Pradyut Hande (Pradyut Hande:): Welcome everyone to this episode of a podcast series. So This is Asia’s first-ever Martech focused podcast where we have insights and thoughts from thought leaders and martech practitioners across Asia. Today on this episode, we are very pleased to have Mr. Azran Osman Rani. Azran is the CEO and the co-founder of Naluri, but he comes with a very great experience of building a lot of businesses being an investor, being a public speaker, author and the list goes on. Azran is also the chairman today at money match he is an investor and advisor with I flicks and his previous association with Air Asia x where he was the CEO with Astro and that list never ends so it’s a pleasure for Azran to have you today on this episode.
Pradyut Hande:– the pleasure is mine thank you for having me.
Pradyut Hande: – Well Azran, my first question to you will be to explain a bit for the benefit of the listener about your businesses and about this vast experience that you have a building business.
Azran– well, I guess I’ll start by saying: first it may look like a random collection of different industries. So from a heavily regulated capital investment like airlines to the faster-moving technology space. it’s been a common thread are common theme all across all of them, which is I focus on coming into this industry understand the existing business model, and figure out how can we make it more much more accessible and much more affordable and convenient especially for a mass-market instrument in Southeast Asia. so we can look into the existing business model and say how can we do this at 10 times lower cost and make it 10 times more accessible so therefore we built everything from the ground up so that’s been my specific forte consistent across all the different businesses. So I am not very good at training premium brands and services but this is more of addressing the mass market in Southeast Asia.
Pradyut Hande: – Perfect. So now I can see your various expenses also have been in that particular segment. Would you mind talking a bit about nalori. Because that’s a fantastic thing and it’s a digital health platform so how does this idea come into your mind and what is it all about.
Azran – OK, I will always start talking about what are the problems we are gonna solve rather than a particular product that is gonna evolve. The problem of the challenge which we take on that’s a big thing. In our case, it’s a couple of things no.1 chronic diseases are a growing part of our health care. They now represent 70% of the total health care system. In chronic diseases few things like IV’s hormonal diseases and cancer. It’s a lifestyle disease. They are preventable, reversible by doing lifestyle choices. Yet people don’t. People have to pay for their healthcare whether it is Government, insurance companies, or private. The employees are really worried to see how the healthcare expenses are growing by over 10 to 15% year on year. The second part is the gap that we see number 1 is people look at chronic diseases separately from mental health but they are instrictably linked. Now we have data to show that those who are dealing with diabetes and heart disease have a strong correlation with depression anxiety and stress the healthcare system addresses this separately and so we want to think about an approach that we combine the two together right to provide digital health coaching in a way that is much more scalable and accessible.
It is specifically to touch mental health. Here is the problem: write the number one there is a stigma, people are unwilling or uncomfortable to talk about it openly. Number 2 we have a major scarcity of supply. Take Malaysia along there are 13 million adult Malaysian wrestling with chronic conditions and they represent brilliants in healthcare expenses. But as a whole country, we barely have 200 + practitioning psychologists and 400+ practicing psychiatrists. If you rely on the traditional model like the face to face therapy or like tele medicines video sessions like this I talk to you I don’t want to talk to anybody else that’s it in the individual professional I only serve about 50 people a month. This means as a whole country we have only the capacity to help 600-700 thousand people and we gonna ignore million people who just simply don’t have access to the service that is why if you reach out to government hospital you are very less than the 90 problems it’s possible to get better at every station in the month only in the private hospital private clinic of therapy session even this price points are beyond affordable that you are since above the population and we didn’t want to use technology and AI machine learning not to replace professional but to augment them to take that 50: 1 ratio and turn that to 500: 1 and going forward as the data self grows to 1000:1 which means that now we can serve 10 times better to the people with the same level of professionalism so that’s a mission that we set up to do because we have seen in the industry.
Pradyut Hande:: So, very well put up Arzan. So you mention those numbers and you really put into perspective those are really scary and I like the way you are using technology to help everyone in the common mass and things are becoming so critical nowadays the chronic diseases that you mention it’s quite common so you are hitting the mass with technology with a wonderful brilliant initiative.
That brings me to my next question in the covid-19 situation that is there it’s a pandemic that also has cause a lot of depression, economic depression, psychological depression .so what are your observation around this and how is that lead to change and user behavior and doing some of these things will come across.
Arzan: well first clearly, on one hand, there has been a spike in the level of depression and also awareness among employers that they need to figure out how they can support their employees, especially in a remote setting. Suddenly digital health becomes something that people are more willing to explore. Of course, also we have a challenge sometimes as much they want to help this struggle to allocate funding because some organizations are really in tight financial situation and they don’t have the ability to make any form of discretionary spending for the employees it’s not an easy straightforward situation that we say there is a lot more of demand than we can commercialize it.
It really hurts that people have significant financial pressures. The challenge does in a crisis situation asking someone to download an app install the app it’s not effective because sometimes it’s there already elevated and so we had to change our model from purely being using a digital app is the main way of delivering our service we decided to add a few things number 1 we have to use webinar create awareness because otherwise if people don’t understand what exactly they are facing and how help can help them they might not be inclined to reach out for help. So that is important the second part is that we started to create personalize phone numbers to create help lines right in a service where every other client organization has a private number where the employees can call so that we can also understand the usage rate like how many minutes down to how many calls the unique calls from one person making several calls possible people making one call each to create a point where people ll can start to immediately reach out to because it is again when you are under a lot of pressure you are not going to download an app or go through that ongoing process I want to talk to a human immediately. And of course, the third part is, when you go through this you realize there are three different categories of people who need help. number 1 people just want to go. People are angry and are frustrated because they are told to work. Other bosses are sitting safely at home, very anxious about their family and they just want to wait for a few minutes right and then the phone call system works very well. the second category of people who say look and really struggling to make this behavior changes right now I am struggling with sleep I am struggling to eat healthy I am struggling to exercise because I am confined in my small apartment it and that’s where the digital coaching can help so there we ask them to download the app get structured multidisciplinary health coaching in a more convenient manner and the third we are now recognizing there are about 7 to 10% off people where it’s a weird situation where it reaches the level of depression and anxiety that is number one effects of impaired ability to carry on his work affects the sleeping lot of your habits gets into disarray there we have to select a different model in our case video therapy station so we do have to get back to the one on one model because of the variety of the situation so we get to broaden the platform that we use to engage with different user
Pradyut Hande:: Wonderful so I see some innovative changes that you have done like the webinars like still, people are doing it for the personalized phone number the digital coaching the one–to–one approach that you have gotten that is quite significant. Azran keeping this thing in mind today what is the most important KPI for you to measure user engagement or retention
Azran: In our case, though it’s more important than engagement or retention is what we call outcomes. We are obsessed with outcomes. This means in the start of the program one has to have quantitative baselines. We actually quantify the level of depression, anxiety, and stress so we can understand the various levels. Some people are normal or very mild. Some people are quite severe, even in very severe categories. Because our client wants to know the initial distribution, and the second part is how then we measure the outcome at the end right to be able to show what are the clinically significant improvements. But it’s not that I have improved but I have improved a percentage that medical knowledge says that it isn’t a clinically significant improvement example easy to understand physical measures. Let’s say our physical traditional health programs read me measure weight where we measure blood pressure we measure cholesterol we measure blood sugar it’s not about just I lost weight but at least a 5% weight reduction Translate to at least 30 to 50% reduction in the risk of diabetes and heart diseases and you can quantify that, but also measurement on depression and anxiety and stress it says there is significant enough Improvement that did he’s a medical reduction in the risk so outcome first matters most but there are some people who have a lot of engagements but no outcomes there are people with limited engagement but with good outcomes. So if you don’t use engagement as a proxy this is why it is very different from other General and Health Wellness applications. What you talk about is engagement in step count or activities. Activities may or may not improve health. What matters to us is the health organ. Our goal is to see doctors prescribed a digital application as a substitute or supplement to medication. An insurance company needs to say this is something standard and it’s reimbursed. In order to get to that stage, we have to prove a consistent outcome because they’re different parents. We then start to understand what digital behavior leads to better outcomes and that’s the essence of Nalluri. So it’s taking outcomes or engagements to different levels. So now we have something more powerful than engagement or outcomes.
Pradyut Hande:: Perfect ! so that’s given a whole new perspective to it because in general feeling it is the more engagement that you create the better the outcomes are but I think you have rightly said, the focus should be more on the outcomes. The engagement could probably align accordingly. The focus should be more on the outcomes.
Azran: correct because there are some people who are so lucky full stop after two weeks I know what I have to do and I am self-motivated and I don’t need digital help and still I can get outcomes these are the people who just love chat engaged but that not making any help to make behavior change that leads to help improve and this is why it is important to understand because now we have a database of thousands of people who started nalori and ended nalori and measured pre-imposed outcomes we can also layer that against engagement. So we cannot succeed against low investment leads against to healthy outcomes it doesn’t so we don’t take engagement for granted so we have to go down one level of granularity
Pradyut Hande:: Absolutely and I think the analysis which you are doing helps you further to improve your program.So wonderful but these outcomes that are there measure I am sure there are quite some challenges so are you facing any challenges in measuring those?
Azran: yes because the reality is even in our solution for example there is a drop of xxx engagements rates for example at the end of one month 11% of people will say not interested anymore because it’s not the right time to go digital and thereafter three or four months we can see 6 to 8% drop off rates. But in the end, 60% of people are staying and we can measure them. That is still very significant because even 50% of people from the program become healthy and create enough values. For example for someone who is unhealthy may cause an organization to $4000 year in medical cost both direct cost and indirect cost direct cause is your medical and an indirect cost is I need sick leaves I reduce productivity and the present scenario of work. Now if only 50% of people achieve 50% off reduction in risk that means instead of causing $4,000 in the whole year that’s $2,000 but only 50% of people achieve that ratio which is 1000 dollar value. But the price of the organization is less than $100 they considered to be a 10x ROI in so you not everybody and so not everybody joints get a benefit of the portfolio the 10x ROI is a very good ROI for people to spend on it . because they know even if they are prescribed medication , medication doesn’t make everyone healthy. It has a 50% effectiveness rate. So that’s why we compare ourselves to the effectiveness of medication rather than only for the sake of engagement.
Pradyut Hande:: yeah !! And these are health-related matters right. It also has to perceive it slightly differently from the other businesses. There are a lot of intangible benefits as well. It has to be looked at in a different manner but is a very good initiative. You mention at the beginning about those personalized initiatives that you are taking speaking to a lot of brands in the last few days, especially where there is a scale required. Personalization has come out to be one of the most important factors which have been considered nowadays so would like to know your thoughts on personalization and how this scale can be used to better or affect consumers.
Arzan: yes! The sure answer is absolutely critical but we haven’t figured it out exactly. So the first step for us is for example the personalization of the experience, because of the mainly able. We call it B2B and b2c model where the end-user uses our product when it locks off comes from power bi B2B partners where they are using on their own employees so the first step is how do you train the app experience and is this co-branded with the B2B partners so there is a feeling of familiarity. So that’s where we are right now. The personalization that matters to us is because it is human coaches they understand what specific issues and individual goes through and they are connecting on a personal level so far example if you have hypertension I might be a dietitian and a psychiatrist my focus will be more on your salt intake actually the sauces that you put in your food, actually have a lot of sodium content and you may not be interested in focusing in your calories and sugar that’s the health issue that you should be focused on and also where the issues come from you are eating out a lot ordering food we are talking about hurriedness better decision versus someone who is making meals at home and so that’s why we’re the personalized interventions come in. But I think we have a lot more to go because we have got to profile it to the point where coaches can be even more efficient. They can send digital signals and develop very personalized personals. Like for example what’s the right time of the day to engage different people to respond to different times of a day right style some people like drill surgeon type of approach some people like more cheerleader type of approach some people like more academic it’s just telling me the fact approach so that’s the kind of stuff we feel is the next level of personalization where we can understand preferences on how we talk to people.
Pradyut Hande:: fair enough I think I like this so today there is technology available where you can use AI to an extent to personalize a lot of this behavior on the app overusing their engagement. have you thought on those lines have there any tool that you are using
Arzan: yeah! So we use our AI and machine learning on the backend because we are big believers on the front-end about who is the user engaging with that needs to be human. chatbots are the way to mechanical so we want an experience where human will engage the user but the AI and machine learning come in terms for providing decision support to augment psychologist simple example we use a lot of natural language process so when you chat will you forward and then based on the sequence of words we estimate emotional intensity score how to understand whether someone is depressed or someone is on a right positive track towards change are there resilient are there close to change psychologist to understand whose I need to focus on today and how do I focus on a boy trying to use we avoid trying to use chatbots. That means we have to sacrifice responsibilities. People will be like I want to chat, but don’t want to chat immediately. What life changes that don’t work economically then it’s very costly as a whole team of first-line psychologists need to respond right and second we cannot sacrifice that because we have one to answer but it’s a 4-month engagement. response within 24 hours and have to have an ongoing conversation there is sacrifice responsiveness of personalization the human touch to it.
Pradyut Hande:: Understand very well portraits using technology along with the human touch and very likely for industrial catering to. Catering to humans about the AI & ML l that you use your backend. Can you speak about your thoughts on the ML front all coupled with predictive analytics or app analytics that can be used for one for your use specific uses and the others generally for brands across.
Azran: So keep it on the first data sets how do we that is that we I will give you example right a big part of our health and physical health and mental health is whether you eat healthy or not it’s in the food journal where we e it’s like an Instagram picture where we will tap the picture. The photo may be subjective on one hand. Yes, we are applying for image recognition AI instant when you take a picture of a meal. We know what chapati is or if it’s a meal but that’s not enough so we need to create an in-depth label. We can label each food with red-yellow-green. Where green means that it’s healthy choice yellow moderate and read it and healthy and the way we communicate to our user to look at your plate whether it is red yellow or green people start the program may be at 48 size is a photo are red software three to four months of coaching the red should be down to 18%red they are making better decisions this also simplify is it because talking about calories and macronutrients I can’t handle so that’s why red yellow-green is something which is easier to handle. red yellow green is also something structured now I can predict red yellow-green as an input to the predictive model same thing with emotions we use font journals food you also log in your emotions am I angry or am I sad am I afraid or am I happy am I joyful or am I calm like creating categories and asking to identify the emotions so that we can have a quantitative data cells so these data point supplements data point from NLP and pre inputs the output will be something or not when you have enough so when you have enough data set you at least have your learning data set so that you can figure out predicting who is gonna be more healthy so that we can compare the behavior to other people who have gone through the program and the hard part is you cannot go from 500 different inputs one outcome which is is it healthy or not. There has to be intermediate variables to create. That’s the hard part. So now you guys create intermediate industries and what’s the reference point. So at last we decided we will stop using the typical black box AI intermediate and what’s the reference point we decided unexplainable using a technique of black box. So we are set to locate certain interesting psychological models of behavioral changes so readiness to check models, resilience models, and various existing models. So we try to then input and see if we can recreate in the seeds approximate outcomes then you may have simple 6 or 8 then it becomes easy and uses that against your outcomes. So that way you can create an explainable behavior model. Otherwise, it’s a black box I just used the app, and some more it tells me and they can be misleading sometimes, and with previous correlations is sometimes means causation
Pradyut Hande:: very valuable insight Azran. Those points of having those intermediate stages being the fine and then also those models of how it is being used in these are some very valuable tips of using technology in the right way and that brings me to the end of this episode it was such a pleasure and I have to say I have done a few episodes before really some different perspectives and some very good though it will definitely benefit our listeners.