SuperLiving: An AI Health Coach for India | The Investment Memo: India | Ep. 1

0:00 So I can guarantee you today that if people have used super living they will not switch to charge chip. >> The company has scaled to 24,000 paid subscribers in just 4 months. Each paid user uses the app about 20 times a month that's almost every day and spends 50 minutes per day which has improved from 20 minutes per day in November. It's crazy. >> So we have this leadership principle in our company which is learn test launch repeat. I tell people that you cannot reach a winning variant in one shot.

0:34 >> Not only is it multi-agent but every single agent actually has its own identity and personality. Yes, definitely. So at the core we believe that health should be as simple as brushing your teeth. >> A tier three which is which has never looked at anything outside dal, roti, chaw is now suddenly talking about I added protein my kids are not eating waggi. >> Yeah. So there is a u you know ritual that we are following that we have to make user calls every week be 10 users be 20 users and it doesn't depend which team you are in even the tech team is making these user callings how are you thinking about long-term defensibility for yourself does it keep you up at night >> I think now we have found the answer so defensibility is actually closing the loop Welcome to the investment memo. a series where light speeded opens up about how we thought about our investments when we wrote the check. today we have super living with us. and super living is basically an AI coach for lifestyle and wellness built for India at an affordable price point.

2:02 founders Manav and Gurjot have joined us today. Thank you so much. as well as my colleague Priyel who's also joined us today to talk about the memo. We're going to spill the beans. You know, when we think about an investment, I would I would argue like most early stage firms, the first thing we think think about is the team. >> Yeah. >> Right. It's it's what we spend the most time on. >> we run Lighteed is notorious for running some of the longest reference check calls on founders. And I'm going to read out some of the things that we that we saw.

2:38 Manav and Gjot move fast and they iterate a lot. they've pivoted twice from pregnancy management app to a platform for lifestyle content and then now to an AI coach from over the last 7 to 8 months. light speeded has stayed closed through all three evolutions and what has not changed was that we saw this team each time doing very fast iterations having a real feel for their users and the willingness to drop quickly what was not working that stood out for us.

3:15 I'd love to sort of dig into what was going on in your heads as you were thinking through these pivots. So I think when we started with pregnancy a content company the goal was to figure out if lifestyle can become entertaining right so all of the last few years that we spent at pocket FM show we understood that users were spending a lot of time digitally online on consuming a lot of content but we always were looking for is there enough content in India market which is more outcome oriented can people also learn while they get entertained while watching that kind of content And to be honest the answer was no. Most of the content was entertaining. You can pass your time but you don't get a result from that. and that's where the first version or the first iteration of super living was started which was a content company. We called it as life tainment category. You think of it as lifestyle which is health wellness everything around it >> combined with entertainment. Over a period we met a lot of users and we realized that okay >> this works but it only works to a certain limit. why? Because users want personalized guidance. While they are watching content, they are loving it. But somewhere the character goes through an arc which doesn't relate with the user >> and then user suddenly starts not watching the content from that point onwards.

4:34 >> So we realized okay now there is a need for personalization on the top of it and that's where we launched our AI companions. >> The goal was to make AI companions feel more human. they should look like a normal human in India. They should not be a celebrity and they should be approachable at all times. We never we never pivoted. I think it was more like you mentioned evolution right. So from pregnancy when the pregnant ladies asked us to help them with their weight loss after pregnancy >> it was a natural shift to introduce weight loss with pregnancy. Then the similar males let's say the new fathers who were going through stress because they had so many changes in their life mentioned about stress management in the app and we launched stress as the second category and over a period as these were younger females and males we realized okay their parents also needed support.

5:26 So we as we kept going deeper into family these categories kept getting lost over a period of time. Interesting. So yeah, thank you for sharing that. And most investments that we make are leaps of faith, right? And there's a very little there's very little signal especially at early stages in companies, right? I want to give you a glimpse of actually how little we have to go by right the importance of health and wellness has accelerated postco and AI enables accessibility and affordability of health and lifestyle coaches. Health is a large category with 15 to 20% of all charge GPD queries and 10% of all Google searches being health related.

6:09 That's the data point. And what will happen at the cusp of AI and wellness? Nobody knows. and so when we invest at that stage, it is literally just that you know the the very small little early signs >> that we lean into, right? And that was the journey here. Having said that, we have our own views, right? And we still believe that this this has the potential to become a very large category that there is latent demand that people don't have access to affordable coaching. how are you articulating the problem for yourselves?

6:48 Right. >> Like you mentioned there is a latent demand. People go to Google, people go to chant GPD, they search about u I am facing STD2 I am am I getting heart attack they search every single thing on social media on even on ress right and what we have seen is that it actually creates more confusion than solving the problem right so how we look at this problem is that people are everywhere they're searching everything everywhere but they get five different opinions when they go on five different platforms now where do we come in is when you are not healthy and you are not chronically sick right so think of that user who says I have a joint pain or I am pre-diabetic now or my kids are not developing or growing well because of their nutrition or my parents have diabetes but I like eating sweets what should I do all these things that are a part of your daily decision making >> which ultimately maybe in a decade or two decades will lead to some chronic illness that's where we come into picture >> and and now if let's say Chad GPT or Gemini or Claude everyone is trying to solve this problem the problem actually doesn't get solved with guidance right so guidance is the starting point >> it's like you are running a race >> and you have just heard the whistle or the gunshot right >> but you have just started it you haven't finished it the finish happens >> towards the finish line and that's where the results matter a lot.

8:21 >> Do you think you are building the market? You're creating the market that or do you feel this market for a wellness coach, for a guide for how to how should I live my life? >> Y >> that market exists but in a different form. >> I think we are majorly creating a market because there is no digital market on coaching >> and all the tracking which is available today >> and there is a need. So these are two different things. There is a need but there's no market. So there's an awareness of the problem because of >> you know the internet and the media >> but there is no clarity on the solution.

8:57 >> Correct. Right. >> Got it. >> Yeah that's actually incredible. And one of the themes that we've actually been excited about is that specifically when it comes to tier 2 plus cities right there has always been this latent demand. Now with AI it is finally possible to solve it at a very low cost with an agent that's always on that's providing a personalized journey to you guys. Right now I'm actually curious to understand did you intentionally go after T2 plus cities >> and if the answer is yes how did you manufacture virality and just the distribution?

9:30 I think it came very very naturally to us. all the success that we had in our previous organizations to be it Misho or pocket. I think one thing was very clear that we were able to understand the user really well because we belonged to the tier two three cities right so I come from a city called Kar which is near Chandigar and most of the people have never heard about it but because I have lived with these people I completely understand you know when they say weight loss we see we see that chronic illnesses are preponing by 10 years in India and a mass from that is from the tier 2 three cities so there is definitely a need to change their behaviors. So that was the mission also we sort of you know started with.

10:13 >> And when you talk about virality I think one thing that you know instantly comes to my mind is relatability. >> because we understand them. I'll give you a little example which you know will make it a lot more simpler to understand. For example you go to a nutritionist u you know you have been habitual of eating paras. Let me talk about you know my Punjabi household here. >> Who isn't? So we are very habitual of eating you know cholelay buras for breakfast. So when to such a person you say cho and you know you eat pha or you eat let's say basin ka chila it feels so off that this is not my palate this is not how I can you know make changes in my life but here if you sit with them and you tell them that you know let's say for the first two days you just don't put butter on your para >> you know then maybe you start taking plain para with paneer because now you want to add protein to it now maybe you know d the now the para can change into a chapati let's say so the ghee has further gone down maybe you add boiled eggs so when you start making these subtle changes being highly relatable to them you know being very regional first I think that is how you can you know crack that market be more viral and they'll also happily accept you as a company >> yeah very very interesting I want to I want to segue into the next section of our memo the other thing that we've covered is execution that's generally a third section in our memo, right? Like team, market, and then execution. there's interesting early traction and strong customer love.

11:49 >> The company has scaled to 24,000 paid subscribers in just 4 months. Each paid user uses the app about 20 times a month. That's almost every day. and spends 50 minutes per day, >> which has improved from 20 minutes per day in November. It's crazy the fact that retention depth and time spent are all improving even as the funnel widens significantly is the strongest signal.

12:19 One very interesting choice that you guys made in order to accomplish this outcome was that you decide to go multi- aent and content first. So for example, if I have to solve for skin or sleep or fitness, there are different agents that are actually doing that. simultaneously there was a belief of actually having short form videos a lot of content first to keep the users engaged. What is the why behind that? Yeah, I think we went multicategory and multi- aent for two reasons. First let me cover the multicategory vala part. U whenever a user comes on the platform they definitely come with one particular mindset. Again taking an example of let's say a housewife here. Now this lady is gaining weight because she's stressed. She's stressed because of all the you know strains that she has in her relationship be it with her mother-in-law could be with her husband maybe she's just feeling alone because her children are now you know out of cities in their colleges so she's really feeling very lonely. So when she's stressed the only solution that comes to her mind is that I'll cook you know my favorite food and I'll eat a lot. So when they start talking about a problem naturally they start talking about multiple things which are now essentially not directly you know related to weight loss >> and that's where the companion you know multi-companion part comes in >> traditionally also I feel like we Indians always have a specialist we have a doctor that will look at your gut your acidity your constipation and then you have a hair care specialist or a skin specialist a dermatologist who will look at to your skin.

13:55 >> Very interesting. I think the second interesting choice was that not only is it multi- aent but every single agent actually has its own identity and personality. >> An example for the same is G almost all customer calls that we ended up doing people were raving about it as if it's an actual human. >> there was a lot of deep trust and I'll say respect. What is what is the thought process behind actually every single coach having its own personality and actually using that for branding and customer acquisition as well?

14:26 >> Correct. I think at this point coach Tara is more famous than anybody of us and I think also the company so even with all these user events that we were having we were meeting with user sometimes you know people don't remember super living but they do remember coach >> Tara >> so the product quality has to be the best for and then of course we can you know layer it up we can put that makeup of how she behaves when you are talking to her. So her talking like a mother to you making you feel very relaxed even if you are stressed about your weight loss you know she'll say that you know beta so using those particular languages or the mannerism is what differentiates her and how then an AI doesn't essentially be an AI it becomes much more than that but I want to switch gears now into the into another section of our memo and I want you to brace yourselves this is a section on risk >> and open questions in your case.

15:29 before I read the snippet out, I also want to say we did debate this a lot and what carried us through was the two of you >> our faith in you and what we believe you can accomplish as well as the execution we had seen. Right. So but I will be open about where we saw the risk. defensibility against we specifically mentioned open AI and Gemini health here.

16:03 >> both of these will launch health in India. we've seen that it is 15 to 20% of JPD conversations already and India is a large market for most companies. How are you thinking about long-term defensibility for yourself? Does it keep you up at night? how do you feel you can start manufacturing defensibility early on? >> By the way, this is Harsha's favorite question. >> Yeah.

16:36 >> So, yes, we thought about it a lot. we also debated about it what really defensibility is. and I really want to clear out a lot of noise around defensibility in the market also. Yes. >> So people think a great UIUX is defensible. It's not. Right. You can build an app or launch an app in a week or even on a weekend now. Yeah. >> Right. So anything can be copied now if whatever is userf facing. Then data I I feel that I personally felt that data will be something that we will always boast about that super living is defensible because of the data it owns.

17:12 >> Yeah. now the answer is no >> right because if you use the best of frontier models today >> they are 80% already there they understand a tier three user >> they understand their constraints and these models also understand India's culture it's not like you can ask fable GPT6 any other model today on what will be the fasting diet of navaratri female right it will answer you it's not that GPD doesn't know right >> so data is not defensible today.

17:44 >> So then the question always arises what is defensibility? And I think now we have found the answer. So defensibility is actually closing the loop. >> So when I say closing the loop, >> how do you use that data to enable your systems and agents to learn better >> and then how do you firstly of course collect high quality accurate data. Then how do you synthesize it, store it? What is your memory model in the back end of the companion and how does every single new user who comes on board gets slightly better in terms of probability of success compared to the last user >> that closing the loop is the most important defensibility. Chat in itself is the biggest dump of data but most of it is verbal data. So people are telling us that I did this, I achieved this but there is no evidence to it. So we have spent a lot of time on building evidence based data repositories first >> where you can correlate a user doing a thing >> or let's say user was guided to do A the user actually ended up doing B which was let's say 60% of A >> user achieved C which should have been let's say 100% of accord in an ideal situation but which was only 40%. M >> and one interesting insight that I have from meeting our users is that let's say there is a user who is using chat GPT Gemini versus super living same user will give a very different data to both of these platforms why because when they are using a free product they tend to get an answer fast they they want an answer fast with our app the same user is conversing a lot so they are not here just asking Q&A it's more like I want to know Costara first or Costara should know me and I will give you a lot of data because I have paid for you.

19:43 >> So the moment a user becomes a paying user >> the accuracy of data increases almost four 5x. So I can guarantee you today that if people have used super living they will not switch to charge CPT >> right especially tier three users or Gemini even if it's free of cost >> because proactiveness combined with high quality >> how old you are >> yeah and I'm challenging GP interesting >> well we believe in it >> so fully aligned with your thesis on this I think what is also very critical about execution in any consumer AI category is actually going to be accomplishing outcomes because that's ultimately what users care about at the end of the day, right?

20:25 >> So talk a little bit more about why does outcome matter? How do you do that in the first place? >> I think outcome is equal to retention. very simple one word answer. If users are here for a problem statement and they don't get an result, they will not stick right. So you can build games, you can build videos, you can build you can get celebrities, build a big brand, you can do IPL advertising, you can do anything but if your product is suboptimal and it doesn't lead to a result for that specific problem statement, user will not stick to the app. When people also ask me why can this app be the same for metro users or let's say US users, the answer is no because >> tier 2 behaves and works in a very different way. But when it came to our users, they said that we don't have a device. U almost 50% of our users don't have weight machine itself in their homes, right? So they go out once a month and check their weight, send us a image and they come back and for the next one month we will not know are they losing or gaining and neither do they right. So then we found out a lot of hacky ways >> where the probability of accuracy was still 60 70% there >> but that gives some direction to us and the user that I am improving without getting access to those devices right because people really don't want to spend money they just want that you tell us at home >> how can we actually lose weight or get a better skin so for example we we built a lot of trackers now inside the app which are again AI based trackers >> which can analyze your skin which can analyze your body fat percentage. and then hacky ways around that you take out simple user example we asked users to take out an inch tape. Most of them have that measuring tape in their homes.

22:13 >> Yeah. >> And you just measure your belly, your thighs and your arms and you upload that measurement with us and we will figure out are you fat, obese, what's your BMI you don't need weighing machine for that right? So these things took a lot of iterations >> but because we were fast and we kept on launching and killing most of the products we arrived at what works the best for the user. So we have this leadership principle in our company which is learn test launch repeat right >> I tell people that you cannot reach a winning variant in one shot.

22:46 >> You might have to do 10 things to get one success >> but for that nine failures to happen you will have to launch those nine products. Right? So and those nine products should be launched in two weeks, right? So you fail in two weeks so that you can succeed in week three. >> On the outcome part, if you can give an example just to make it real for us, all of these coaches, what outcomes were you running for? Any story that comes to your mind there? So there's one story I can think of which is a story of a individual starting the app and somehow it went to the whole family and everyone was using it and we actually did a video call with the whole family to understand how it evolved. So the story started with around a 45 yearear-old housewife I think she was she was living in Agra so she came to the app just to manage her weight because she had thyroid as an issue. So she was not able to lose weight very fast. for the first 14 days she saw around 1 kg weight loss which is very sustainable but what she mentioned to us was in the first 2 months she lost close to 5 to 6 kg which she felt was the best she has ever done and she was only using coach Tara for most of her use cases. Now over a period of time we started seeing her time spent increasing on other coaches except coach Tara right so we understood that she was actually hooked to three more coaches one was a kids nutrition which we recently launched and because she got results from coach Tara she felt that now my kids can also get some form of benefit from another coach and because everything was a part of a subscription and we have seen that people really do their best app.

24:28 So basically this is how tier 3 works, right? So they're very value conscious. So if you have paid 200 bucks, you will try your best to take out 2,000 rupees out of this app by getting everything that you can get on this. When we connected with the family on a video call, they told us that we are also using your app for our parents basically the children's grandparents because they are diabetic, they have joint pain, they have a lot of other problems and at the end we are a joint family. We all cook the same food. The best we can do is add variety or delete some things. And the interesting anecdote that came out was that the coach Tara had guided them to cook the same dal for everyone.

25:12 >> But the dal gets extracted for the grandparents before tarka >> after tarka for kids. And then kids also get additional paneer added with the dal. But the grandparents are only having dal. the whole family is eating rice because they love rice. But now the rice is shifted to brown rice than white rice. And they're all happy doing it. And everyone is adding curd and salad to their meal so that they eat less of carbs and more of fiber. So this change in a tier three family which had never talked about macros >> and we were surprised because I think 5 years before now I was also never talking about macros honestly right >> a tier three which is which has never looked at anything outside dal roti chaval >> is now suddenly talking about I added protein my kids are not eating maggi >> that was a big win for us as a company but I think very interesting that five users with just two subscriptions get the best out of this app.

26:08 >> Fascinating. Very interesting. >> Very, very exciting. I actually have one question, Harsha. So, there were a lot of teams building in this category. What got you focused and excited about us? >> I'm going to let take this. I think first of all both of you are from pocket of film and when we actually ended up speaking to people over there who worked with you very closely we actually realized that you had actually cracked a category that was so difficult at that point in time which was the fantasy genre which actually ended up becoming one of the most important categories for India for pocket FM.

26:38 >> Yeah. And I would say every single conversation that we had with you guys right three things that actually stood out which we heard from pocket FM folks very consistently was as even Harsha mentioned obsessive focus on the customers listening to them I would say very very keenly >> the second one was just the shipping velocity right because the amount of evolution that actually happened across all of the phases that's actually not natural that comes in when you're shipping with a lot of intensity and I think third is >> building the business with the right focus areas. So at the end of the day, what really matters in businesses like health and lifestyle at the end of the day is actually retention. So having a very keen focus on that which kind of is what we heard what we saw in every single conversations and additionally when we spoke to the customers I think it was we like to talk about it internally it was actually one of the most positive set of customer calls that we actually had where people spoke for like 30 minutes with us just talking a little bit more about >> what has Taraji actually impacted in their life the outcomes that they've actually achieved be it in terms of weight loss be it in terms of improving sleep so that was unusual And that got us very excited.

27:49 >> So this is what we heard from people at Pocket FM and our own experience. I'm curious to understand is that how you guys see yourself as well and what are some of the core values with which you've been building the company. >> Yes, definitely. So at the core we believe that health should be as simple as brushing your teeth. Now these are decisions that you know you you know subconsciously take. So you get up you have certain habits. Now these habits can be very simple. This could be something like you know you have you get up and you go out for sun so you spend 5 minutes in the daylight so that your circardian rhythm can you know get the gist of it then you come back and carry on with your other tasks. So we want to make health which comes very naturally.

28:30 It doesn't feel forced or it doesn't feel like somebody has you know put a lot of restrictions for you in how you live your life. and lastly of course you know as pre you and Harsha were also mentioning we have always been very close to our user. >> So that is a core principle and that is how you know our team essentially works. before every project you know every data that we analyze we try to anyway we draw insights but we try to correlate that to how the user is feeling about them as well. So there is a u you know ritual that we are following that we have to make user calls every week be 10 users be 20 users and it doesn't depend which team you are in even the tech team is making these user callings just to understand how they are interpreting interpreting the product if the product is making sense to them and if you know it is being as useful as we are you know making it to be for them >> that's absolutely wonderful >> very interesting >> yeah actually I'm also very keen to understand there was a time when you guys also had a lot of offers in the market right everybody was chasing you >> >> what made you select us choose us talk a little bit more about that >> got it >> I'll take this it's a very interesting story we had been connecting since I think almost 8 months right so we have you have seen us how we iterated from preced to now series A round u when you guys gave us a term sheet we were 99% % sure at that point in time that we don't even need to think about it right why because I I could see the kind of conviction coming from calling you guys meeting you guys multiple times across that round and we saw that you guys also felt that this can be built the way we feel right because it's very important for us to find a partner who feels the same way about the user about the problem as the founders right because otherwise we would be just convincing a partner why this is right to be done and then I think now the posts signing phase where a very interesting thing happened very recently is that I met hersa for a one-on-one session I think it was a one and a half to two hour session first time we got a chance to spend so much time without any distraction and there were a couple of takeaways which are of course secret for the world but we both know that there were a couple of takeaways which I think completely changed two things our mindset and company's trajectory both in a positive direction right so I think I was pushed a lot in that conversation to think very differently from what I usually do u and the moment I went back from that conversation I spent seven days just thinking about what happened right and when I came out of that 7-day period I met a few investors and I told everyone that I think my ROI with light speeded is already sorted like I have taken out enough from that one conversation than I would have >> from let's say one year with some other partner right and that was the impact and now if you look at just last couple of months of our MIS it actually shows a sustainable growth happening at the right places now if I look back I think I'm 200% sure that I did the right thing >> thank you thank you so much for sharing that anecdote mad I just want to say This is a little in Hindi and I think that understanding that getting an investor on board is a partnership and half the job is in brainstorming and talking about problems openly. So I want to thank you for your openness.

32:16 >> Thank you. >> maybe we'll switch gears a little before we start crying. that brings me to the last section of our memo where we try to wear the founders's hat but obviously we're not good at it. to imagine what this will become. >> Yeah. >> What's the vision? >> Thank you. >> We will be taking a forward leap of faith on health and lifestyle dam. >> As I mentioned earlier, there's not enough data points.

32:50 and we believe this dam will open up with AI coaches who will bring in affordability and accessibility for India. We feel super living will eventually end up at high retention as it is multicategory which is already played out. 15 specialized agents and 20 plus categories have already launched and it includes higher stickiness lifestyle categories like personality development, anxiety management and confidence building. why we're confident that the team will get here. We believe that a habit that survives the weeks when motivation dips is a far stronger habit that will see stronger attention. and we believe a multi-category multi coach approach enables one to achieve that.

33:38 I'm sure we haven't imagined 1% of what you're imagining. So please tell us what you think. >> What is your vision for the company? >> Got it. So I think from multicategory to multi codes the next version of superliving will be multi-user within the same family. Right. So the biggest unlock that we you would see and I think it's hardly one or two months away from today is building a family first app where you can integrate a lot of your family members and your coaches understand each family member's needs separately without hallucinating >> and this can unlock a lot of use cases not just people who live together but also people who don't live together.

34:21 Think of it as my parents who don't live with me here. >> I can manage their health right from my app. This can go up to people living outside India whose parents are here or parents whose kids are in college. It can unlock a lot multiple other different problem statements for us. So that's something that is immediate and you will see it live very soon. >> post that we are now working a lot on building new memory models. so so far we have launched five different memory models. the latest one being close to 70% in our locomo benchmark testing which is generally a memory benchmark score >> and the next one which is going live in next month is already scoring 86 in our internal testing.

35:04 >> Wow. >> Which which reaches which makes it very very proactive but without and without being too intrusive in life and also makes it more human right. So every single memory model that we launch makes the AI human and it also makes the hallucination go very very low. So so we are expecting the hallucination will go to the levels of 0.01% in 9 months of app usage. Right? Eventually once these two things are out and all of this happens within this year >> you can look at our company in the next one to two years to become your personal health agent. which started with guidance, moved to engagement, then moved to remembering everything for you, then booking your services and now in future you can get access to the right doctors, the right hospital, you can check vacancy in hospital, you can maybe even book ambulances in future.

36:00 >> but the core focus of doing all of this will be retention first. >> If any product increases the retention, it sticks otherwise it gets killed and that's how we work right. So you have to prove retention to make your product go live. That's the one thing for all product managers who join Super Living. >> Yeah. >> Yeah. >> Very very exciting. God, I have goosebumps. with that we'll we'll end today's episode. Thank you so much for being so open and you know we don't generally we've historically not opened up our memo to anybody and really glad we started with you guys. It was a very nice >> frank and in-depth conversation. So thank you for your time.

36:46 >> Thank you so much for your time. >> Thank you so much. We also really had a wonderful time having this conversation. >> Yeah. Looking forward to another one next year.