Transcript
Speaker: Five months after we launched, we got an email from a 13-year-old girl from the US saying, i have depression, I tried to commit suicide and you're the only thing which is helping me hold on, so thank you. That's when the penny dropped. started believing, oh my God, this is something serious.
Speaker: WHO says more than a billion people are living with a mental health condition and only one in three with depression ever gets scared. Ramakant built Vyasa to close this gap, an Indian AI mental health app used by 7 million people globally. On this episode of the Founder Thesis podcast, Ramakant and Akshay Dutt discuss how to scale a services platform globally.
Speaker: We were adding 3,000 users a day, which in mental health context, imagine a public health system serving maybe 3,000 people a year. And here we were adding 3,000 people a day. I think one of the largest providers of mental health in India is an astrology bank.
Speaker: Oh, really? It's not in our culture to admit that you need help for mental health related issues. 40% of our usage was coming from the US.
Speaker: Ramakant, you are the founder and president of VISA. I would love to first of all understand this beast that is VISA. It has ah different elements. As we were just chatting, you told me that there is a clinical part of it which comes under healthcare and then there is a digital part of it which comes under wellness. ah Just first, can you define the business areas in which you operate?
Speaker: Correct. no no And very nice, Akshay, that you mentioned a beast called Viza because it actually is a penguin. So not quite a cute, chubby little penguin at least. That's the mastermind.
Speaker: That's the avatar. And ah incidentally, Wiza was also, it's a inside joke on how we came across the name. A lot of people ask me saying it's a very unusual name, by ah why Wiza?
Speaker: And it's is' loosely based on Eliza, which was the first chatbot which was which came out of MIT in the 1960s. So it was a dumb chat pod in the sense that it was a series of questions and very open-ended questions which were simulating what is called Rogerian therapist, so which is basically doing therapy by questions.
Speaker: And asking questions and then and allowing of that to make you feel. yeah and how And what makes you think that or why do you think this is so? So it then prompts the process of self-discovery. So um when we were launching, we thought, oh, well, maybe this could be like Eliza, but only Wiza.
Speaker: Wiser. So that's how that happened. And the penguin was an accident. So I think we were thinking about mascot which would be accessible. And we just happened to do launch with the penguin. And at least in the early days when we were in direct to consumer, people loved it.
Speaker: So and we then got very strong feedback saying, can we have a plushy toy which we can hold and don't change the form factor, the penguin it is. And we tried to make it a little more grown up and thinner and taller. And we got a major backlash.
Speaker: So it's very much as a penguin. But of course, now that we're going into health care, it's going down. So ah to answer your question, um so I think it's useful to think about this as um started as a direct to consumer app, but very rapidly over the last eight, nine years has grown to become what we call ai as mental health infrastructure. So it's like a platform.
Speaker: I know it's an overused term, but it really is a platform ah which has multiple components or modules, elements, whatever you might want to call it. um Which starts from if you think of a patient journey, right? So you have somebody who's presenting with some patient or a user or journey who's coming in with distress and wants to have a conversation. There is an initial part of what we might call e-triage or context gathering and navigation, in other words.
Speaker: And where you are really understanding the user and saying, okay, what are you looking for? And what is your objective and what kind of services to say if you're working with a client or an institution and then helping them navigate whatever is possible or whatever is available and reach a certain point of um care or give them a certain point of care.
Speaker: Then there's an element which is if suppose the ah the need is at a certain acuity level, which is below a threshold, then you say, oh, this is suitable for self-help. And then you can give them a self-help ah module or a set of modules which can be delivered on an app or web, whatever. And this is where Wysa incidentally began. It began as a D2C or Platter-to-Consumer app.
Speaker: So that's kind think of it as patient support, self-help. And then there is a third element, which is suppose now the need progresses or the the person is looking for something more. Then you have an opportunity to provide a hybrid model of care, which is introducing or integrating the digital piece ah with a human support.
Speaker: And now that human support, now the tech platform is something which can be actually used to deliver human support. So say if you work with an insurance company or we work with a public health system, they may have therapists of their own.
Speaker: What we can do is we can give them a software, ah what what is what we internally call co-pilot, which can allow the clinician to stay in touch with the patient or with the user and deliver services.
Speaker: which also include the WISA chat and the WISA AI services. Well, so for example, if I'm i'm taking therapy, for instance, the therapist is in in touch with me, they can also send me homework on ah on an app, which is basically the WISA chat.
Speaker: And they can give me little meditation modules or conversation modules. And that can be interspersed with... ah the actual delivery of a conversational support from the clinician as well.
Speaker: So you can you can deliver that hybrid model of care and in that case, in that situation, you can either offer just the software, the the digital platform, or you can offer the entire service including the clinician.
Speaker: And last year ah we made two acquisitions in the US, which enhanced that part of the business. So we actually added on um services which allow you to deliver that hybrid multimodal kind of care, human plus supported by digital.
Speaker: And the most important thing was that we were able to ah now reimburse with a payer in US, which is a in a sense, the gold standard in terms of what a digital health startup can do. And there are multiple ways in which you can monetize, but that's the one which is most stable.
Speaker: And it is also really, really large. By payer, you mean insurance company? By insurance company. That's right. Okay. Okay. Okay. So I just want to recap. So there is a therapist part of the business. ah There is a self-help part of the business, which is the WISA app. And then there is a co-pilot part of the business. This co-pilot is like a white label solution that you offer or? It it can be white label or it can be integrated. It's not therapist part of the business. I'd call it just human support.
Speaker: Because there are various types of human support. so And in fact, even in the US, what we offer as behavioral health support is not therapy. So it is a non-licensed, so it's and it's not a licensed therapist, it's a behavioral health care manager.
Speaker: And there is a certain model of care called collaborative care, yeah specifically in the US, um which is reimbursed by an insurance company, but it's not therapy.
Speaker: So you can actually deliver, a andi we can we'll talk more about that in detail, but you you can actually then expand the kind of services possible ah because you don't have to depend on a licensed therapist who, as you know, on ah globally are in short supply.
Speaker: ah For example, in the UK, if you wanted to go and see a therapist in the NHS, the National Health Service, the public health system in the UK, a wait lists are up to one year long. So you have to wait almost ah almost a year to see a therapist. so there are And even in high-resource geographies, very well-funded, advanced healthcare systems, and there is a huge demand supply gap.
Speaker: So models like collaborative care help plug that model. And of course, the work we are doing in AI ah starts to the very, very initial part of the spectrum or base of the pyramid and delivers an even higher level of support at a very early stage of need.
Speaker: So the idea, of course, is that ah we started Wiza out of a very personal journey ah with mental health. And both my co-founder, Joe, and I had a have background in social entrepreneurship. So what we wanted to do was take a crack at solving a global problem, and but do it in a sustainable way, in a way which actually can just sustain itself and hence scale.
Speaker: um And that's how Wiza was born. Okay. ah Help me understand the mental health the TAM here and how it is broken down. So like you told me that ah there is a difference between a licensed therapist solving mental health problem versus a ah like a manager ah a care manager solving a mental health problem. so So what's the overall market of mental health? Like just define the problem area and the TAM and how it breaks down, like different segments in it.
Speaker: It is, i mean, if you think about, say, if it ah there are various pools of or various markets within mental health. So, but if you really abstract it out to those 30,000 foot view, I think the easiest way of saying is how many people are in distress at any point in time.
Speaker: And ah of course, during COVID, I think it was everybody who was in distress. Right. WHO, the World Health Organization came out with statistics saying at any time, ah one in four people are in distress and need help.
Speaker: So that's one very ah broad way of defining GAM. Then you can also cut it by severity of need. and different specific conditions within mental health.
Speaker: So there are other statistics would say, you know, 20% of people people are suffering from anxiety and or depression. So that's another way, but that's one in five. um So there are various ways of cutting this problem or cutting the need and cutting the time.
Speaker: A very interesting way, of course, and for a startup, ah that is useful to a certain extent saying, yes, there are a lot of people who have ah who have need, but that's not enough to base a business on.
Speaker: Because there, then you need to look at time very, very differently. And you can say, oh, i have a certain business model or an operating model. For this kind of operating model, what is the irreversible market? Or what is the serviceable market?
Speaker: And there you asked, i think your question was specifically around what I said about collaborative care earlier. Now, a very interesting statistic is that 50% of all mental health need, and especially in the US, which is the geography we are serving with this particular model, ah shows up at the primary care level.
Speaker: So basically what that means is a person doesn't say, oh, I'm having anxiety. Let me go and speak to a therapist or one in two people may do that or try to do that. The other person goes to their normal primary doctor, primary care physician, a GP. And, you know, the person you'd go to for a headache or a foot sprain is also the person you'd go and saying, hey, you know, I'm i'm worried or I'm having anxiety.
Speaker: Now that poor GP or the PCP in healthcare parlance, the primary care physician ah has a problem now. Because a person is shown up and saying that I have some symptoms or some I'm in distress.
Speaker: Now, one, they're not trained to do anything about it. And two, even if they don't know where to send the person. and So there at the primary care level, ah given that one in two people or 50% of need shows up there, the collaborative care model was invented to actually intervene at the primary care level in such a way that you're not having to depend on therapy therapists who are in any case in short supply.
Speaker: And the way the model works is that ah you have a behavioral health care manager who is not a therapist, but who's trained and who operates under the umbrella and supervision of a psychiatrist.
Speaker: ah But The patient still belongs to the PCP, belongs to the doctor. And this behavioral health care manager is providing support through a phone call, say once every few days. And whatever support is being provided and whatever is happening with the patient is then going in as data into the patient record in the electronic health record, which the PCP can see.
Speaker: So one of the ah other big issues, if you really go really deep into the healthcare system, is that when when a patient is handed over from one specialist to another, your original owning physician, if you will, loses control or loses track and sight of what is happening to the patient.
Speaker: So the collaborative care model actually solves that problem because it is delivering a good high-level ah model of care and it is doing so under the care of a psychiatrist.
Speaker: who is also in very short supply, by the way. So here the behavioral health care manager is actually expanding the reach or the number of patients the psychiatrist so can support. And it is happening under the umbrella, care umbrella of the PCP, who's owning the patient relationship in the first place.
Speaker: So okay advanced model, but it's really useful. It really solves the problem at scale. And it is very interestingly, it is reimbursed by an insurance company.
Speaker: And as you know, you know in the US healthcare system, that $4 trillion dollars healthcare economy, a large part of that is fee-for-service and is is reimbursable. and that And that's one of the most stable ways in which you can monetize.
Speaker: ah Am I correct in this understanding that ah the... ah Based on the severity of the mental health the problem, the patient can either be recommended to work with a behavioral health manager or be sent to a therapist. Like more severity means sent to a therapist. Possibly there is an even higher level of severity where they go into a healthcare institute or a... like Or a helpline. If there is extreme distress, if there is SOS, then you get you get escalated out completely into a specialist therapist.
Speaker: And, ah and in fact, that's one of the weaknesses, which um after chat GPT and after Jenny, I came about, I think that is one of the weaknesses which happened is saying that one ah it's a point solution. There's no handoff. And the second is how do you manage things like acute distress and thisera safe escalation pathway?
Speaker: And the third is how do you identify the distress? and And you do that and so on. So these are all and these are all clinical safety guardrails you need to build into a solution to make sure that it is available and appropriate for deployment into in in a context like mental health.
Speaker: Okay. So this ah co-care ah model where there's a behavioral health manager and the person the relationship continues with the primary care physician, ah is this... ah ah is there some monetizable benefit for the primary care physician here ah or ah like there is no benefit other than the relationship there is and um ah so effectively what we are doing is we are a service provider to the PCP we are a doctor right so the doctor is the person owning the patient relationship they are the people billing they are the people raising the claim with the insurance company and they people getting paid
Speaker: ah Okay. Then we are we are charging a service fee for delivering the service. And here, um it's not just human support because you are integrating the why at a digital tech with the human support and that combined service is what you're delivering at the turnkey solution to the PCP or to the doctor.
Speaker: And this is a remote solution, right? There's not an in-person. Yeah, it's remote, which is also one of the reasons why it can scale. Right, right, right. Okay. So essentially then you are a sub-vendor to the PCP, but you are a sub-vendor who is approved by the insurance companies and therefore the PCP can offer this to the patients and then raise a claim.
Speaker: So, ah you know, we don't need to be approved as a service, but yeah your clinicians need to be having certain certifications. And for example, your psychiatrist needs to be accredited and licensed in each state which you're offering. So there are there's nuances there, but broadly, yes.
Speaker: That you need to be a service which is of a certain standard, which the PCP can adopt. But otherwise, generally speaking, it's the PCP who is raising the claim. So as long as the doctor is satisfied and happy with the service you're providing, you are one step behind in terms of the insurance company.
Speaker: So it's basically the doctor who's managing and owning the patient relationship. And ah so this is not like an existing market, right? There's like a market you created from scratch.
Speaker: No, it exists. It exists also. um um And sorry, I'll use an acronym here, but collaborative care is called COCM. So collaborative care management. So COCM codes ah were introduced by CMS, by the health authority in in the US, who sets reimbursement pathways, saying which services can be reimbursed by whom.
Speaker: And ah so I think they're about two years old. So it's not very old. So therapy, of course, is very established. And there the big movement happened during COVID when CMS allowed therapy to be reimbursed remotely.
Speaker: So remote therapy or online therapy was being was able to be reimbursed. And that created an entire... ah market out of scratch. This better help ah as any podcaster, anyone who listens to US podcast would have surely heard their ads. Got it. and So, and that that, I think just created an entire market landscape.
Speaker: And every year there is this slight, I think, flutter of a heartbeat where people say, oh, will CMS withdraw that? ah It was called an emergency provision saying that you will allow online therapy to be reimbursed.
Speaker: And everybody has a little bit of a palpitation whenever this is up for renewal. And it it does get renewed. I think now um we've also seen that usage has shifted and usage patterns have shifted to online therapy. And now, of course, with AI, it's totally different.
Speaker: all Right. Okay. Okay. Got it. Okay. And... ah How big is the time for this collaborative care ah opportunity? I don't have a specific number ah for you, but it is in the billions. I mean, there's no i doubt about that. Because if you say 50% of all mental health therapy and then you have a certain LTV around $1,500 per patient,
Speaker: So it is it is so large that whatever we are doing and whatever we can do will be a drop in the ocean. So it's ah it's a chance to build a very, very valuable business and a very large, high-scale business. How do you grow this business? Is it ah through signing up PCPs?
Speaker: And state by state, because for each state, you would first need to have the paperwork and the approvals in place. And then you go and sign up the PCP. Is that how you grow this? Correct. And and that's the, since we're talking specifically about the U.S. healthcare care part of a business, um that, and for any anybody anybody, any listener who's, say, an Indian digital health startup, and they're thinking, saying, hey, ah we got to go out and monetize in U.S. healthcare.
Speaker: um I think one thing is not easy. It is actually, there are lots, there is obviously huge amount of regulatory barriers and things you need to know. And there are things that which us as founders in digital health did not even know that we didn't know. um So, which is why our acquisitions really helped because we got access to teams and and structures which are completely set up for monetizing in US healthcare in this reimbursement pathway.
Speaker: One, I'll give you two examples here. right So one of them in terms of how it is special or how it is different from the way a digital health founder or otherwise typically you would think about Asia or India.
Speaker: um One of them is ah the fact that there is something known as a corporate practice of medicine guideline, which basically means that any services which are being provided to patients need to happen from a clinician owned entity.
Speaker: And typically now, if you are a startup and you may not be a clinician, I'm not a clinician. So how do you then get around things like those or constraints like those? It's basically on setting up what is called a PC MSO entity.
Speaker: So PC means a professional corporation owned by a clinician. And MSO is you a managed Managing Service Organization. And you then have an agreement between the MSO and the PC saying that even though clinicians are managed and deployed out of ah the PC entity,
Speaker: The MSO is the person who's controlling financial planning, ah patient recruitment, policies, procedures, and administering the PC's wishes or the PC owner, the clinician's wishes.
Speaker: So it's a very, very fine balance where you are retaining, in a sense, dotted line control. from a financial point of view, but clinical control is completely out of your hands. And you have to be very, very careful in your contracting and in your legal documentation to not even give the appearance that you have control over clinical ah service provision.
Speaker: Otherwise, you contravene and you you flout those laws and then you are in for trouble. So that's, I think, one very ah key element. And the other one ah is also this whole, and there's an anti-kickback statute, which basically means that if, for example, you're working with another pc entity or another clinical services provider who have access to patient flow, you cannot go there and saying, give me a patient, I will give you X dollars.
Speaker: Hmm. that's ah That's incitement and that is under the ethical guidelines in US healthcare, you cannot do that. So you need to be very, again, both from an ethical, what, you know, doing the right thing and doing things right.
Speaker: You need to be careful you're doing both. And so structuring and creating a legal governance structure or public structure for U.S. healthcare and reimbursement to U.S. healthcare is not easy and it's a specialist skill.
Speaker: ah How much of your revenue comes from the U.S.? I would say about 70%, 60%. Wow, amazing. so ah this ah U.S. revenue, one part of it is, of course, the
Speaker: Sorry, I keep forgetting that term. COC? COCM. Collaborative care. Okay, collaborative care. Yeah, right. So one part of the revenue is the collaborative care. How much of that is this? COCM contributes how much in that? COCM plus the the other acquisition we did, Akshay, also was expanding this. So COCM is basically remote behavioral health.
Speaker: And we actually took one step further and we acquired a company which is also delivering physical therapy. Physical therapy. Yeah. Physical therapy, you mean like recovering after a surgery? Surgery, and so exactly. And or chronic pain management. And the reason we did that was because ah in earlier on in Wiza's history, we had actually seen that there's a huge overlap between physical and mental health.
Speaker: and Any chronic condition has a biopsychosocial element attached to it. So, for instance, if I have cardiac disease or or I'm living with diabetes, it is almost a ah Psychological ah management is also required. Right. So being living with diabetes is like being on a perpetual diet.
Speaker: And that has and yeah if you don't manage the psychological aspect of it, then you will fall off the wagon and your clinical out outcomes will be bad. MSK is even higher. It's a huge amount of overlap and there's also physiological overlap and comorbidity between depression, anxiety as well as physical pain.
Speaker: So imagine someone who has fibromyalgia who's going through dialysis or who's just had surgery and is recovering from surgery. So their ability to adhere to a plan, to a physical recovery plan,
Speaker: As much of it is about overcoming mental barriers as not. And we did a three-arm clinical trial about, I think, in 22, 22, 23 on patients with chronic pain.
Speaker: And ah we were able to see that delivery of a WISA program ah was able to achieve outcomes in depression, anxiety, physical function and management of pain at par with the gold standard, which is actually speaking to a therapist.
Speaker: So and it and it led to both kinds of outcomes and that led us to believe and understand saying that, oh my God, this is a white space where there is and any kind of chronic condition. Think oncology, for instance, right? Think cancer care.
Speaker: ah Pain management is a very, very large part of cancer care along with chemotherapy, but very seldom do people do it well. So these are the kind of opportunities where we want to now take that COCM++ model and take it and and combine physical therapy support as well and take that combined solution ah into specific ah therapeutic areas like oncology.
Speaker: So this would be like on-site, not remote. This can also be virtual. So physical therapy can be virtual as well. So if you combine this whole concept, then roughly about 50% of our... Current revenue is coming from ah the US healthcare reimbursement services and the remaining 50% is a mix of ah what I would call population support.
Speaker: So which is whysa the original VISA chatbot as it started, but now being given to say a million people in ah who are policyholders of Aetna or the entire population of Singapore.
Speaker: So the the Ministry of Health in Singapore offers, WISA is powering one of their large early stage mental health intervention, which is being made available to the entire population. so So, and plus the work we're doing in impact in India. So all that, if I call non-US healthcare, that's about 50%.
Speaker: And US healthcare is about 50%. And the healthcare piece is growing. It's growing at an amazing pace right now. but That's the massive market, right? ah the usl you said I think you said 4 trillion. ah That's the US healthcare. a yeah US healthcare. care Yeah, right, right. Okay. Okay. So ah very interesting. So the ah COCM is 50% of your revenue, which is mix of- I'd say US healthcare is 50%. Yeah. COCM plus physical.
Speaker: Together. ah Physical therapy is also, I'm bucketing it on the COCM. ah No, it's actually not part of the COCM model because the COCM model and the reimbursement pathway is very specific. So we've added another reimbursement pathway, which is physical therapy. Together, I would just, I think a better way of calling it is it's US healthcare reimbursement.
Speaker: That's the point. Okay. Okay. Okay. So in the physical therapy model, is it like a different kind of ah gatekeeper? Like in COCM, clearly it's a PCP, the primary care physician. Yeah. Here also there is a physician.
Speaker: So it may be orthopedition, it may be a pain management specialist, but again, it's a doctor. So effectively, and I think you asked the question on saying, how do you grow the business? um So it's actually, it's pretty much a ground game where you have to go and speak to surgeries or primary care, doctor clinics,
Speaker: And convince them, show them the value to the patient, to the provider, make sure. i think one of the big issues again is, or be big fears of adopting any services. Will I have to do more work?
Speaker: You know, it's not even financial. It's mainly saying i have limited attention spans. I can't, ah So you have to fit within their pathway, their clinical care pathway and their process. You cannot introduce anything new. So things like yeah EHR integration, ah making sure that your electronic health department has departments. So integration with that, making sure that they don't have to juggle or toggle across multiple systems.
Speaker: these ah These sound very minor, but they can make or break a business. And in healthcare. So it's it's extremely, extremely important, especially for people who are building for things like US healthcare or for doctor adoption ah is to not underestimate the the power of removing doctor's pain.
Speaker: And the ah physical therapy solution is also built by the ah the the primary doctor or the but doctor. No, in fact, they this has gone one step further where we are ah able to build directly.
Speaker: Okay. ah With the pair. Okay. But the recommendation is still in both cases is coming from the doctor. In this, I guess it would be very similar to how pharma companies do their sales. Like they have these medical, I mean, in India, they're called MRs, medical representatives.
Speaker: I don't know what's the term in US who go visit doctors who educate them. Yeah, I know it's a great, it's a good analogy. But we're also very conscious saying that don't send a non-clinical person to talk to a doctor.
Speaker: Because then you're, ah it's very hard to, I think, create any kind of ah ah impact. It's much easier if you have a clinician who's talking to a doctor. For example, in our COCM business, the account management function, if you will. So, and typically a sweet spot, we have also learned along the way is that going after and speaking to say a large hospital system, say a Mount Sinai or a ah mass gen or people like that is it's a very high risk high written return proposition right because it's a very complex sale some 10-15 people will have a say in whether or not their patients in their hospital system will be sent to a startup to service um and you might get through eight of them it'll take you a year
Speaker: But then number nine will say, hmm, I'm not sure. And then the deal is dead. okay great so it's And that's also, I think, one very hard one learning. Saying that we very quickly after, think, multiple false starts with many, many, many people, ah we've realized that a sweet spot is actually a ah preferably a physician-owned provider group who has, say, imagine ah a surgery or a doctor's office who have 10 clinicians, 5, 10, 20, or multiple sites with maybe 50 clinicians.
Speaker: and ah But it's the decision-making is very quick and it's very small. the decision make ah The set of decision-makers is small, so it's much easier to make a sale.
Speaker: and And I think that is really, really important from a GTM perspective. So this is like a standalone, uh, healthcare center or a medical center or something like that.
Speaker: Uh, standalone or, uh, in the U S you also have a lot of, uh, uh, folks who are now increasingly being rolled up with within in private equity transactions. So you have chains of clinics who have now been acquired by a private equity firm. So then those also become interesting.
Speaker: um The other one is you go to a health system, but then you have a clinical set of clinics who are affiliated with the LK system, but they're not part of the hospital itself. So that's a good way of acquiring patients or a good way of acquiring clinicians as well.
Speaker: Okay. Got it. Got it. Okay. Okay. Understood. how How big is your team in the US? And like how many of them are into what kind of function? ah the our Total across the world, we are about 150.
Speaker: 150 is super lean for ah like a $10 million dollars heading to $20 million dollars business. I'm so happy you said that. I keep losing sleep over this thinking we need to be leaner, especially in the age of AI. You're too many people and maybe we should be, you know, one person should be doing many more things. So that's internal pressure at least.
Speaker: ah But, and out of that, I would say about one third 20 to 30% is in the US in terms of the number of people.
Speaker: So it's outsized, you know, 80% or 70% of our revenue, but sort a stuff And the last part the engineering and product teams are in India. Right, right.
Speaker: But yeah the ah behavioral managers, the BHMs, are they not on your payroll? This 150? Yeah. I would have.
Speaker: So including these people, it's 150. Okay. yeah That is very lean. And these would be based in US or are they also all over? No, they have to be US. Regulatory requirement for them to have some...
Speaker: ah like some license to operate or something like that. Absolutely. So this part I understood. Now the other part, which is the population services of your business. um This is essentially what started as a direct to consumer, like a consumer app for healthcare. ah Just take me through that journey. Like like you you said that there was a very personal reason why you got into this space and and you started doing this when there was no...
Speaker: uh, like this was a pre-chat GPT, pre-transformers era. Uh, there was no such thing as a foundational model. ah I'm curious to learn what made you want to get into this and how, uh, the, uh, the, the foundational models and all that's happened in there, how that has changed the way your product has gone through. Yeah.
Speaker: Yeah, no. ah So two or three questions rolled in one. um So I'll tease them apart and I'll keep it short. it's it's ah It's a very interesting story. I think most people in healthcare start out of personal need or there's either a physician who's seen this and saying, I don't want patients to suffer or there is somebody who's suffered and said, I don't want others to suffer.
Speaker: and And then you jump in because building in healthcare is not for the faint hearted. and they're I think easier or other ways of making money. But here there is always a mission attached. or Nine times out of ten, any founder I've spoken to always has a story.
Speaker: And the same with us. ah we um So although i' am both my founder Joe and I, um we don't have mental health backgrounds. but And so I'm a consultant, so strategy consulting, investment banking. So as far away from mental health as you can imagine.
Speaker: So I probably caused more mental health issues than solved them in my earlier life. So in terms of the work I did. ah But then and Jo, my founder, she's the product person and I'm kind of the the person wearing the business hat.
Speaker: And we are also married to each other, by the way. So we are a founder couple. Our lives took a very interesting turn in 2009 where ah we got the opportunity of moving and working in social entrepreneurship ah in the Middle East. So there was a foundation being set up. Moving from India. You were in India prior. You were in.
Speaker: So I was, am my last corporate job was at Goldman Sachs in London. And then we had an- But you grew up in India, like India's- You grew up in India. So then I went overseas and did an MBA, then worked in strategy consulting, then banking.
Speaker: And ah the startup, the foundation, social enterprise was being set up under the foundation or umbrella of ah ah the World Bank in the UN. n And with a large endowment from the Qatari royal family, the Queen Shekha Moza was on the board.
Speaker: Very luminous, very high-powered board. So Nobel Peace Prize winners, the ex-commissioner of, high commissioner of refugees, ah the UNHCR's ex-commissioner, and one phone call away from any kind of leader you can think of. And the mandate was to look at conflict areas in the Middle East and create livelihoods and economic opportunity for young people.
Speaker: as a way of intervening so that socioeconomic unrest can be minimized. And we spent five beautiful, beautiful years there, really transforming both, I think, professionally and personally. This was the time of the Arab Spring.
Speaker: If you remember, the the root cause of that was also of young vegetable seller, a a graduate who whos emulated himself because he didn't have a job. So literally without realizing it, we stumbled upon one of the largest issues of its time at that particular point in time, which led to huge number of revolutions.
Speaker: And in the midst of all that, I helped build a $100 million dollar impact investment portfolio around microfinance and microenterprise financing. Joe did support about a million people in skills and jobs.
Speaker: And our learning was very stark, which is what we brought to Wiser, which is that you know, to make an impact or to move the needle on a big problem, um you need creativity more than money.
Speaker: yeah And as long as you're being creative and you're you're leveraging tech in some way, shape or form, that time we did a lot of work on mobile. And now, of course, has completely taken the world by storm. um You can really move the needle. You can really make lots lots of differences. And ah with that understanding, and I also then discovered how these things can scale only when they're sustainable.
Speaker: You need to have a very strong sense of financial sustainability. So ah with that broad framework and learning over the last whatever five years we spent there, we came back to India with the intent of working in that space and saying, let's do something of our own.
Speaker: And so we were in a sense hammers looking for a nail. And then mental health happened. We had personal experience of suicide and depression and friends and family. My father had bipolar. He was living alone when we came back.
Speaker: So and so that was also my experience as a caregiver. um And that is when we realized that this is a problem which needs to be solved and maybe tech can have a handle in solving it.
Speaker: And long story short, we went through multiple pivots and Viser was set up. We went live in 2017. So ah January 2017 is when we kind of threw it open. You threw open an app.
Speaker: Yeah, as an app. So we put it ah out as an app on the Play stores and... And initially, we weren't sure what would happen. And this was a paid app? No, it was free, but it was still a behind an invite code. So people had to email us to ask for a code.
Speaker: And I was a skeptic, to but to be honest. ah When we launched, I was like, who will it talk to an AI chatbot? And those days, they were customer service bots or bank bots or travel bots.
Speaker: booking boards and they were pathetic. They were really, really poor. So I said, who would want to talk to a bot about their mental health? And I was guinea pig number one ah when we were developing the product. So we said, hey, we will use ah principles from something known as cognitive behavioral therapy.
Speaker: um And ah which is also very suitable for delivery in a chat format. And saying using CBT principles, we will construct a con conversational agent, a set of conversations.
Speaker: And for people who are going through distress and we'll see what happens. Because people who are going through distress find it very hard to reach out and take help. I think those problems are very well known. Affordability, stigma, accessibility, so on and so forth.
Speaker: Sometimes even knowledge, right? You don't even have the vocabulary to understand what you're going through. So and ah we went through a process of development and suddenly ah um me being the guinea pig, I was like, OK, maybe I'd use something like this and we put it out there behind an invite code.
Speaker: And within a matter of days, hundreds of emails started flowing in. And that's when we realized, oh, there's something happening here. So we threw it open. We just couldn't keep up. And so that was early 2017.
Speaker: And usage just exploded. So within matter of days, we were adding 2-3,000 users a day, which in mental health context, imagine a public health system serving maybe 3,000 people a year. And here we were getting 3,000 people a day.
Speaker: How were you funding this? Because this means a lot of server costs. Yeah, this was completely organic in terms of the traffic. So we were not doing any, where people were finding out because they were searching for help online.
Speaker: but But there would still have been the cost of product development, compute costs, cloud costs. We raised a small amount of money. ah The compute costs were not high. They're still not very high for us. um And ah the ah initial, we had some very, I think, ah prescient angels and early investors who almost said, look, we like what you guys are doing. This needs to be done.
Speaker: So here is a check. And that has now grown into something which is now on the way of becoming a very valuable business. So I'm very gratified that we are able to yeah hopefully give them a very, very good return. um But at that time, nobody had a clue. you know Even we were, we set up to solve a problem starting from very, you know, n equal to one. So sample size of me.
Speaker: And from there, we said, what do we want to do next? And ah there was a community there ah which was in pains. And they were using these words like therapy, anxiety, online, on Google Play, ah looking for help. And Vyself was there.
Speaker: And because of which, and even then i couldn't quite believe what was happening. So I was still like, I have no, no, no, this can't be it. This can't be it. Five months after we launched in May, um we got an email from a 13 year old girl from the US saying, I have depression. I tried to commit suicide and you're the only thing which is helping me hold on. So thank you.
Speaker: That's when the penny dropped. And that's when I started believing saying, oh my God, This is something serious. The second um inflection point was COVID.
Speaker: So 2019, 2020, I think the world literally was on fire and mental health became center stage. Till COVID you were surviving on that but one angel check, no monetization. Yeah, angel checks that we had raised a seed and we had started monetizing and trying to monetize in B2C, which is usually the first port of call for any founder.
Speaker: And how how, how was that like like some free plan and a paid plan or. So we were, we had a set of therapists, international therapists based out of India who are offering therapy on text. So there was the, why the chat bot with a plan for digital only. And then you could also, because we also wanted to create that step model of care where it's not a point solution saying, sorry, I can't help you go and figure out something else.
Speaker: So we also felt that there was a responsibility to handhold the person and Either provide higher levels of care or point them towards where they can go. Rather than holding up your hands and saying too bad, i can't help you.
Speaker: um So that was 2019, 2020. And one one quick question ah here again, your traffic, did were there like some major countries responsible for it or was it fairly spread out? or US, I mean, interestingly, i would have thought that we we were based in India and I have no idea how the algo worked out that this global product, 40% of our usage was coming from the US.
Speaker: And which was very counterintuitive and and people were coming back and saying it's empathetic. I'm not feeling I'm feeling hurt. I'm not feeling judged. The Penguin work because people like, oh it's a very chubby, cute character.
Speaker: And people who had bought saying is very body positive. It's gender neutral. And we had not even thought about things like those when we had chosen the Penguin as a mascot. So I think serendipity, I think when an idea's time comes, it comes. So I'm a firm believer in, you know, putting an intent out there and things happen to coalesce around. I think it happened to us.
Speaker: and And that was it. At 2020, we got an email from a clinician in the and NHS, the National Health Service, the public health service in the UK.
Speaker: ah The UK is a single payer system. So this is... Post COVID or like once? Or during lockdown. Just around COVID. Just pre or just around COVID. And this lady was a, she was a psychiatric nurse and she said, are you the founders of Wysa?
Speaker: ah Because do you have any idea what you've done? And we were like, oh my God, what happened? I thought it was a hoax. So then I looked at email and it was NHS.co.uk and I was like, oh my god, it's it's actually real. It's serious.
Speaker: And she came back and said, ah my patients have been struggling to find help when they are on a waitlist and waitlists are really long in the NHS and you're the only thing which is sticking onto their phones.
Speaker: So who are you guys? And we were like, oh, uh, right. Okay. Um, let's meet. And next time we were in the UK, we actually met her and she said, ah do you know how to sell in the NHS?
Speaker: We said, no, we have no idea. She said, let me help you. And you need to do this. And that was a start of our journey into clinical safety. And, uh, then we've done now, we have the world's largest evidence base around outcomes.
Speaker: So more than 40 published research papers, um, and ah which are peer-reviewed research. and But that was the start of the journey. And after that, so that was kind of phase two where institutional adoption started happening.
Speaker: And we started getting inbound interest from employers, from um insurance companies. we have We now work with nine insurance companies from across the world. ah We have a partnership with Swiss Re, one of the world's largest reinsurance companies.
Speaker: who reinsure or reinsure insurance firms. And they had done um research of their own and they'd realized that a large part of their portfolio um book of business had mental health as an unmet need.
Speaker: And it was not being serviced properly. Health systems were not struggling to keep up and deliver services. And that was leading to risk in their portfolio. So they came to us and said, look, we we are a firm believer that there's something like this needs to exist.
Speaker: So we co-created one product with them as well, which is based on the Wiser platform. And that is now being deployed within the Swiss Reclient book as well. So things move from one part to the other. And that is the the what i I'm calling the population part of the business. And I think the second thing in 2022 is 23, 24 is when I think we made a one further step in that part of the business where we said, how do we solve for mental health for India?
Speaker: Because people are not going to download apps. They're definitely not going to pay for them because monetizing and in digital health in India is is really, a really hard problem. But how do you then create something which is sustainable and can scale and actually solve the problem, which is our impact work. So we can talk more about what we're doing in impact, but it's not digital only. It's not it's not an app.
Speaker: So that's the journey. That's been the journey. So yeah, we'll revisit the impact work. um Do you still have like the B2C business ah where people directly download the app? We do. It's a very small part of a business. i think maybe from a revenue point of view, it's maybe 5%.
Speaker: ah But from an impact point of view, it's really huge. So we have ah a plugin on Slack where every time somebody writes a review, So we have a channel on Slack where all reviews show up good, bad, and ugly.
Speaker: And we also keep count of the number of people who have come off, starting from the 13-year-old girl, number of people who came in and saying this was life-saving. And that count is about 500 now.
Speaker: um And so it's great because it makes you get out of bed. And during those dark days in a startup when you don't know what's happening or what's going to happen, it's pulled me through many, many tough spots.
Speaker: Your B2C product is still a world-class product with a free plan and a paid plan and the paid plan and a pathway to a therapist also, like as a standalone product.
Speaker: Pathway to a counselor. Yes. Yes. So we are also very careful to say, um so these are licensed therapists in India. But if you have users from overseas, so we actually call it counseling. and okay Okay. Okay. But yeah it still exists. And it's it's also a great sandbox for us. So because you have such a large pool of users.
Speaker: um So we still add about maybe a thousand people a day. And we don't do any advertising. It's completely organic. And you help people. So it's it's it's good for morale. What's like ah like an MAU number for that?
Speaker: We don't track MAUs at all because in mental health, it's the wrong metric. right im an i man I don't want to talk about how sticky the service is. You don't want retention. And typically, mental health service usage is about 12 weeks and then stops.
Speaker: And you want it to stop. The more important thing is how you're determining outcomes is the person being helped. Okay. so So total number of signups is how you see impact? Like how many people downloaded the app? and We're now at 7 million across the entire platform.
Speaker: Okay. about Had about a billion conversations on the platform now. So it's operating at scale. I'm just wondering why that ah did not monetize or did not scale the the direct-to-consumer business. I mean, in an alternate world, it could have, like people pay for Netflix, right? And if you have to pay whatever, like say 500 rupees for mental health for two or three months, ah it seems like a reasonable ask, but it did not work out. what There must be some macro reasons for it, which you would have learned.
Speaker: I think it's, look, the need exists and it's incontrovertible. um You can't ignore it. But I don't know how much of a pain point it is from a financial point of saying, is it big enough?
Speaker: If I have a broken bone, I will move heaven and earth to make sure that i the bone gets fixed. If I have a heart attack, I'll pay 10 lakhs, 20 lakhs and get a bypass. But if I'm in distress... And there are multiple ways in which people try to solve it. Religion is always there.
Speaker: ah ah Meditation, religion. Other people say, hey, you know, family. are a close family-knit society. Family is, of course, very often the cause of my distress. But whatever whatever reason, it's still there.
Speaker: So there are alternative ways in which people try to solve it. And which is absolutely true. And I don't think there is a... Even though it is... Now slowly being accepted increasingly as a thing which needs medical support.
Speaker: um I don't think it's there. ah Very interestingly, for example, I think one of the largest providers of mental health in India is an astrology platform. Oh, really?
Speaker: Yes. Because in astrology, people when they're in distress and they're worried about the future, they go to an astrologer and say, Swamiji, what will happen? I won't get job. I won't get a marriage.
Speaker: And so on and so forth. And it is, you you need a Trojan horse. You need some way of disguising, making it palatable, making it acceptable for India.
Speaker: um And which is, for example, I think it's a very natural segue into the work we're doing in Impact. We are not delivering a B2C app. in our impact space. What we are actually doing is we are constructing a book, something like this, which is like a workbook.
Speaker: ah And ah for adolescent girls, girls, adolescent girls are particularly vulnerable because of their age and and physiology. And um delivering that through the school system as a physical product, physical plus digital, where it's like a workbook which talks about emotions, talks about emotional self-regulation, building self-resilient skills. It is delivered by the teacher.
Speaker: endorsed by the government and it has QR codes embedded. So now imagine a a girl in Aligarh, for instance, of a certain socioeconomic ah situation, doesn't have a phone of her own.
Speaker: Her phone usage is very carefully watched and monitored by the entire family. ah She probably uses a phone borrowed from her mother and for an hour a day. Now, in that scenario, how do you help the girl?
Speaker: So you give the book and then you have QR codes embedded. it Say you're talking about anger and saying, how do you what does being angry mean and what does anger look like? How does it feel like? Does it have a color and things like those? And then you have an entire conversation about ah anger and when are the times you feel angry? How do you respond and so on, which can be happen ah on a vizier digital front end, but which can be accessed with a QR code and is anonymous.
Speaker: So now you're you're combining both and the the book is called Dream Kit. And it's incredible. We've rolled this out in municipal schools in Maharashtra. Girls have come back to us and said, this my family. The book is part of my family.
Speaker: And we want the the when the girl gets married and goes to a new home, we want her to take this book with her. And there was another instance where one of the girls, so her mother was illiterate. She was a daily wage laborer.
Speaker: And the mother came and said, can you please do WISA invoice in Marathi? What happened? She said, name me my my daughter came and told me saying that WISA tells me that I am fine the way I am.
Speaker: So don't scold me the way you do. And she says, I want to find out what Wiza is saying. And I also want to talk to my daughter the way Wiza is talking to my daughter, but I can't read. So can you please deploy this invoice?
Speaker: So stories like these and the way we are monetizing here is we are going through the school system and we're saying we'll go after the education budget because everybody has budgets for books. Nobody wants to buy an app, but everybody has budgets for books. So if you are able to create social emotional learning and include this into the career, that's the vision.
Speaker: For India, the vision is 50 million kids um and get to them through the school system. Very humble socioeconomic background. This is not a top of the pyramid kind of service.
Speaker: We really want to help the country. and help them at a stage and at a stage of life where helping a girl, manager regulation, regulate her emotions better, building those skills, then helps the family, helps the next generation. It's a multi-generational impact as well.
Speaker: So not only is it good from a financial point of view, but it really, really will change the move the needle. ah for the country. So that's what we want to do in India and that's what we are doing there. Right. so From what I understand, it's ah cultural thing to admit.
Speaker: i mean, it's a it's not in our culture to admit that you need help for mental health related issues and Therefore, the like like that's the basic thing. you're You're willing to go to religion or astrology. Those are culturally accepted.
Speaker: Definitely not pay for it. I mean, 20% of our usage still comes from India, but it's all free. ah So LTV to CAC ratio, right? For a B2C startup.
Speaker: and And actually it's not just an India problem, to be honest. um Globally, even folks like Headspace who invented the online meditation space or Calm, they started in direct to consumer and they the people who made, I think, 30, 40, $50 million ARR businesses in B2C.
Speaker: But the cost of customer acquisition is so high that everybody has now moved to enterprise. And there would be high churn also, I'm assuming. High churn, high CAC.
Speaker: So it's not an easy space. And ah the your price points are such that it's very hard to then ah balance that LTV to CAC ratio. So it's not the unit economics of the business in B2C are not easy.
Speaker: So if there are i think there are people who are trying to do it ah and if people actually solve that, then I would be the biggest cheerleader. For a startup, it solves the problem. But we, I think, reached the conclusion saying it's very hard to solve.
Speaker: A lot of young people, I believe, use ChatGPT for this. Yes, they do. um that's way we We can talk a little bit about Gen AI, but I think there was research which said that came out in MIT saying the top two use cases, one of them was mental health.
Speaker: right Then Anthropik recently released a... um A survey of, I think they did they interviewed 81,000 people, their own users, and on what they would want to use AI for.
Speaker: And personal transformation was number two. 14, 15% of their user base said that we want to use it for personal transformation. The other one was also to make life better. So, Kitchis also has a mental health component to it. Together, they accounted for around 25-30%. So, one third of user base said that they want to use for something like mental health. So, yes.
Speaker: Then the question is, why does there need to be a VISA app for the consumer use case when an LLM can do the job?
Speaker: I think the one word answer would be safety. OK. And um saying that mental health, LLM, horizontal use case by its very nature is something which is used by a lot of people. Can somebody go and develop a specialist LLM or foundational model? There are people who are trying to.
Speaker: And there are startups which are very well funded are also trying to. ah Wiser started life as um actually a not a Gen AI, not LLM.
Speaker: ah In the early days, it was a decision tree. a Simplest way of imagining it was so rules-based conversational agent. ah Like like what you what's called an expert system, I believe. Correct. So, and which is linked to a decision to NLU models, which are understanding user context, invoking responses from a content library, which is written and monitored by clinicians.
Speaker: And with an IFTT, if this, then that kind of... So then and invoking and then and creating that entire conversational tree as the conversation progresses. And of course, you can jump from one place to another, but it's very tightly controlled and so on.
Speaker: And you have escalation paths out at various places. So that was kind of the original architecture. And when Gen AI came about and obviously the conversational capabilities exploded,
Speaker: But at the same time, these are all black box models. So how do you then control them? And there are, I think, also very well-known and very tragic instances which have happened. where So things like ah bias or um inadequate or improper escalation, not being able to recognize ah how or not properly handle high-risk situations.
Speaker: So what we did was we said, look, we need to harness the power of what Gen AI is giving us. We don't want to build foundational models, but we want to do this in a very controlled, tightly regulated, hardness kind of environment. So and there is an architecture which is ah possible in which you can do this, which is what is called a neurosymbolic architecture where and it has two components to it. So one is a neural component and a symbolic component.
Speaker: So a neural component effectively is using the ah robust conversational depth which an RLM provides. ah But it is tightly controlled using a symbolic. Symbolic architecture is very direct.
Speaker: So it is basically saying there are very strong guardrails saying go in this direction, don't go here, don't go there. And of course, then starts going into very proprietary territory here. But then we we use a lot of techniques to make sure that a black box LLM invocation and we use a portfolio of models. And of course, which also helps us deliver this in multiple languages.
Speaker: But There are, for example, filters on an API call going out. So saying, is any PII being exchanged? um Is this a high risk or a low risk scenario?
Speaker: Then having filters which are clinician set and clinician monitored. ah Then the response coming back is also filtered. saying that this is suitable to be played, then escalation pathways, constant and periodic security checks, um especially for detection of high risk scenarios. So we draining data sets, which are used to do this.
Speaker: So there is a lot of infrastructure we've built around a black box LLM invocation of an API to make it safe. Okay, I'll try and ah just tell you what I've understood and you tell me how much I've got it.
Speaker: ah The original VISA was if this then that hard coded expert system that this is how you solve somebody's distress scenario. And I'll maybe ask you a little bit more about the underlying ah theories that it worked on. You mentioned CBT, maybe I'll i'll come to that later.
Speaker: um post chat GPT moment, it made no sense to have that kind of an expert system approach, but it made sense to leverage the power of the large language models. And so now you have an LLM with a checker in the loop kind of a model where the queries, ah in in that conversation, there is a patient who's asking something or saying something. And before that goes out to the LLM, maybe it gets appended with some additional prompt that while the patient is saying this, but there's an additional prompt or additional context, which goes to the LLM and the response of the LLM, which comes back also gets checked to that, is this okay to show
Speaker: to the patient. Have I got it correct so far? that's That's one aspect of it. So that broadly is correct. And there are other aspects which are, you know, safety by design in terms of the architecture in which this is happening.
Speaker: So designing that particular architecture, there is also periodic testing, ah which yeah you need to do to make sure. um So I'll give you another example of hallucinations, right? So it is also very well known that the longer a conversation happens, the higher the chances of hallucination.
Speaker: um So how do you control for that? So that we have a proprietary way of in which we control. um So which is a architectural feature. so it's not quite just a filter of input and output when you're invoking an So there's there's more to it. But broadly speaking, i think directionally that is. yeah And the word I would use constantly to describe this whole model is neurosymbolic. Symbolic architecture, which is creating guidance, control and safety.
Speaker: and inside which there is a neural architecture which is delivering ah depth of conversation. And you need to have a fine balance between both. i'm I'm assuming my checker-in-the-loop analogy is somewhat correct. And so I have a follow-up question there that this checker-in-the-loop would itself be an AI agent?
Speaker: In some cases, yes. But the AI agent is set up with a pathologist. um so and So, there are two or three things we do, but that is one of them. yeah Okay, okay. So, that could be an AI agent or it could be hard-coded if this, then that kind of a, or some sort of an expert system. Yeah, correct. And then also, if you want to comply with, in some cases, medical regulation,
Speaker: then PII storage, PII extraction or scrubbing. So those things need to happen as well. And increasingly what we are also, I'll give you another example of where this gets very nuanced, right? Is ah is SOS, handling SOS.
Speaker: ah Now, ah most and um most horizontal LLMs, for want of a better word, have also caught flags saying, oh, you don't handle SOS well enough or or they're also very scared of liability.
Speaker: So what is the most logical response for a startup? It is to say any hint of SOS... Yeah. yeah yeah google Google does that a lot. Something which is slightly question mark. Gemini cannot. Sorry, sorry, sorry. I can't handle this. Here is a list of helplines you can call. yeah no I think and everybody is obviously doing this with best intent and it is good for the startup to cover liability. But as a mental health intervention or service, that may not be the best for the user.
Speaker: I guess how many people actually make that call? Even if they were in distress. It's probably 10% or so. So the remaining nine people are in a sense left high in drag.
Speaker: And now if you're a mental health service, you're offering an AI-based enabled intervention, you need to think about that a little bit. LLM, this market niche is not big enough.
Speaker: Or maybe it is and maybe you you will have, you know, an Anthropik or a Gemini come back and say, hey, we have a specialist mental, it's a chatbot and and that's it. um But I think it's a niche where you have to have depth and we've chosen to build within in that depth and monetize there.
Speaker: The market size is not large enough for, ah i think, for a horizontal element to say, oh, we're going to go really deep here. But somebody needs to do it. And what we're doing, WISA is actually globally and we're working with IEEE for creating global standards on AI and mental health.
Speaker: And we're also thinking about safety planning. So there is an intermediate step between... When SOS is detected and an escalation happens saying, please call a helpline. And which we think is completely underexplored and underexploited in terms of delivering a service.
Speaker: So if I say that I am extreme distress and I'm thinking of harming myself. Now, there is a concept called safety planning in the literature and mental health literature where...
Speaker: yeah at the time when you're not in distress, you have a conversation with the user saying, if you are ever in distress, what are the things which help you step off the ladder or step step back from the ledge?
Speaker: And is it calling your mother? Is it, you know, going out for a walk on on grass? Or is it immersing yourself in water? Or whatever it might be. And ah create a plan saying, if I am ever in distress, then this is what I would want to do.
Speaker: And I will remind you. And the process of creating and it's pretty sophisticated. So you can create a very sophisticated safety plan and you nudge them at the right time and so that you also don't kind of cheese them off.
Speaker: ah But at the same time, it is there. It's waiting. So if you're ever distress, then you invoke, VISA will invoke the plan and say, hey, you said you'd call your mother. It's this time to maybe make that call. and And now these are this is an example of nuanced intervention or a depth of design, which is both mental health first and privacy safety first.
Speaker: which To answer your question, why not just chat GPT for mental health? I think there is a space and a need to build something which is more intentional, nuanced and deep. You've probably got a lot of conversation data through the consumer app, a thousand people getting added every day. Is that conversation data a serious moat for you? Are you able to use that in any way to create a and competitive advantage? Yeah, we would never monetize. We never do and we will never will.
Speaker: Not monetize, but you can, for example, use it to build a ah small language model. We could, ah we and we are thinking about in that direction. um And it is also very good for um safety ah because you are you have a data set which you can use to test.
Speaker: And it is also informing human process design because then you have conversational data which you can look at, analyze. um And ah it's also ah very useful in proving outcomes.
Speaker: Because it's the B2C user base of the conversations which are happening, you are actually then also able to run real world trials or real world evidence and construct saying, hey, in this particular use situation, this is what happens. So we recently published a piece of peer-reviewed research which said that I forget the exact, I think ah a person who interacted with the bot was, i think, 3x or 4x more likely to complete a um process of therapy as well.
Speaker: And so, which I think constructs a very strong case for a hybrid multimodal level of care. So if you want people to engage and actually complete in dimension or ah that having AI in the loop helps. It's not enough, never would, but it helps. So ah things like those, I think the data is very useful for Things like that.
Speaker: let's Let's talk about how ah the therapy works. ah you You mentioned CBT as the underlying technology when you started. Is it still the same? And what exactly is CBT?
Speaker: ah Yeah, one um also clarification. um so we don't offer CBT as such. So this is CBT inspired. So we borrow from principles, but and we would again, because then that starts also straying into medical device territory. So this is very clearly ah CBT inspired. And we whiched ah use various themes from various techniques. Mindfulness exists. um Then motivational interviewing, behavioral activation are some of the other techniques and principles which we use to drive a lot of these conversations.
Speaker: um But then the there are small elements of DBT which we also inform. But again, it is not it's called dialectical behavioral therapy. so But again, this is not therapy. That's why I'll re-emphasize and say it once more.
Speaker: ah This is really important for us to also highlight and we also tell the user this. Saying this is a conversation. It's supposed to be a wellness conversation to help you in self-reflection. It's a guided journey. But this is not therapy. Please do not use it as such.
Speaker: So help me understand how, like if someone comes and says that ah I lost a loved one and ah I'm just not able to focus on anything, i have no energy to do anything, waking up in the morning is hard, et cetera, then what happens? What what kicks in does ah Do you have some approach of, okay,
Speaker: Since this is the problem, then this is the solution or this is the style that we should use or this is the underlying. Yeah, it was. And this is where it'll stray into the territory where I ask my psychologist colleagues to weigh in.
Speaker: I say, what would you use? But it is informed by. clinical practice. It is informed by a clinical view saying this is the right thing and this is a safe thing to say or this is a safe direction to go down and there are places where after some time you would say this is from this point onwards you need to speak to someone.
Speaker: So and there are those boundaries which we establish within the con conversation as well. so in And sometimes you also use assessments, which also then indicate degree of severity.
Speaker: And that then also gives you a ah pointer towards saying, do you want this person to stay within um the solution or do you want to escalate them out into human support? And these assessments are built into WISA?
Speaker: Into WISA, yeah. In some geographies and so on. So there is a, there's lots more nuance there, but yes. yeah Everything is nuanced because this is like healthcare. This is... Yeah, right. Got it. Okay. ah Help me understand what what exactly for a consumer, like if I want to tell somebody in my family that, okay, you should use WISA. What am I asking him to use? What's he getting?
Speaker: What's he getting in the free plan? What's he getting ah beyond the free plan? So in direct to consumer, I would say that it's a reflective so ah space. So it's a space for self-reflection and which is guided.
Speaker: and and guided with expertise and with safety. So there is a so i would say there is an outcomes or you will get help. And the second statement is you are safe.
Speaker: when you're getting help, this is not the wild west. So I think both of both those, it's a double-barreled message. um And ah from a free plan point of view, it is the basic conversation stays free on direct-to-consumer. And the ah the freemium model, the paid plan, then gives you access to a certain set of tools.
Speaker: which are more specific um conversations for specific use cases. So that then is behind a paywall. And the third element, of course, is if you want to speak to someone and all of this, by the way, is anonymous and direct to consumer. So you stay nameless.
Speaker: um And which is which is extremely useful. And then you can also get escalated and speak to a therapist or a counselor who's a qualified counselor and stay anonymous at the same time. Because again, you can do this on text.
Speaker: So yeah I call myself Rajnikant or Superman and have attack on it. That is also what you want. OK. So the ah the enterprise version of this product ah where you're selling to either governments or insurance companies or large employers, ah There also, do the users stay anonymous or do, for example, employers get a report that ah within your team, these are employees facing some issue or whatever?
Speaker: it's The population base is all anonymous. Okay. Okay. So employers don't come to know that. No, they'll get aggregate reports at a population basis um and or so cohorts. But then we also define saying there's a minimum size of cohort at which you will get.
Speaker: So if you if you say three people complained about the manager and then you're like, I know which who those three people are. okay you don't get You want to get a report that says that. In case three people have to complain about the manager before you get to know.
Speaker: Okay, but but it would flag if if there is some toxicity in the workplace that you would be able to flag that. Yeah, I mean, we have ah we had a client situation where it was a very large healthcare care system with about 200,000 employees in the US.
Speaker: And i think they run about 50 odd hospitals in the US. And they were going through a cost transformation process and one of their sites. And we have these word clouds which are extracted from emotions in the conversation and anger started showing up.
Speaker: And what cloud demand so we were like, what's happening here? Because, like you know, three months ago, there was nothing. And oh, my God, actually, there's something happening here and we need to intervene. So it then constructed. So it was, I think, a beautiful example of analytics, giving them advance notice and which kickstarted an intervention, which hopefully helped people.
Speaker: Right, like like attrition, prevention, employableness. Or at least hi or preventing things from bubbling over. We're actually doing a lot of work on productivity and burnout now.
Speaker: And on the population side, we're also working with travelers. I can name them as a client, which is insurance company, and they do workers' compensation. So anybody who's had um who's had ah gone off work because of an injury, physical injury,
Speaker: Again, process of coming back to work is the barriers are psychosocial, you know, and feeling underconfident, out of the loop. Manager will ignore me. My coworkers have moved on.
Speaker: ah Catastrophizing, saying, oh, my God, that pain in my back has started coming back. So all of these are psychosocial in Asia. So what we actually saw and demonstrated, we we ran a clinical trial and we saw that ah lost days at work actually reduced by approximately one third just by introducing WISA.
Speaker: How do you ah get the Actual usage, like like's say you sign up an employer, lets say an Accenture, which has, I think, a fairly large in lax kind of a headcount. ah How would you get employees of Accenture to actually start engaging on the app?
Speaker: So we're in fact doing a lot of things now. So we want to integrate with MS Teams and Slack and so on. So to to enable discovery, ah because you want to go where the employee is. Right. And yeah um in some cases, that could also be an intranet or a employee portal where people go very regularly and you're there.
Speaker: um So and the third is you then do a lot of campaigns, internal marketing awareness, outreach campaigns to make people aware that something like this exists. the the The HR function would like kind of do that internal communication. Okay. yeah Have you discovered a sweet spot? Like you told me you discovered a sweet spot in US healthcare of the kind of business which would partner with you. Is there a similar sweet spot on ah this side, the population services?
Speaker: are and what we're also seeing is people one where people are assets really assets like services businesses services businesses the second one is ah very interestingly where it's not people are assets but ah PNC risk is high so from an insurance but point of view property casualty so um imagine an oil worker on a rig managing $10 million piece of machinery, you better be sure that he's actually hale and hearty, both in body.
Speaker: Because if not, or ah in fact, there was a very famous, I think, example of a FedEx a truck driver who had an accident because I think he was sleep deprived or he had just gotten off a fight.
Speaker: and And he went and hit some TV personality in the US, which led to a huge claim and lots of very bad PR. So it's a risk mitigation approach. So one is productivity and asset maximization. The other is minimizing risk.
Speaker: So those are the two places where productivity and risk are the two things which i would we are also honing in on. Mental health intervention. Because again, there is a lot of rhetoric on mental health, personal and organizational. But then you do need to connect it to some kind of ah ROI.
Speaker: So in the workers' compensation return to work case, it was actually saying people come back to work faster. we're doing This is not just a nice thing to do. It actually makes business sense.
Speaker: So I think that is really important to to keep demonstrating. Otherwise, it's ah population health, digital health. People are a little wary and they're also a little tired. I think of point solutions. and There's so many people who've been trying to do something in this space. So there's a little bit of fatigue there.
Speaker: And how much does it cost an employer? Like a large employer? Like a couple of dollars per person per year? or Something like that? So it varies, of course, but by volume and scale, but it goes from a few dollars per month to a few dollars per year, depending on scale.
Speaker: And depending on scale. Got it. So how do you grow this business? Like this is through, again, enterprise sales team. Correct. Enterprise sales teams, in fact, our partnerships, GTM partnerships.
Speaker: So our partners like Swiss 3 have really helped us get to scale references, referrals happen as well. So when Swiss Re is working with a corporate for their group health insurance plan, for example, that they could recommend you something like that. Or they're working with an insurance company and taking risk of their books because because it's a reinsurer. So got it. Then they go back to them and saying, hey, you have a book of business. So many lives covered and mental health. so What are you doing?
Speaker: We have a solution, a partner. Are you interested? And okay God, God, interesting. Okay. ah So it seems to me like ah you are a deal maker. A lot of what you have unlocked is by making deals, be it those acquisitions or be it like the Swiss Ray.
Speaker: ah Is there some advice that you have on deal making? The art of the deal? know Oh my God. Yeah.
Speaker: In fact, not the art of the deal. Think about the art of the deal. Yeah, yeah, I'm just pulling your leg here. No, I really do. And ah like now, of course, im I've got many more, many years and hardly any hair left on my head. But whenever somebody comes and asks me for advice, ah even around networking or how do you ah figure out making relationships, I said help other people.
Speaker: And if you help other people, it's good. It's not just good karma. It actually comes back. And I have seen this. Now I have had a 35-year career. I've seen this multiple times. So, and again, it's very similar. It's not just doing good. It's actually helps financially as well.
Speaker: And all our clients, all our advisors, our investors, everybody has come on board and joined the journey because they've believed. And belief and trust takes you a long way.
Speaker: um I don't think I would have been able to raise my first check, my first institutional check. So her first institutional investor was K Capital. So Sasha and Vidushi, who's now on on our board.
Speaker: And Vidushi believed. I still remember that coffee house in which Joe and I went and met her and we said, we want to work in mental health. Many people said, what? At that time, it was basically, what are you doing again? And of course, now that question has changed from what to how.
Speaker: What has been answered. I think the world has answered. Belief. ah how How big was that first round? The institutional round? Your series A, I'm assuming you're talking about. No, this seed. So it was, I think, about a million.
Speaker: And then after that, how many more rounds have you raised? ah So we're a series B funded and we raised about depending on how you use the exchange rate annual anywhere between 20 to 27 million.
Speaker: How much in A, how much in B? And what kind of investors like Indian? or ah it B was about 13 million. um And ah we've been very capital efficient. So we still have a large chunk of that still in the bank.
Speaker: And we've now been in existence for nine years. Builder 150 people, 10, 15, close to 12, about to be 20 million ARR business.
Speaker: And I think we spend maybe about 1.1 and a half million per year on an average of investor capital. How did you fund the acquisitions, those two acquisitions? They were stop swaps.
Speaker: Okay. So the ah founding teams are still there because now they have equity and very strong teams. So very, very strong teams. So, and I think they've been a huge addition to the capability and culture of Wiser. So very lucky to have that.
Speaker: How does... ah the ah what's the journey to an acquisition like? it You know, it just, to me, sounds like you need balls of steel to say, okay, I'm going to go out and buy a company. I'm just wondering. There's so much theory which says that acquisitions rarely work out, right? Right, right, right, right. There's many reasons want not to do an acquisition.
Speaker: oh It goes back to the same thing in my mind, which is that it's it's about culture and trust saying I would never ever do an acquisition if me and the founder didn't get along.
Speaker: but But there must have been some trigger that, okay, we need to acquire. I'm just asking you that journey. like so I think it was basically that. ah So, of course, there are strategic filters and then there's a cultural filter. And then there are economics filters, right? So the strategic filter is pretty much...
Speaker: What is the gap in our business model? Where do we want to grow? Where is growth possible? And yeah that was very clear. saying Increasingly, it was that we do need to have a reimbursement engine because that is where the money is and that's where the commercial opportunity is.
Speaker: And it needs to be linked to what we are doing. There's a strategic linkage. um So and of course, you're enhancing or expanding the model, but it needs to have a very strong foundation or overlap between the two.
Speaker: So those are and so that's kind of the strategic filter. And then ah even before you get to the economic filter of what is the price, what is the swap ratio and so on, you go to the cultural fit, saying both the founder, the team, the cap table.
Speaker: And will you get along? ah So I think there's a lot. and Joe and I spend a lot of time thinking about that, which I think is one of the reasons why we are very blessed to say that these have worked for us and they're actually net accretive and hugely value creating for us.
Speaker: Because I think if you get that wrong um and incentives as well, allowing incentives for for both the incoming shareholders and the founding and the founding team. ah So we spend a lot of time thinking about that.
Speaker: ah How do you align incentives for the incoming team and shareholders? So the stock swap and in some cases doing top ups from a stock options perspective, making sure that they are fairly and reasonably compensated.
Speaker: So, and from a, both from a compensation point of view, and also there is a whole, you know, freedom versus control element saying, uh, we've heard so many horror stories of starters getting suffocated when they get, uh, so be careful not to do that.
Speaker: Uh, so, and, uh, I think the respect, mutual respect is really important. So, because these guys are, i mean, the people, the founders themselves have built businesses in their own right. And they're really, really established entrepreneurs. So, um, them agreeing to join hands ah is also a leap of faith for them.
Speaker: so And respecting that. and And you make sure that it you communicate that respect as well and you get it back. If the person, ah there's some people who will actually see that as a positive and some people who may not. So you also then, you look and then you also judge saying, is this person actually acknowledging the fact that this is happening and recognizing it? And that's a signal.
Speaker: So that's how you cross the cultural filter. And then, of course, there's an economic filter. The price, the deal terms and so on. But I think doing number three first is a bad idea because then you're an opportunistic acquisition may not work.
Speaker: Did you ah basically pitch Indian equity to them? Like is your headquarters in India or like? Yeah. Oh, wow. Okay. So they they came in for Indian equity. that That is itself a sign of like confidence about India.
Speaker: It is. And that actually is very true because um i remember early days of VISA when we launched and we had incorporated in India at that time. And a lot of people told us saying flip to the US.
Speaker: do a delivery C Corp, there an entire established playbook and either out of distraction or laziness or something else, we didn't.
Speaker: Till that time as we go to a certain size and it also then became prohibitive from a a tax point of view. And then we said, okay, fine, we're it. This is it. And this is the structure. And a lot of people have said, you know, your capital pools are much smaller.
Speaker: And then India changed. I think what India used to be in 2019, 2017, and what it is in 2026 is very different. ah The startup ecosystem changed, the depth of capital available, the willingness of people to back deep tech,
Speaker: um All that is very different now. And India is on the ascendant and there are a lot of other places where um stock markets are not done that hot. IPO markets are not that hot. Exit scenarios are not that easy.
Speaker: So it's been very, I think, a little bit of luck. But then probably it's a it's very visible. I think this is a very tangible ah example of where how the Indian startup ecosystem and how Indian founders and companies are seen.
Speaker: So it's the the people literally voted with their wallet. Your equity raise has been largely Indian funds or like? ah Yeah, mainly India, Asia. Google wrote a very small check ah earlier on in our a history. There's one fund, W Health, who invests in the India US corridor.
Speaker: ah Otherwise, yeah, it's been Indian funds. Health Quad was, you know, let a series B. So again, people, VC is backing something like mental health. ah was I don't think I would have said yeah that this is possible ten fifteen years ago. What's the roadmap to an IPO? By when do you think you'll be EBITDA positive? and ah We are hoping to be EBITDA positive in 18 months.
Speaker: Okay. um So, if if you had a it's a very strong trajectory of doing that. Now, whether it's going to be an IPO or something else, who knows? um So it's still early to call.
Speaker: ah But yes, the IPO market is hot. So inshallah is this. Amazing. Amazing. Thank you so much for your time, Ramakand. It was a real pleasure and more power to you.
Speaker: No, thank you so much, Akshay. Thank you a lot. Thanks a lot for having
