's a different setup, not clinical, but at least there's oversight and there's nurses. and And they didn't have the workforce either.
00:42:23
Speaker
So they said, we can't take you. So she was sent home. Right at the point where she's probably at the lowest that she can possibly be in her life. um This is happening. This is eight to 10 weeks. There was almost no clinical medical sort of oversight um and in a system that's obviously public sector, very different from India, one would argue.
00:42:41
Speaker
um You can pay for service. um She didn't get any. um And two days before it happened, I was i texting with her and It's completely bizarre, right? Out of the blue, we're like all preparing for her to come back to the office and so on. um She basically got a lung infection um and passed away.
00:42:58
Speaker
um So after she beat the cancer, after she was done, after we expected her to come back, um she lost her life to the, to in this case, the NHS. um And I've lived in the UK in the 90s. I've seen what the system was capable of and I've seen what has happened to it.
00:43:13
Speaker
um and and and and And my CTO shared the same story, having worked in Germany at a pre-IPO fintech. He was sort of head of engineering for the largest, one of the largest banks in the u US as well, um and used to work with us at IMO as CTO.
00:43:27
Speaker
um He described his his experience of of, you know, this brought out memories for him because um they lost their baby in the German healthcare care system at one of the top hospitals in Europe, ah the the number one hospital in in Germany and top three in Europe.
00:43:41
Speaker
um because they were waiting in the waiting room for six and a half hours and nobody came and the wife was in pain. And, you know, they just, they just, she, she could have died on the operating table. Like she was, she was in, in, in lost so much blood and, but they couldn't save the baby.
00:43:58
Speaker
And for them, it became, a again, for him also, it's super personal, right? It's just not everybody understands this side of what's going on, right? There's a, there's a, there's a, there's a ah complete, um,
00:44:11
Speaker
an utter ah breakdown of the healthcare systems at this point in time. um and And one of the biggest reasons, and this is sort of what you know really opened up my eyes, like the the the shortage of workers is just is just astronomical.
00:44:25
Speaker
right You're talking about almost 85 million people short in the global north. Out of that, almost 20 million coming from healthcare care and and and old age care and social care. shortages and and it's only getting worse because the economies are getting older and older and older, right?
00:44:40
Speaker
I was just on a job ah talk a call with one of my clients and in Japan now. It's like 50-year-old average age now in Japan. 50 years is the average age in Japan, right?
00:44:51
Speaker
um Replacement rates are down, right? So fertility rates are down. There not enough babies coming. It's 0.9 to 1.3 versus 2.1 to stabilize the population. And the youth that is graduating don't want to do these jobs because these are hard jobs, right? A 75-year-old dementia patient, right?
00:45:09
Speaker
First day on your job, throwing feces at you is not what this generation is signed up for. They would rather park $25,000 in Bitcoin and sit on it. right that's That's very happy with that. All they drive a delivery truck and deliver groceries. They're happy. but like It's just not an aspiration job. It's in the rural areas and small towns where there's and old people. They don't want to go there.
00:45:30
Speaker
So these towns, these villages are getting deserted. right So there's a massive problem on this side. So Germany, the UK, Japan, US, Canada, a lot of it, very, very similar shortages. right At the same time, I sort of came across and I started to, you know we have we have a nanny in London, this girl called girl girl girl from Gujarat in Ahmedabad.
00:45:55
Speaker
She's an operating theater nurse. Her dream was to work in the NHS. The exact same problem I just talked about. And she found an agent to help her get that job.
00:46:07
Speaker
He paid that guy, paid that guy, paid that guy, paid that guy, kept paying that guy. the They had to sell their family house. The father had to sell the house so that they could pay the 50 lakh rupees required to get her to the UK.
00:46:21
Speaker
She comes to the UK and realizes the job is fake. There is no NHS job. And she's screwed because now how do you complain? The moment you complain, your visa invalid because it's currently in a payroll scam. They pay the money, they pay she pays the money back and that keeps the visa going.
00:46:41
Speaker
Complain she's out of a job, she's out of the country, right? Or fake job, whatever. And and this is there's a reason this is happening, right? Because you've now got one of the largest oversupply emerging, right? in a country in In our country of India, 1.4 billion people, 105 million of them are either self unemployed or actually unemployed. And women have essentially opted out of the workforce, 100 million of them have opted out. I'm not even including them right now, right? So it's crazy.
00:47:08
Speaker
So there is this aspiration slash desperation um to be able to find employment. And of course, a lot of good work is happening at a national level, at a state level. We're seeing it. like i'm I'm working very closely with the the ministries and different states.
00:47:23
Speaker
But it's not enough because there's not enough employment to be had for we cannot absorb that many people. So you've got one of the biggest oversupplies emerging. India has, don't know if you know this, like we've doubled our capacity for nurses in terms of, um ah you know, nursing college it seats and medical college it seats in the last five years. We've doubled it.
00:47:43
Speaker
So this is where the problem became very apparent. You've essentially got a fragmented supply of talent in markets like India and Africa, Central Asia, Philippines, Latin America, India being one of the, I would say the the the most well positioned, um but also poorly positioned in a different way.
00:48:01
Speaker
One of the biggest oversupply is emerging while you've got the biggest shortage that we have ever seen in skilled workers on the planet. It's happening at the same time. So the, the the obvious answer here was could I now look at this problem as it all started happening in front of me, you know, the last last period of of of building IMO, like, you know, financial upside aside, this has to be solved.
00:48:28
Speaker
Like there is no chance that, you know, we can allow citizens of of of our country, ah you know, and and to go through the six or seven layers. and And, you know, people are essentially being sent to the front lines in Russia because these kids are so desperate ah to find a job.
00:48:44
Speaker
And they're saying, look, at least if I die, my family is fed. This happening in present day, present India, right? ah There's a reason why movies like Dunkia are being created. So I think the government is doing a very good job in trying to clamp down and actually create more ethical systems. But the reality is informal sector is so large and so um ah driven by profit and driven by money that, and and and the other flip side of this, think about it, you do get a job tomorrow at Google in India and the headhunter comes to you and says, okay, I got you this job, now pay me 25 lakh rupees.
00:49:16
Speaker
Makes no sense. Why is the burden of revenue on the talent? Makes no sense. And this is how the category operates. just So we've essentially flipped the script and said, look, the employers have to pay.
00:49:28
Speaker
right It is not about, so it's again, a C2B model, right, as we described. So constraints apply, demand that is really, really, really desperate, but also slightly disorganized, not really modern um and connecting those tools. So essentially build an AI platform that does everything from the first phone call, um including counseling, by the way, to a WhatsApp document and data gathering, kind credentialing, verification um to the actual 30 to 60 minute video interview.
00:49:55
Speaker
All of it happens on AI.
00:49:59
Speaker
We of course got a team that is supporting in terms of ah counseling, relationship, you know training, language, et cetera. But all of this platform essentially built on AI. So as an employer, let's say NHS Birmingham, right I don't need to go through the six and a half lakh people in the turn database I can just look at you know the eight people that are pediatric nurse qualifications, that have the NMC exam done, that have the language, they have the technicals, they have the visa documents, and they are a 90% plus fit for that particular role, and they have gone through the interview process already.
00:50:34
Speaker
right And I can pick them up and I choose seven of them. And that's what's happening today. right We've got 90%, 88% conversions ah from receive to offer. um And once that's done, the the entire supply chain on logistics of visa, immigration, arrival, post-arrival support up to six months, all of it, housing, credit, if you remember a lot of that coming back again, um is supported, provided to the individuals so that they have a soft ah landing into the country and able to integrate well um into their new ah new home. So this is what we've built essentially, a if you if you will, an Amazon ah for healthcare workers, for nurses, doctors, for, you know, physio, rehab, radiology, dental, blah, blah, blah. So that's sort of what we've built.
00:51:18
Speaker
um and And the reason I built it was actually deeply personal. It has nothing to do with money. It has nothing to do with, you know, just in terms of financial upside is, of course, is this enough and more happening? I just,
00:51:29
Speaker
you see a problem like and that's the problem with sort of I think some founders are founders that tend to be obsessed I got obsessed with the problem and thankfully it's it's paid off um you know we sort two years into the journey we are growing 40 percent month on month ah the businesses is is you know doing over 200 to almost 250 crores in revenues and and sort of accelerating fast and just it's because I think the problem is big enough to be solved um I couldn't I couldn't look away
00:51:59
Speaker
this A lot of questions I want to ask. 250 crore is your ARR? Yeah, annualized. This is money in the bank. So you're saying in about 40 months you've reached a 250 CRR?
00:52:15
Speaker
Yeah, so we're at 24 months now. 24 months. Okay. Okay. Okay. Amazing. um What's your headcount like?
00:52:25
Speaker
So we're 80 people within the within the team, largely India-based. So I would say 50% India, 30% Germany, and the remaining 20% between the UK and the UAE, Dubai.
00:52:42
Speaker
And how many people are you moving each month or like what's the annualized number of people getting immigrated? So we've we've we've had thousands, we've not sort of talked about the numbers yet publicly but um for for a few reasons. I think it's sort of, um you know, the government's also trying to maintain control over that information politically as well.
00:53:05
Speaker
ah So we don't talk about the exact numbers. It's largely into Germany. So I would say, Our business is 60% Germany, again, 30% in the UK, 25% UK and 15% roughly into the GCC region. We've done a bit of pilots in Japan the US now, which we're sort of expanding into soon.
00:53:25
Speaker
um But yeah it's it's it's it's in it's in several thousands today. ah Less than 10,000, but above 1,000. In a month, it would easily be like a three-digit number per month.
00:53:38
Speaker
yeah so yeah Yeah, exactly. So in some months, it's been two digits last year and some months it's sort of closer to three digits now. Yeah. Okay. So how do you do so much with such a small workforce? Because we're talking of 100 people moving each month and you need to...
00:53:58
Speaker
give them certain amount of hand-holding. And you need to also possibly, there could be training requirement. If you're sending people to Germany, I doubt you'll find German speakers in India. So that language training would be needed. Maybe etiquettes, whatever. like like It's very different in the West in terms of etiquettes and soft skills. So I'm sure that soft skills, etiquettes, polishing is needed.
00:54:19
Speaker
And then with the soft landing, everything that you're
AI in Recruitment and Talent Assessment
00:54:22
Speaker
doing over there. you So how you doing it with such a small workforce? Yeah. So I think this is the difference between sort of companies built pre-2022 and post-2022.
00:54:33
Speaker
um I think for us, we're an AI native company, right? um So what I mean by that is literally every single team member either has to learn to do their tasks with AI or...
00:54:47
Speaker
basically not be here, right? ah So there's a very clear mandate to the team, whether it's marketing, whether it's supply, whether it's training, whether it's you know operations, every single step of the value chain essentially has to be run with a right? And for us, that was a very clear intention from the beginning because I did not want to build 10,000 people company again that was doing a billion in revenues, right i mean, to me, it makes no sense um because the the dollar ARR per FD metric which is where the world is going, is abysmately poor. of them right So our goal is to get to a million dollars per FTE.
00:55:21
Speaker
um Let's see if you can get that. Pure tech companies are coming out of the valley, which are doing sort of two people companies generating 100 million. So we'll we get there. I'm pretty sure we will also ah do our level best. I think, but it's not about getting to that number. I think it's more about the intent and how you build.
00:55:36
Speaker
And the reason is I think the marginal cost of dev is heading to zero. right So the way you build companies now do not need to be, okay, now i have got this idea and six months later, tech team will have some capacity to build it. It'll take another three months.
00:55:50
Speaker
We're building it on the fly. right Our product teams are essentially, and and and this is where it's very interesting, the ratio of product to tech in the past has always been one is to eight, one is to 12, one is to 10.
00:56:03
Speaker
That's been the ratios. The ratio we are moving towards is one is to 0.5. I want one product manager for every 0.5 engineer. I'm not taking engineers out of a job, right? The best ones, the elite ones will always be working with us. But I mean, if even if we don't get to 0.5, 0.9, we're at 1.1, the point is it's really the different way of building companies and it's starting now.
00:56:27
Speaker
The way that compounded companies are getting built as opposed to single product companies. right As I mentioned, marginal def margin de cost going down to zero.
00:56:39
Speaker
Even if I need to build a product that is doing um phone calling to the customer or talents, I will never do it with a human at this point. The humans will do it to learn and then they themselves, so that the team themselves start to build stuff.
00:56:54
Speaker
Right. And then obsoleting that particular topic. Right. We essentially moved it so far up the funnel. Right. um Similarly, on training, language training, you mentioned. So we don't. we We actually did run German training classes and we continue to run some German training classes. it Takes sort of nine months to get people to be two level, which is sort equivalent to ah speak fluently um in a profession like nursing.
00:57:18
Speaker
um But what we started to experiment with was was AI facilitators, right? So the constraint there for, let's say, a market like Germany is not going to be ah people willing to learn German, right? The challenge are there enough German teachers?
00:57:32
Speaker
Can you create enough German teachers fast enough? And that's not something within my control, right? that's ah that's ah that's ah That's a capacity foundational problem. So can we fix it in a different way? Okay. We said, okay, instead of a one is to 10 ratio between teachers to student, how do we get to one is to 100?
00:57:48
Speaker
How do we get to one is to 1000, right? So my co-founder comes from EdTech. He sort of built one of the more successful impact the spaces. It's a company called Avanti Fellows, sort of did distance learning, partnered with Harvard, sort of you know social entrepreneur of the year multiple times in India and global education.
00:58:04
Speaker
sort of built urban company and then as well with with the team. Abhraj and the team is part of the leadership there. ah Built, you know, as load share, chief operating, sort of ran the and the entire business. ah CDC company comes from BCG Healthcare, care sort of family is in, ah essentially in, you know, two two two very famous doctors in India as well.
00:58:21
Speaker
ah So for us, And for him, and the education side of this ah did make a lot of sense. as a very clear ah path on how we would we would build it out. But for us, the bottleneck was not the teaching. It was really about how do you scale, a sorry, bottleneck was the teaching. How do you scale that instruction from 1 is to 10 to sort of 1 is to 100?
00:58:39
Speaker
where we started to use AI. Like you got one, so instead of 420 hours of lessons to get to let's say German v one ah by humans, it just, you want to bring it down to almost a hundred or 80. And then the remaining sort of get filled um through AI facilitation and practice sessions. So that's what we built. We have apps where people can be traveling from the hospital to the home and just practicing. it's like a Duolingo for nurses.
00:59:00
Speaker
i mean, that got spun up within a month and a half. Like it's just, it's insane what is possible now, which you couldn't even, like the constraint is human imagination. The constraint is not dev. The constraint is imagination. So I think for us, that's been, and I was presenting to one of the governments um recently um because we're sort of in in certain projects, we're competing with a very large players like Palantir now ah for for for landing some of these projects as well.
00:59:24
Speaker
um And they looked at our product and and this is the chief AI officer of the country essentially saying, I, I see tens of AI companies every day, um but I've rarely seen something as impressive and as close to human interaction as as you guys have done.
00:59:40
Speaker
um and And my point in all that, it's not about you know the words, it's really about the depth with which we have gone into building the stack, right? It's really, you especially the interviewing piece and the competency assessment and the you know the actual ah sort of two-way video interview or underwrite, if you can call it similar to Karl's, the underwrite of the person.
00:59:59
Speaker
um We have built this with clinicians, with doctors with 15, 20 years of experience. ah We've built this with nurses that have sort of stayed in head nursing roles, built this with healthcare recruiters. We've built this with people working at RanchTad and at Echo and all these wonderful companies with the AI technologists, the teams, up you know the the the product guys and the engineers working alongside them, right? And this has taken us you know over a year to build that particular model.
01:00:23
Speaker
So there's certain models we've gone super deep in because we know this is the version seven of it, that that AI needs to be able to respond under oath in front of a judge.
01:00:36
Speaker
That's how robust it needs to be. Because essentially what you're doing there is essentially finding people, this particular country wants us to create this talent quality filter at a national level, not at an operator level.
01:00:47
Speaker
Because they've got people coming from 70 countries to come and work there. And they're saying, if that's happening, this is a national security problem because I'm unleashing foreign... healthcare professionals as wonderful as they might be, but I don't know what competence they actually have for the requirement of the role in this country on my citizens, right? So for us, this was an eye-opening moment. and Like, holy moly, it's not just us solving for ourselves, but actually if we build this correctly, there's a much, much larger play, which is the original mission.
01:01:16
Speaker
Can you create a transparent new age way of talent finding opportunity and for countries and healthcare systems to find the best best qualified profession? So This is coming to life now much faster than we had anticipated because, and and I don't know if answers the question, in a very convoluted way, um but we've built it AI native.
01:01:38
Speaker
What are some best practices of AI? Making air products that work, like like ah one could just go to ChatGPT and give a prompt and say that create an interview bot and it would create an interview bot, but it would be pretty apparent on the other side that I'm talking to a bot.
01:01:57
Speaker
So, you know, what what is it that is over and above just a prompt on ChatGPT to create a good quality AI product? Of course, no, I love that. um So I think I'll tell tell you what we are doing. And I don't know if it's best practice. It is a practice.
01:02:12
Speaker
It's our practice. So I think two, three things. um One is, there is, you need to understand, we need to understand that the first 80% very easy.
01:02:25
Speaker
Whether it is dev, whether it is debugging, whether it is um design, whether it is legal opinion, whether it is interviewing. yeahi there's like ah There's a thousand AI recruiters out there. I mean, honestly, you go out there, you search on online, you'll find a thousand AI recruiters. right The point is not, or interviewers, the point is not to find somebody that can mimic a human conversation or can ask smart questions.
01:02:51
Speaker
The point is really in the tuning. and Is the specialization. I'll give you an example. So the reason why this particular government is so excited about working with us and building this out for them is like we've gone into the detail of having 80 different specializations of nursing.
01:03:06
Speaker
Right. And being able to create an agent associated with each of them.
01:03:12
Speaker
right? We are able to assess 28 different attributes for each individual specialization across hard skills, soft skills, communication, empathy, leadership, you know, ability to learn, blah, blah, blah.
01:03:24
Speaker
And if you think a three-dimensional matrix, then it's also the number of years of experience, right? So, i have three-year experience, pediatric nurse, um you know, which which on on empathy, right? What is the level of competence and how do you assess it and how do you It's really about verticalizing. So the best practice, the how we are approaching it, look, if it's a product that is non-core, which means it doesn't directly impact the quality of talent being produced or assessed, and it's an operational sort of product, then
01:03:56
Speaker
you would pick up Bolt or Cursor and team would go in and essentially just spin up and ah ah you know a design and an app, test it out with the internal teams, because largely that's where the client, or the customers, all external clients, test it out, learn from them, and then only go into building.
01:04:09
Speaker
So we would never go into building, and and the building in that case is essentially the last 20%, ensuring the interoperability, ensuring the infrastructure architecture works, and ensuring the data access works, and ensuring it doesn't fall apart in coroner cases, right? So that's really what the work happens.
01:04:24
Speaker
But we don't do any of that, and until we've created some kind of validation for the thing that we spun up in an hour, right? um On one of these one of these products. So we we tend to very heavily use some of these products for non-core, I would say, operational efficiency gains. I'm assuming this MVP, even a product manager could do.
01:04:43
Speaker
You may not even need a dev to build an MVP. We don't need a dev for that. Essentially, that's why you have more product managers because you want to tell your product manager, if there's a problem to solve create an mvp and uh if that mvp works then your your dev will convert it into an actual feature in the product exactly exactly so i think the approach there is really looking at i would say very high caliber product folks that are coming from founder backgrounds um because you're essentially not you're not
01:05:17
Speaker
So the product role essentially is a founder CEO role, right? That's literally my job. Like i I cannot delegate it to anybody either, right? But having said that, you need people that think and and behave like founders in those roles because now the product role is gonna become the most critical role in any company.
01:05:34
Speaker
Any tech company, the product role is gonna be the most critical role. Largely because they will be mini CEOs building out their stacks by themselves, figuring out with consumers by themselves and launching by themselves with sort of, you know, obviously elite engineers to help out on on the actual um implementation scaling architecture side, right?
01:05:56
Speaker
um And the final 5%, if you will. you will but The product role is becoming so central. And I don't think it's ever happened in the history of tech this much that and it it used to be a thing. It was never a thing. that Back in the day, you know was there was the business and there was the tech. And then the product managers came along and became a consultants and all sort of wonderful, smart people.
01:06:15
Speaker
But they did not actually have many a times the real world building experience. um So for us and for me, it's been a very clear choice. that these have to be founders that are very curious, that have built products before, um that know how to iterate fast, that have full ownership over the stack. So it's not like five layers of product managers now. It's literally the one person. We only have one person per business unit or per stack, and they do everything.
01:06:42
Speaker
And there's no confusion as to them sort of now jumping in here, jumping there. They have four layers of people sort of doing small. No. You don't need that anymore. You need to have very high quality. I think very similar to engineering engineering as well, the direction of travel is going to be more and more ah sort of engineers that have, that can do the final 20% versus the first 80. And yeah, I don't know if that helps, but that's sort of where where how we. That's best practices on product development.
01:07:08
Speaker
What about, ah so that's more on the development side, but you said verticalization is key. Yeah. It is possible to verticalize using AI. So for example, you can give a prompt and say that create manuals for pediatric nurse and then feed that to create the agent. Is that how you do it? Like, like like you know, does does the verticalization also happen with AI giving training data to the agent which is getting created or like?
01:07:38
Speaker
Yeah, no i like i love i love I love the idea and I'll probably use it in some way. Thank you for that. I'm not paying you for it. I'll probably pay you for it over over over dinner. um Look, there's three or four areas of how verticalization can happen faster, right?
01:07:51
Speaker
um So one thing, as I mentioned, is um having interview agents that can, so one, creating synthetic data and then feeding into this one with a human in the loop. Without the human in the loop, you're you're you'rere creating a completely Irrational outcome potentially, there's a very dangerous outcome possible. Again, operating in healthcare, operating in categories that are dealing with patients directly, we carry a very huge responsibility to make sure these agents and and AI products are built correctly.
01:08:21
Speaker
um with a level of depth that, as I mentioned, they can be under oath in front of a judge and they will give the same answer, right? That's an extremely hard thing to do. As we all know, if you've tried asking ChatJP the same question on five different days, right?
01:08:36
Speaker
Things do change, right? um So what I mean by verticalization is the if you're creating synthetic data, let's say there are 20 attributes, there's a male doctor with oncology with 18 years of experience with some Arabic language and English um and has degree from, i don't know, UK, MRCP, blah, blah, blah, blah. There's 20 attributes.
01:08:55
Speaker
That synthetic data of that individual can be created by an agent saying these are the answers this person would give. That answers have to be fed into the next agent that's going to assess it with a human in the loop to fine tune it.
01:09:08
Speaker
right So these are very important steps. You cannot miss these steps. The moment you miss this step, you've essentially created an infinite recursive loop. And I'm oversimplifying here in terms of explanation, which you don't know where it'll end up.
01:09:20
Speaker
The hallucinations will basically compound. and That's a problem. So that's one of the things that we've been very intentional about. That's why saying we have a team of clinicians, a clinical AI team essentially does only this. Their job is to look at the data, the output coming in, tune it and feed it back.
01:09:34
Speaker
and And which is why we're being approached by ah sort of some of the larger foundation labs now, um the AI labs to do the verticalization for their foundation healthcare models as well. So as without naming names, that's the one because we have access to a database that people don't really have access to, right?
01:09:51
Speaker
If you need an internal medicine specialist with like, you know, 12 years of experience can also speak Arabic. They don't exist. We have to essentially use our database to recruit them. So we have actually now moving beyond the general sort of data labeling like company next scale AI to now verticalized and deep. And that's where I think our database is also coming in super handy.
01:10:10
Speaker
um Some stuff we're doing is actually very, very interesting in that space of helping foundational models get built in the US and and other markets. ah You said there were a bunch of ways in which verticalization happens. One is synthetic data with human in loop. What are the other ways?
01:10:26
Speaker
So the other ways that we are looking at verticalization um is ah is is is user, right? um Is the interviews that are coming out without sort of having personally identified information because of GDPR. We're essentially able to take a lot of that real data coming in, the interviews that are really happening, um using that to continue to improve the models.
01:10:44
Speaker
um compare that with notes in terms of how would how did the actual ah client or the HR person review that particular person ah adjust accordingly and then keep improving take feedback from the head you know you did not read this correctly I was so shocked recently where we we Actually, i actually hired for a role for myself, which was like a founder's office hire.
01:11:07
Speaker
um We got over 400 applications, 105 of them extremely good. On paper, I would hire each each one of them extremely strong. And we had a version one. like the The clinical version is now version eight or version nine. we have a version one product now um on on this founder associate that the team unleashed on.
01:11:25
Speaker
And because the rest of it is the same, right? It's basically the same. So this would be synthetic data with human in loop version one. Exactly, exactly. Synthetic loop with human in the loop, a synthetic data version with human in the loop. Exactly.
01:11:37
Speaker
And we unleashed upon, you know, this set of people, like 105 people went through the process. um No, almost no HR interaction, WhatsApp document data gathering, and then the interview process. These are very smart people that have built big tech companies or have been part of ah very successful outcomes, MBAs from some of the best schools, love blah, blah, blah.
01:11:54
Speaker
And it was so fascinating that I saw somebody that I actually knew the person personally, right? I knew them. So I know the person has an incredible experience in tech and consulting and sort of work in various countries, in Southeast Asia, Middle East, Europe, India, like, and and big branding, right? So, um and that person got a 6.2 out of 10.
01:12:13
Speaker
And I'm like, this doesn't make sense. Something must be wrong, right? So either he did not complete the interview or and the questions were asked poorly, blah, blah, blah. And then I went through the actual interview myself, right? So because of the recording, so i went through the 20 minute interview.
01:12:29
Speaker
I realized that the context was created was answered, the outcome was explained, and the business impact was explained. They did not explain the actions in any of the answers, the five questions that they've given.
01:12:42
Speaker
And in my head, I'm thinking, oh, wow, that's um that's that's a pretty superficial answer, right? I mean, that doesn't, there's not, for a person that can um experience and that quality of work should have gone in depth. I kid you not, that was exactly the words written in the AI interviewer response, right? So what the recruiter, the the AI recruiter had had responded with, for somebody with that level of experience and that sort of variety, they should have been able to go in depth and the answers are very superficial. Literally the same words that came to my mind.
01:13:12
Speaker
So like, and this is version one.
01:13:16
Speaker
So for for for me now, we are finding that for clients, the recruiters, the recruiters we've built are actually going beyond what HR can ask. We're able to ask questions at a clinical level that you know a doctor would potentially encounter in their life. So we we're having to dumb down a little bit of our interviewing to not sort of eliminate some some some very high quality people also or high quality people also. So yeah, that's some of the other approach is to use live user data.
01:13:45
Speaker
So... yeah You know, the what what light this throws on future of work is
Impact of AI on Work and Adaptation
01:13:51
Speaker
two things. One is that there is a high chance that ah what you are doing today, if it can be transcribed, then you can be replaced.
01:14:01
Speaker
You know, any action which can be transcribed means you're replaceable. So be it interviews or whatever, will be it education. If it can be transcribed, then it can be fed to a model and the model can be trained on it and you can be replaced by a model.
01:14:15
Speaker
And the second insight this gives me is that a lot of work then is going to be more about quality. Like ah if you are replaced by a model, what is the quality level of the model? And being able to understand how to improve that quality is really what AI skills are all about, right? Like ah what you want to hire for is people who can look at what a model is doing and say, this is shit, we need to improve it and should also know how to improve it.
01:14:47
Speaker
Exactly. but think i think it's um it's a very obvious thing. I think the way we are looking at it, and at least I look at it, um is that AI not replacing humans, but AI is replacing humans that don't use AI.
01:15:08
Speaker
If that makes sense. yeah And I think the reason why this is so important to remember, and I keep reminding the teams, we keep reminding our nurses, we keep reminding our clients as well. The same thing is it's not coming, it's here.
01:15:24
Speaker
right It's like the internet came about um and people still do not know how to sort of use and browser um and are stuck. like this is This is bigger than the internet.
01:15:36
Speaker
This is bigger than the internet and faster. The way it's on on a daily, weekly basis, the amount of progress we are finding in models is just completely off the charts, right? So, and if you still don't operate in an AI native era, right?
01:15:50
Speaker
By the don't have any sort of assistant scheduling, anything that happens. and and and and And obviously my my my my life is very chaotic. I'm sort of all over the place. I have four different tools, AI tools that I use for ai for email management, for scheduling, for literally um a lot of my prioritization, my traveling, um different tools, right? Currency management, I do a lot of that. so I just think you you you almost have to get into that zone that now i have to start using the internet. Now I have to start using the internet.
01:16:21
Speaker
And for us, it becomes so comfortable right now with whatever we've been doing that we don't think about it. But the reality is it's here already. So AI is not going to replace the human intentionally. It will replace the human that doesn't use AI. right So let's that's for me would be the one big sort of, and you know people far smarter than me have said something very similar.
Tools for Efficiency: Granola and Fixer
01:16:41
Speaker
What the tools that you're using? I'm curious.
01:16:43
Speaker
So I'm trying out a few very different products. I love Granola. I think it's one of the most- Yeah, I've heard a lot of people talk about Granola. It's one of the most remarkable- It's like a constant transcriber. It's transcribing everything. Exactly. It's doing it right now. So while we're talking, it's also, yeah it's like an always-on thing, right?
01:17:00
Speaker
Right, right. love it. it's it's It's on my phone. It does the same thing. It's on my... It does the same thing on my... On my um um this video as well. Video calls, it automatically detects the video call and tells me. um I use... um i've I've tried Fixer, F-Y-X-E-R. It's a UK-based startup. Obviously, it's a of a bit more biased there um for email management. So, it does sort of a lot of the... So, i think after a week, it started to write emails like me.
01:17:26
Speaker
It almost... Like, i would literally... start typing and it'll draft it automatically and it'll keep it in there and say, you know, let me if you want me to send it. And I'm like, yes, please.
01:17:38
Speaker
It is pretty intense. it It goes to the point and this this sort of partly spooked me a little bit, but partly impressed, or very impressed me was that there was a certain conversation with a board member around know' not but one of our investors around some number of shares, and they were asking the question.
01:17:54
Speaker
It answered the question in number of shares because it identified it in a particular email PDF attachment. And I was like, holy moly, is this going too far? So I don't know. So that's another one I um i got very excited about.
01:18:08
Speaker
ah and so you've told me two ways.
Verticalization and Regulatory Compliance in Healthcare AI
01:18:11
Speaker
One is how product managers use it to create MVP. Second is ah in terms of data, synthetic data, human in loop, and then showing comparing with actual ah recorded performances, et cetera. Is there anything left in that in terms of verticalization best practices?
01:18:26
Speaker
Yes, yes. um I think um the other two, three areas, so there's one additional point, right? um And there's sort of branches branchs in the two areas. um One is given we operate in the healthcare space and there's a huge amount of responsibility on regulation.
01:18:44
Speaker
um So what we end up doing is we end up feeding a lot of the actual regulatory documentation ah for that particular agent. um Let's say for pediatric, there's sort of, you know, various boards that we need to be compliant to. If there's a specific country we're addressing, then that country's ah data ah from the from the regulator sort of gets fed into it.
01:19:03
Speaker
um And I know it sounds obvious, but it's not because you would assume that that data is already a part of a AI recruiter, let's say, or AI interviewer or AI assessor.
01:19:16
Speaker
um When it is about pediatric nursing, it's not. right So you actually have to manually tune it and add that context um um very intentionally. um So that's sort of the 3A, if you will. The 3B, I think we are very aware that...
01:19:34
Speaker
um we are not the experts long term, right? And and you know eventually it will be a partnership ah with with the healthcare systems in this case, right? um And so a lot of the work actually gets done ah with clients wanting a specific kind of prompts. I'll give you an example. So one of things that came in from a client was um if we hire 3000 nurses this month um using your platform and we are running a lot of parallel interviews, what if one nurse picked up the questions and shared with the other nurse and cheating happens?
01:20:12
Speaker
And I'm like, They can't because the questions will change, but I get your point. um So then there is a work that we're doing on proctoring, right? It's cheating assessments, right?
01:20:22
Speaker
Is understanding how the eyeballs are moving, first of all, understanding page refreshes, um understanding if um the the the person is is basically reading from a script, whether there's the same person that was part of the process, has actually attended the interview, of somebody else has attended the interview, um and obviously having a curated list of of of you know questions being generated for a bank of questions essentially versus singular. So I think there's a lot of that feedback process in verticalization that is super, super critical because you might build it in ah in ah in an isolation in a lab, right?
01:21:00
Speaker
um But if its if it's not really going deep enough, through levels such as that, customizations such as that, I think um it's not going to work. And the AI is not going to be strong enough, right? It's not going to work and in in various applications, so various different systems. So that's something we've learned. And I think the last thing is, fourth one is kind of obvious.
01:21:19
Speaker
ah but it's also sort of understanding the landscape of how the the the systems integrations are going to happen. And, you know, in some cases, um and leaving that too late, obviously an enterprise is going to be a massive issue. That's not doing verticalization, that's more implementation, but I'm sharing it um out of a context of of making what you've built now successful.
01:21:38
Speaker
Okay,
Tern's Vision for Revolutionizing Human Mobility and Healthcare
01:21:39
Speaker
interesting. ah I want to understand the business of Tern. You said you are competing with Palantir from... My understanding, you are a staffing business, an AI-first staffing business. So how are you competing with Palantir?
01:21:56
Speaker
Okay, great question again. um I need to be mindful what I say. um But just ah just on this, right? So I think it's obvious, or maybe not obvious, but I'll explain.
01:22:08
Speaker
ah Look, we we started with a a thesis on... um international mobility, right? That's what you're solving for. um You have a a massive information arbitrage um within which a lot of these agents and brokers and, you know, experts are playing and taking advantage of ah both parties, to be honest. um it's It's like the Nike analogy of of ah sweatshops in China.
01:22:34
Speaker
um And, you know, them saying, we don't know and we don't care. No, you do know and you do care and you are liable if kids are being used in those sweatshops. So it's like it's like the accountability of the supply chain does rest on on on on the on the eventual employer of the healthcare care system. and so So, you know, collapsing that supply chain and and and creating a two-sided platform in a marketplace ah was an obvious step for us to take because we needed to build our business, right? Which was, as you call it, recruitment staffing business.
01:23:03
Speaker
Very quickly, we realized that there is a um two things. Number one, there is a cap in the number of people a country is going to absorb on an annual basis.
01:23:15
Speaker
right So let's say healthcare in Germany um and nurses tends to get the biggest. There's around 40,000 to 50,000 nurses that are going every year. Right? And from ah from a market sizing point of view, and of course that's growing, right? So that'll keep growing to 50, 70, 100K, et cetera, et cetera, sure.
01:23:32
Speaker
But like from a market sizing point of view, if you think about how, where the market will correct over the next several years, you might make an average ticket size of let's say $5,000 as a platform or as a, in this case, a broker, AI enabled broker.
01:23:44
Speaker
um That meant that you were talking about a total market size in German healthcare of $250 million. dollars right Now to build a business, and um we're not we're not we're not interested to build a company that does 100 million or let's say 50 million revenues, right?
01:24:02
Speaker
Or 10% market share already very good, right? 50, 25% market share exception. um What we're trying to build is a generational company, right? We're trying to build the future of how human mobility happens on the planet, right? That's what we're building.
01:24:14
Speaker
If that's the vision, And to be able to create a staffing player in, let's say, and and and and to do that, you want to but you want to have at least a TAM size of 10 billion, right? That's how you' would approach it.
01:24:26
Speaker
We started to look at that and said, okay, to get to TAM of 10 billion, you're essentially having to have 40 country category combinations. right roughly 25, whatever, 250 times 40, which means you have to build regulation compliant human mobility operating systems right and and people behind it for 40 of these so eight countries times five five categories, which makes sense. I'm and sure you can we don't do it.
01:24:55
Speaker
That sounds exciting, but I think if you take it one level upstream and say, why don't we now do it at a system level? Why don't we, instead of us doing it one by one by one for everybody, why don't we do it at a system level and present almost a government level solution or a healthcare system level solution, national level solution?
01:25:14
Speaker
um That's where this all started began. And we started getting demand from, ah ah you know, the governments and the and the and the and NHS systems we are working with, the UA government working with, ah sort of the German healthcare care system we're working with.
01:25:26
Speaker
um So why don't you build this at a national level for us? I think that's where I mean by competing with Palantir. Obviously, it's a... company that has in incredible value created, $450 billion dollars sort of, you know, does billions of revenue from from government contracts. But the reason I'm mentioning that is we're actually starting to um get involved and and and and and compete and win versus his versus his players like that as well.
01:25:51
Speaker
So you are essentially white labeling what you have built and ah offering it to an NHS for cross-border recruitment or for all recruitment?
01:26:03
Speaker
All. So essentially it started with cross border because that's the toughest one, right? The context is very different. People are coming from different countries, language levels might be different. But then they pick that up and say, look, why don't I just assess not just recruitment, but why don't I use the same product to do a workforce assessment of my entire million people workforce that a McKinsey will take five years to complete, it'll cost them 150, 200 million.
01:26:29
Speaker
I can do that in a month and a half, two months, instead of five years at one 10th the price, right? So it's, I think the way of SaaS is also shifting rapidly towards almost a McKinsey plus Infosys model, right? So given we had built the product, and I'm not saying this is our primary business yet, but it's starting to accelerate quite rapidly, right? And this is this is potentially one of the one of the very big outcomes that we will see but originating from the original sort of engine which is the international mobility piece.
01:26:58
Speaker
Do you see this becoming a bigger piece of the revenue pie in a couple of years?
01:27:06
Speaker
I think so. I mentioned compounded companies before. What
Compounded Companies: Innovation and Expansion Strategies
01:27:10
Speaker
does that mean? I wanted to ask you that. You said compounded companies that previously. i forgot to ask you. So compounded companies and there's compounding in a traditional sense of 1% more every day.
01:27:23
Speaker
um But the way I think about compounded companies is that you have a core central thesis or a core central engine. And around that engine, you find concentric sort of circles of value that you're able to create.
01:27:36
Speaker
um And because the marginal cost of dev, as I said, is almost going down to zero, ah previously it would you would only focus on that core and that's what you're building. But right now the defensibility of that core is so poor, given the cost of dev is so cheap.
01:27:51
Speaker
or is becoming cheap, that unless you build companies which have those concentric sort of spheres, which are very tightly built around that core and have a lot of synergies, you will essentially be out of business in a couple of years because that's how fast innovation is moving. So what I meant by compounded companies is having these concentric spheres of value.
01:28:13
Speaker
ah One of them in this case being sort of a B2G or a B2E enterprise platform solution. um Another one, which I didn't talk about in the call um that we're very excited about is, is is look, we've got these these healthcare systems that were designed post-war in the nineteen forty s They've not changed at all um in the last 70, 80 years.
01:28:34
Speaker
um What if you could build a parallel health system by leveraging now a workforce that you have unique access to globally, which is the world's most talented um you know upwardly mobile and sort of aspirational workforce, which is Indian indian healthcare workers, nurses, doctors,
01:28:50
Speaker
What if you're able to build a parallel health system as well? So that's something that we are we are beginning to experiment with as well, which is essentially asset-light way of a hospital-at-home model um and really sort of creating a parallel system there. So Don't want to confuse you with sort of too much information, but what I mean by compounded companies is exactly that. And the third thing I did mention earlier is essentially AI labs, some of the top companies in the world, requiring a much more, a deeper database on specialists um in AT specializations in healthcare care um that we have unique access to, and then using that to of you know more and more tune their models. So beyond data labeling, it's the human in the loop and and sort of much more specialized ah feedback to their models.
01:29:36
Speaker
those are That's what I mean by concentric sort of spheres. I mean, Nick at Revolut does this beautifully. I mean, he he calls it big bets. As we're building a company with a core thesis and a core value prop, and if you're able to build big bets around it, you can actually create um outsized returns like what you're seeing with Revolut being valued at what, 65 billion now.
01:29:55
Speaker
in dollar terms. And and that's sort of one one thing that I sort of always have an eye on is is how much value can we actually unlock. And that can be now possible. As I said at the beginning of our catch up, it's limited by human imagination now. It's not limited by you know the tech priorities anymore. Much to the dismay and annoyance of my off my product and tech teams. I think it's more limited by us.
01:30:19
Speaker
Aren't you doing too much for a 24 month old company? Yeah. So I hear you and I agree with you. um Look, I think, as I said, the rules, the playbook is gone.
01:30:32
Speaker
So I often, you know, i had ah had a chat in London School of Economics the other day ah with with a class that was graduating. And and the question asked was, um what would be one advice you give us as we leave?
01:30:43
Speaker
And my comment to them was forget whatever you've learned because the the when you entered your course and when you exited the course, the world has changed. It is not the same world anymore you're entering. um I don't know, Akshay, to be honest.
01:30:55
Speaker
um I do know that having... Experiments that can have billion-dollar revenue outcomes um is something that I'll keep doing. um i may not Not everybody might will get it. um Some people will, and that's fine. are the people that will come on our journey together as investors and board members.
01:31:15
Speaker
Some won't, and we're fine with that. um But I'm not here to build a company that's doing 100 million revenues in 10 years, and you know it's just in the middle of it. It's not a bad outcome, by the way. it's it's I won't call it a zombie company, but we're here to truly change how work gets done in the future and and and solving for um the human mobility. um and and And this is the way that we we feel we can. right And if I can unlock one or two themes and you know if we need to carve out companies later on, if you can do other things with it, of course, we'll do all of those things.
01:31:46
Speaker
We call it turn group for a reason, because I also think the product, the the category is so vast. um that there's this insane amount of of of ah upside and value to be built.
01:31:57
Speaker
But yeah, I mean, many of these bets might die out and they might not pan out and that's okay. I'm fine with that. and But the ones that will, will create those, you know, multi-billion dollar revenue companies. So that's really the outcome we're envisioning.
01:32:11
Speaker
your Your ambition is inspiring. um
Ethical Recruitment and Training Initiatives in Germany
01:32:15
Speaker
Okay. ah The ah revenue model is just a pay per hire or do you have other opportunities also? Like say in Imo, you had multiple ah places where you were able to make ah money. So is that similar here or is it just, you said eventually it'll be like a $5,000 per hire.
01:32:31
Speaker
Probably today it's higher, I'm assuming. Yeah, it's close to 10. Yeah, yeah, yeah. So look, I see um as it's it's so interesting, right? I use this phrase, which it can sound crass, but if you understand it, probably you get it as well.
01:32:47
Speaker
We're birthing a worker in a new country.
01:32:52
Speaker
That worker did not exist. That baby did not exist. You essentially, you don't have a baby, you've birthed a worker. The moment you birth a worker, A lot of things happen. That worker needs credit.
01:33:04
Speaker
The worker needs a house to live in. The worker needs um services. The worker needs a career planning, education. ah The worker needs, ah you know you know, food. They need everything, right, essentially.
01:33:17
Speaker
Now, I'm not saying we tap into everything or everything is is is materially important to to to ah spend effort on. um But some of the bigger, they need to remit money back home.
01:33:30
Speaker
There is a lot of this that actually is very high value, right? So housing is very high value, for example. um The career upscaling, education, continuous education, language training, et cetera, is quite high value. um You know, there's there's value around ah major life decisions like cars and so on.
01:33:45
Speaker
There's decisions of remitting money back home as well. ah Services there, you know, I don't know. is there ah Is there a stable coin linked card? I don't know. Like just, there is a lot of stuff that is starting to open up.
01:33:57
Speaker
We're not acting on all of that. um We're starting to experiment here and there. ah But as you correctly pointed out, and now suddenly they've gone from making 15,000 rupees to 3 lakh rupees a month.
01:34:09
Speaker
ah So they also have a um a different sort of requirement and expectation on financial side. um So we we are seeing those streams opening up. We are finding revenue opportunities there for sure.
01:34:23
Speaker
um But I think um to to us, that's it's It's more about value creation, right? So what I want to build there is a perfect counselor, right? So somebody that can, you know, three years before and they even think about moving, right? They're in final year of nursing college.
01:34:36
Speaker
They should be able to pick up the phone, speak to us on an AI conversation, ask all kinds of questions. and that And that person that... calling it AI a person, um that counselor sort of holds the hand um until the person is settled in and in the future as well. So that's if you build that, then you open up a lot of opportunity of of but revenue maximization and expansion into services. So yeah, to your point, yes, it's very similar. I would say it's even more than the EMO case. Yeah.
01:35:03
Speaker
Currently, like, are you monetizing the training that you give? Like, say you're teaching people German language, is that monetized or are there other aspects which are monetized? So um ah you we don't monetize the training. So and because we are one of the very few licensed players in Germany, as an example, um we also commit to ethical recruitment.
01:35:27
Speaker
And today the definition of ethical recruitment in Germany includes zero cost recruitment plus zero cost training and the training essentially being subsidized by the employers um on the other side.
01:35:39
Speaker
So it's a working capital situation, but we don't charge. ah But there are things within that that, you know, as as soon as you start to create counselor sort of model, ah there can be a premium plan, there can be a basic plan, there's more credit to get, blah, blah, blah.
01:35:52
Speaker
um You know, on the on the ai interview prep side, the mock interviews, a lot of this, we're not monetizing, we're starting to do some experiments. um But honestly, the as I said, the the the the burden of cost, I don't want to keep on the talents. I want to shift it more and more towards the employers. And I'm actually very happy that that's the way it works in Germany. And and it's not true for all markets, but at least some markets have understood that um dynamic between the two sides.
01:36:19
Speaker
You know, you have markets where and there is ethical recruitment as a mandate and and they don't expect talent to pay. But then there are markets where, like say, Middle East, traditionally, it's always been the agent charging the talent and not charging the employer, especially for like blue collar, gray collar kind of roles. so So do you feel like you'd be able to compete in those markets?
01:36:39
Speaker
Yeah, no, again, very, very well-researched question. um So I think there's a key distinction between, um so there's two vectors to consider. One vector is English speaking markets versus language lock markets.
01:36:53
Speaker
um That's one vector to consider. ah The other vector you touched upon a little bit as well is ah blue collar versus highly skilled. When I say skilled, right? um and And that's another vector to consider. And the reason I mention is if you pick up the English speaking markets, let's say the UK, US, Canada, the UAE, Australia,
01:37:13
Speaker
um
01:37:15
Speaker
There is a the the kind of problems being solved, needing to be solved there is slightly different. um It's an efficiency problem. It's a transparency problem as opposed to an availability problem, right?
01:37:29
Speaker
So they the supply does exist. The supply is super inefficiently sort of distributed um to find that exact person you need out of those, i don't know, 250 applications um is what you're solving for, right? There's a lot of desperation. Nurses will give the exam. Doctors will give the exam.
01:37:46
Speaker
um But you don't know how to find that one person. So as I mentioned, that's why it's ah more of a system level problem as opposed to a staffing or a physical mobility documentation problem right it's more about the efficiency of can the supply meet the demand right um you know in a super uh quick um and and and transparent fashion um whereas if you come to a market like germany or japan for that matter or you know belgium netherlands know you can pick up any market in in
01:38:17
Speaker
greater Europe and or even South Korea. um It's more of a, and the ecosystem itself, there's not enough supply of German speaking. There isn't, there is a, there is no country that speaks German out of dark region, right?
01:38:31
Speaker
So Germany, Austria, Switzerland. So that's also a manufacturing problem as well. so you almost have to create the availability of talent that can speak the language. So there we actually had to get involved in building up our own language curriculums, actually getting, you know,
01:38:44
Speaker
from other countries like North Africa, getting people down to train and nurses, sorry, train the teachers um in India with our partners, with ourselves, and really build up that ecosystem. And then we sort of do a lot more of the ah finishing and sort of school approach ah using technology.
01:39:01
Speaker
um And similarly similarly for Japan, and there's there's a there's a product market fit, there's a timing point in this, way You can't go too early because people don't really want overseas workers. Now everybody is saying we cannot survive without, um nurses in this case or professionals like health skilled professionals like that.
01:39:19
Speaker
um So that's the one vector, right? So for us, we are not looking at a market like UAE or the UK or even the US from a supply building point of view.
01:39:31
Speaker
We are looking at it as systems point of view, which is what mentioned where we compete with Palantir now. In the NHS, we're competing with them. In the GCC region, genry competing with them. And we see other sort of English-speaking markets in the US and so on. We find very similar sort of challenge requirements of pinpoints.
01:39:45
Speaker
um When it comes to the, um I would say the blue collar versus the bit bit more gray collar, skilled labor or soft white collar, we were very intentional that we actually did not go after the blue collar for a couple of reasons. Number one, I think For better or worse, um there is a level of informal, the the informal sector works to get these people placed in construction, um in logistics, in sort of you know manual labor jobs, maintenance facilities jobs.
01:40:24
Speaker
It works to an extent, um but again, the challenge there is more an efficiency problem. So I would say a system solution is much better placed than trying to become another agent.
01:40:35
Speaker
um To be frank, the margins also aren't there, right? If you want to do this ethically, the margins don't exist. So what we would rather build is a two-sided marketplace where people can find the employer directly and they can use the system, go through the AI interview, go through the uploading of documents and are able to connect with the employers.
01:40:52
Speaker
and that employer essentially being facilitated by the government. So there's government level filter. That is what I would rather build ah versus again, trying to build an agency where margins don't exist and shouldn't exist. I mean, these people are coming from really difficult backgrounds, right?
01:41:06
Speaker
They're not really in place to pay that much for i shouldn't be paying that much, ah which they do pay to to several operators in the category. um Whereas when it comes to skilled labor, specifically around certain markets I mentioned already, the other the other um point to remember is there is a quality filter as well, right? Your education, your technical background.
01:41:28
Speaker
So there's a consistency in evaluation that we can achieve um towards the employers, towards the clients and a fairness that we can achieve towards the talents, right? um Whereas in blue collar, in some cases like real estate, like construction, it can be, there's not as much of a difference between where the supply is coming from, what the person's qualification is,
01:41:50
Speaker
For good or bad, right? um I'm not not saying. um And and i just we just didn't feel we would add as much value there, um given it is the informal sector is working. um I do think the system solution is possible, and that's where we would rather spend our efforts on.
01:42:04
Speaker
How did you crack these government accounts? Like, how did you get an entry into NHS or the German medical
Strategic Partnerships and Government Insights for Healthcare Solutions
01:42:10
Speaker
system? Was it investor introductions? Yeah.
01:42:15
Speaker
This is fascinatingly weird to explain. No, my board members actually aren't. So we do have some folks that have been sort of part of, as investors have been part of these systems. So the former chairman of the NHS England is an investor in company. He's sort been a mentor to me when I set up the company in the beginning.
01:42:35
Speaker
you know The deputy health secretary, government of UK, the CEO of AXA Healthcare, care um the chief executive of the Conservative Party in the UK. one of the former health ministers of Germany as well, right?
01:42:47
Speaker
So there's people that have really worked hard on landing right up front and creating sort of interest about this solution. And they are coming to us and saying, if you are successful, you can actually save the country if you if you do this well.
01:43:03
Speaker
So for them, this is a very personal mission also. Apart from that, I think it's honestly been a lot of... following the threads um as well. like So those intros turn into other intros. and and so on. And if you remember right up front, I explained sort of coming from a government family, both me and my my wife, ah my father sort of being head of human resources or equivalent at the Indian railways, a lot of the learnings of how these systems work, how the decisions get taken, how a workforce should be operated. And you know he's also a PhD in organizational design, literally this topic.
01:43:35
Speaker
He's held you know governments of Sri Lanka. He was part of the Air India restructuring before the Tata acquisition. so It's sort of in the blood in a weird way, um I think, um because that would be dinner table conversations.
01:43:46
Speaker
um So I think that's helped, you know, as well ah to to proceed further and land a lot of these MOUs with governments like Japan and the UAE and also our state governments actually in India, some really forward thinking governments, to be honest, that we're working with.
01:44:01
Speaker
All right. Awesome. Thank you so much for your time, Avila.