Podcasts > Ep. 103 - Predictive algorithms for aging gracefully
Ep. 103
Predictive algorithms for aging gracefully
Dr. Hugh Rashid, Managing Director, Xavor
Friday, October 08, 2021

In this episode, we discuss Xavor’s entry in the Smart Home Nursing Market and specifically the development of applications of Smart Home Technology for the prevention of falls among elderly populations. We also explore the opportunities and challenges facing foreign entrepreneurs who seek to enter Chinese markets.

Our guest today is Dr. Hugh Rashid, Managing Director at Xavor. Xavor is both a service provider and a risk taking innovator, with expertise in machine learning, IIoT, Cloud Computing, UX Design, and Embedded Engineering.

IoT ONE is an IoT focused research and advisory firm. We provide research to enable you to grow in the digital age. Our services include market research, competitor information, customer research, market entry, partner scouting, and innovation programs. For more information, please visit iotone.com

Transcript.

Erik: Welcome to the Industrial IoT Spotlight, your number one spot for insight from industrial IoT thought leaders who are transforming businesses today with your host, Erik Walenza.

Welcome back to the Industrial IoT Spotlight podcast. I'm your host, Erik Walenza, CEO of IoT ONE, the consultancy that specializes in supporting digital transformation of operations and businesses. Our guest today is Dr. Hugh Rashid, Managing Director at Xavor. Xavor is both a service provider and a risk taking innovator with expertise in machine learning, IoT, cloud computing, UX design, and embedded engineering. In this talk, we discussed Xavor’s entry into the smart home nursing market, and specifically, the development of applications of smart home technology for the profession of falls among elderly populations. We also explore the opportunities and challenges facing foreign entrepreneurs who seek to enter Chinese market.

If you find these conversations valuable, please leave us a comment and a five-star review. And if you'd like to share your company's story or recommend a speaker, please email us at team@IoTone.com. Finally, if you have an IoT research, strategy, or training initiative that you'd like to discuss, you can email me directly at erik.walenza@IoTone.com. Thank you. Hugh, thank you for joining us today.

Hugh: Thanks for inviting,

Erik: You're obviously not Chinese, but you're now in China on the beach in Shenzhen, can you just share with us how you ended up here?

Hugh: Oh, to China or to the beach?

Erik: Well, let's start with to China, and then we can work our way down to the beach.

Hugh: Literally, back in late 90s, I used to come a lot every month to Hong Kong, we were doing some banking IT projects, and then slowly, some visits to China. And then last 10 years, I used to come for teaching. I used to teach MBA courses; innovation intrapreneurship, come for a month, two weeks. But last two years, I would say is really when I'm spending more time almost 60-70% of the year when we set up our own office here in Shanghai. So it's been kind of slowly gradual, but now pretty much 60-70% here.

Erik: So if I understand your core business, you're providing IT services to Silicon Valley-based companies, and you have your technical team based in Pakistan. So how does China fit into that picture?

Hugh: One of the core areas for our business was data integration. And one of the top companies, still is pretty much way up there in the top 3-4 is TIPCO. They were big in the financial sector, banking, Stock Exchange, mostly all over the world use their technology. So Standard Chartered Bank was implementing that in Hong Kong. So that was a huge four year project for us. So that sort of really got me coming to this part of the world.

Then I saw a lot of opportunities happening on the mainland side and, so grew from there. But integration was the link into China, and we were doing other projects in Asia, also Singapore and in other places. But that was what our secret sauce used to be 20 years ago.

Erik: So, for Xavor is that still kind of the cash cow that pays the bills integration, or has the core business also shifted?

Hugh: We started trying to move up higher value and in more business application. And by that time, ERP was quite mature, and it was too late for us. But PLM was really picking up, which is Product Lifecycle Management, where companies started to not just automate their manufacturing or CRM sales processes, but product development processes. So that became and still pretty much a big part of the business, is helping specially medical device companies or any company, chip companies, which are producing high value products to really streamline their new product development process, the engineering processes. So that's really a lot of stuff we do now in US.

Erik: So you're coming out of a background of understanding these processes and the software used to define them. And now you're basically extending the business into building your own technologies. So what's the backstory then? When did you make the decision to start developing your own solutions?

Hugh: When I looked at China, there are a lot of good IT firms, large ones, the prices are very competitive. So I knew that the same thing we're doing in US would not work here, let alone the whole consulting work, and IT work has to be in local language and culture. So day one, our thinking was it has to be around product innovation. And having been teaching innovation intrapreneurship for a while, I really thought that especially the attention the government is putting into the future of China that they have to move higher value products, and especially using digital technologies, IoT.

I just felt this is really where we need to invest and focus in China. And that's sort of what we have been doing last two plus years. And last year, we really realized that even that is too huge to buy it and we were a little bit distracted with different industries, and we realized, really, we got to narrow down. And that's where the lessons are pretty much year and half it's all about home care aging technologies products.

Erik: So Hugh, you then focused on the topic of aging. If we're looking at IoT digitalization, AI IoT or application of AI to traditional industries, there's 20 different industries you could have selected. How did aging float to the top of your list of priorities?

Hugh: So what happened was pretty much a couple of years ago, a good mentor of mine, and a friend of mine, Dr. Ramesh, and I went to UCI, also from a master's, so he's a dean of the computer science there and also heads the Center of Future Health, and he was coming to China a lot every month advising national labs in China on digital health strategies.

And one of the national labs is in Shenzhen called Panchang Lab. And so I used to meet him a lot. And we started to talk and we realized, wow, this aging challenges China is unlike anywhere else in the world. Basically, they're projecting by 2050, over 475 million people will be in the aging category. That's the bigger population than the whole of US. And of course, then the national policy is around, they call it, internet plus social and virtual nursing home which is your home will be your nursing home, so we really felt that technology has such a role to play, and can make such a huge impact. And it would be really fun trying to solve some of the problems. So, that kind of got me interested through him.

Erik: It was a really interesting market to focus on here. Because on the one hand, you have a traditional healthcare system that is not very mature. So you have some good cheer one hospitals, but outside of that, the hospitals they’re okay, they get the job done, but they are extremely crowded, service levels are pretty low. And basically, you don't want to spend too much time there. So you have on the one hand, the healthcare system, then you have this huge population of aging people, so hundreds of millions of people that are 65+.

And you also have in China a very high receptiveness to new technology. So, as opposed to maybe Switzerland where you might also have a high proportion of older people: the openness to adopting a new technology is quite high here. So it's really a perfect storm of aspects for supporting technology adoption and aging, but there's also probably some regulatory challenges as well. How do you view kind of the, if you were to weigh the opportunities but also the challenges that you especially as a foreigner would face in developing solutions that are touching quite a sensitive area, so personally identifiable information around healthcare topics?

Hugh: No, that's definitely a huge topic, especially these days where just November there's going to be another policy on personal data, and they're really going after and making sure what are you using data for, and you limit that use. And a lot of the big companies had pretty much open hand for a while. So now, when I look at their policies, they're actually even softer in some ways than European policies on this data privacy and those topics. Well, there is a sensitive thing for foreign companies, is about this data storage that you're not allowed to store it outside of China. And also, there are some categories they call it important data, which is sort of national security.

So for us, this product we are developing we call it Care Companion, it's a home robot, it’s going to be a local startup in China with Chinese investors, which is sort of how we look at this startup. It has to be localized. So yes, there is technology. We are licensing some from open source labs, like Dr. Ramesh runs from the Lab Center of Future Health. But eventually, it's a local company which is going to be developing this product and launching it. There is going to be no data outside of China. We won't have a lot of the multinational issues that are quite tricky when they have global operations.

So as far as the value, there is tremendous willingness from government to find preventive solutions. What is happening is this aging population is going to create so much burden on the existing health system. Hospital is, of course, the big one. But also, the nursing homes and all just no way, you can let people become so dependent that they need all the full time care with all these hundreds of millions of people. So prevention is like the old saying is the cure here in terms of long-term solution, which means keeping people healthy in their homes.

So a lot of the technologies we are developing is really all about preventive care, and predicting early symptoms, and doing something about it rather than letting it turn into a disease and then it's a societal cost and the family, of course, the individual aging, suffering through that disease issues long-term.

And so with that, we have a good roadmap of the kind of things we can tackle, the ones which are preventable category. And mobility is one, keeping people on their feet, and detecting early symptoms, which can lead would make them high fall risk. And then that's our first focus. Then mental health is another one of those where there is a lot of technology to predict early what's going on and manage it.

After a certain point, it's just point of no return with these issues. So a lot of the AI technologies we are developing are all about as much early as possible predicting symptoms and doing something about it while the person is in their home.

So in US, they call it independent living aging in place. In China, they call it virtual nursing home. That's the official policy where the government has a lot of incentive. So I feel the government itself is encouraging the use of digital technologies in home.

Erik: This is a very critical area for the country to manage. So you've mentioned a couple of use cases. So it sounds like the priority one right now is predicting when somebody might be at risk of a fall. And then the second is mental issues. Maybe we could focus on that first one. So what does the solution look like that would help list with prediction here and then how do you move from a prediction or some kind of diagnosis to a particular set of actions that the family or medical personnel or that individual should take?

Hugh: So basically, what is happening, and if you think about smart home these days, especially where it's going, it's really a lot of sensors keep getting added to the home, to collect data, what's going on, mostly historically has been for convenience and safety. But the same data can be used to understand the activities of daily life. What is going on, the movement, how long person is sitting, how the movement patterns are? So we sort of kind of started to look at that as a source of early symptoms, and not just understanding detecting these movements, but understanding the pattern of changes because pattern of changes is the one where the magic is.

And there are subtle, small changes in the way person walks, the speed of the footsteps, the distance between footsteps, the balance, subtle changes there. And to those kind of things we feel are really things can be noticed. And that's actually the heart of what the center of future health is, the UCI, they call it ‘Personicle’, which means chronicle of a person which is like a lifelong journal. And then through that, the word Dr. Ramesh uses pattern mining, understanding the patterns and events and then mining them for some kind of causal understanding of what might be happening.

So that's the big part and we call it mobility index, which is sort of an index we develop about the health of a mobility function. And then over time, you sort of notice the decline. For example, a person is low risk. We want to catch that person somewhere when they're at medium risk, not let it go to high risk. Because medium risk, maybe you can reintroduce some early preventive rehab, some exercises; if it's a balance issue, some balance exercise; it’s there muscular strength issue, some strength. So that's where you want to catch.

Unfortunately, what happened is that it's not only high risk, it's after the fall that the rehab gets introduced. The person gets a broken hip, and then they're back from the hospital, and then two years, they're doing rehab. By the way, 50% of those also never fully gain back their mobility function; at older age, it's very hard. So for us, it's all about early pattern detection.

So mobility index has many, many aspects. And we are doing version one of it, which is both predicting decline and also guiding recovery. So, if they are in the rehab situation just using mobility index to reintroduce the right rehab activities, and letting their rehab physiotherapist know how the index is operating and then they make changes to the care plan. And then some other fun features, gamification to keep them engaged. Actually, there are a lot of emotional symptoms, when people have these fall incidents in recovery time, anxiety or falling again, depression, and may never walk again. And that's why the recovery is actually very low.

So the way we are trying to do that is through, technically, it doesn't have to be robot: we could understand this data from a lot of the existing smart home devices and maybe a few other devices, the ambient solution. But at the homegrown solution is a robot, it's a care companion robot, which is not only noticing, but also engaging in terms of positive behavior change, and conversational interface to really understand if there's some pain issues or other issues, just let the right people know.

And interestingly, Amazon introduced a robot just a week ago, but it's coming from their angle of home convenience, and home safety and those kind of things. So we are betters that there'll be enough devices in the future to understand activities of daily life. And the secret of ageing, and home is really using those activities of daily life pattern changes to detect many things, mobility, mental, other issues, so that's why we are quite excited.

In some ways, really, the Chinese leadership has redefined smart home for all people. When I hear their policy document which talks about virtual nursing home, they're really talking about the smart home for the elderly.

Erik: This comment that you made about, you could call it, pressure peer support maybe, but having a social aspect to this makes a great deal of sense. There's maybe a company you might be familiar with called Magic Mirror, which has basically a mirror that you can activate, and it turns into a screen and it uses machine vision to track. So it really to tracks your movement. So you can do yoga, you can do different exercises, and it will tell you if you're doing them correctly. And some of my team uses this, and they say the thing that really gets them engaged is that they're in a group with 100 other people, and they can see over an hour long yoga session how they rank relative to their peers in terms of the number of movements that they did correctly. So there's a lot of pride, if they're the top 10, then think, okay, I'm doing something right here.

And I suppose similar dynamics for an elderly person if they're isolated, they might feel like the thing I'm supposed to do is sit here and watch TV. But if they see that, actually, their friends, their community is out there doing daily stretches, and so forth, and they can say, okay, actually, I should be doing what Sally down the road is doing, and she's actually doing her exercises.

Then we get into this question of the data. So, on the one hand, hypothetically, the government wants to support this smart home initiative. And you can create algorithms that can help to diagnose issues if you have the data. But there's still this tricky issue of how you accumulate the right data. So if you have your own robot, I suppose you have some sensors there you can acquire some datasets to work with. There's a lot more data in the smart home and there's devices being added every year. But actually getting access to that data can be quite tricky.

So how do you approach this? Do you form bilateral relationships? Do you try to create a platform where other OEMs can opt into this platform and monetize data in different ways? And then how do you deal with the ownership issue related to the actual individual who owns that device because I guess hypothetically, the data is owned by that person so they would have to give? So how do you manage this data ownership or access challenge?

Hugh: So first, what you're saying is really what our long term vision is, but we feel we have to independently establish value. So the current robot really doesn't need other devices. We have enough sensors to especially the mobility functions, depth cameras, and some other sensors, we can really understand. Even like you're saying, guiding the exercise behavior, monitoring, or they're doing it right or wrong, lot of those sensors, where that's why we felt we need to have this self-contained product.

But we’re also recognizing we would love to have more data because the more data the better the model of activities of daily life. So variables are easy to integrate. So we have been able to integrate some common variable things. But other smart home devices, we feel, eventually will have to demonstrate the value. Here's the product, which improve the quality of life of aging in place, and then open up some care API's for other people to connect with us.

So one thing we are doing is really adding lot of power to the edge computing because that is crucial, the videos, the processing of pictures, and all, we really don't want that going anywhere into the cloud. We want to process those algorithm indexes right at the edge so that way we feel it's going to still just, beside the legal issue, just the trust and confidence of the family and if you want to keep that way a lot of the data stays at the edge.

As far as monetizing, there are a lot of amazing things going on under this umbrella of data economy 2.0. And we are working with some companies in Europe specially, it's a consortium of about 12 universities which have huge European Union grant to address exactly that issue. How do you establish data ownership, and fair value exchange monetization into these kind of micro transactions? So, we have that in our plan, but again, it's kind of version two topic.

When we built a data model, our business model for version one of the product is really about product sale, recurring monthly subscription, and some referrals. If you are able to refer certain value added services, we get a cut. Data products, we have almost left it like a couple of years out because there are the things you are mentioning are not so simple. But we are fortunate we are connected to some really best companies who are working on this exactly what you're saying. How do you make sure that the ownership is very clear and then these kind of transactions really there's a nice fair value exchange in terms of monetization?

So one of the companies called Data Swift, it's UK based, there are some others, but they all are borrowing the same body of research, which has been going on. So one of our key advisors, Dr. Jung Jin, he’s part of the research and he's giving us the best ideas on that.

Erik: The business model, is it primarily direct to consumer, so maybe a child buys this for their elderly parent? Or is it sell this to insurance, and then the insurance can have a lower risk that one of their customers is going to have to use insurance so they can maybe have a better financial return on insurance package? What business models are you looking at to bring this to market?

Hugh: Well, B2B sounds like would be the first. But you're right about the children. We really feel that's going to be eventually the huge part of the business. Maybe we can look at it like B2B2C because children may want to also get some credible recommendation about the product. Does it really actually prevent fall? So we thought we have to work initially with some reputable legacy insurance companies, some nursing large developers which are building these high end apartments for aging, and even some rehab hospitals, we are in touch with, some government departments who are focused on aging. The challenge is all of them love it, they say, yeah, give us for trials.

So we are at the point where we cannot have enough of those meetings. And in some cases, they've done a press release that yeah, we're strategic partners for this and that. So I'm like feeling this is not a simple a quick app product; you're building a robot, a lot of AI edge computing. So now we are prepared a pitch deck and really are going out to raise money for this local Chinese startup because to do it right and to work with these few giants who are very interested you have to have a staying power, produce few 100 robots, give them for trial, collect data, validate the efficacy, all of that. And then the commercial model will follow in a huge way based on the size of the potential of the market.

But B2C is clearly everybody's telling me you have to target the rich. China has now 6 million millionaires, and there's enough of that population where for them it's going to be a peace of mind for children, especially if they're living far away, okay, there's something which is doing something in terms of prevention, rather than me getting this. Nobody wants that call: your dad or mom has fallen and broken and hip and you're 2,000 miles away. So we feel that some market, but initial start will be B2B.

Erik: So I know, actually a few other foreign entrepreneurs in the healthcare industry, so it's certainly one of the areas where there's in general more participation from foreign entrepreneurs than in a lot of other markets, I think, because a lot of the expertise originates outside of China. But are there any barriers that you see in building these partnerships or because you're incorporated as a Chinese entities, is it basically an open playing ground, and you just need to make sure you have the Chinese team to support any language issues?

Hugh: Well, we basically have some really nice local Chinese, there are professors here on homecare, but there are also professors in us, Chinese American who are part of the project. So they have been able to open a lot of the doors in a very credible way. So that helps a lot to navigate these things.

And secondly, we are kind of an American company, Xavor, which has a [inaudible 26:49] setup here, which is sort of incubating this project. But the startup hasn't been established, it will be established with the investment. So we are kind of in that mode, is it going to be Shanghai company? Where are we getting the government support? Where is the investor coming in? Is it amazing company? So it's an American incubation kind of thing in China, but the startup will be done according to where we get preferential policy investment, and early beachhead for the trials.

Erik: I think this is interesting for me personally also as an innovation model because we at IoT ONE, we work with a lot of corporates who have a legacy, just as Xavor has a legacy in providing certain services for decades, and in then looking at how do we extend into new high growth markets. And for a lot of corporates, the model that you're taking actually makes a lot of sense of saying, let's bring in external capital, let's set up an external entity and run this as a local firm. I think that's probably the most likely path to success for at least some situations, not necessarily all. But it's a very difficult path for a lot of corporates to take.

Because I know you have a lot of relationships here, have you talked to any larger American or European corporates who are also in the space about partnering? I imagined would be some interested in seeing how they could collaborate with you based on this model because I think, quite difficult often for them to do themselves. Or you focus right now on the Chinese ecosystem?

Hugh: No, I think you're really given a great idea. I haven't thought of that. There are companies, American, European in this business for decades with all kinds of infrastructure resources, who may find it. A couple of large consulting firms, one or two of the big four or five, have offered to entertain and from the investment point of view, to see if they can connect, but I didn't look at the foreign corporate, which I think is a very good idea. I'll have to think about it, and we can talk maybe.

Erik: So it sounds like you have a roadmap going forward: you're raising capital right now. For capital raising, maybe just we can dig into this a little bit for the benefit of the audience because I think a lot of people don't really understand the venture markets here in China. Of course, all the big foreign VCs they have funds in China. And then you have a lot of government funds, which might be allocated through universities, through tech parks that are located in different cities.

And of course, you also have the local VC community and then you have the corporate VC community, which could be either foreign or local. You've already mentioned a few large consultancies that have VC arms that are maybe potential investors here. Do you have particular segments? Are you saying we need to have state owned VC because we need that political connection, or are you kind of agnostic from the capital side? What would be your preference on the capital side?

Hugh: Just because of aging and the social dimension of the project, everybody recommends me that it's nice. In this case, it's a good positive that there's some kind of aging social, some government fund or whatever, which is behind. And there is a private local Chinese VC or fund, which is kind of part of that which is connected to the ecosystem. I mean, that's the ideal combination that there's a private fund which is focused on this, and they have a lot of the government funds available, which means the government is behind them to promote these ventures.

But really, I haven't had those discussions. We, last week, finalized the pitch deck. In fact, our financial advisor, who does a lot of these fundraising in US gave us a whole pipeline and process and how we should do the meetings and take the notes. So that's what we were discussing, actually, in the morning, that we are just ready to start that activity. The pitch deck is ready process, but all the things you are mentioning is really what I'm going to find out. What kind of reactions we get? And what kind of model eventually works out? But we have done our best to answer typical questions that investors look for.

Erik: Let me know once you have that polished and ready to go and happy to make some introductions. Hugh, can we go into the tech a little bit more? So you have the robot. Okay, so I guess a lot of the sensors there, you can get off the open market. If you look at the AI side, maybe let's just talk about what's challenging here.

So we have sensors, and then we have edge computing for processing locally. We have actually building the algorithms, I guess, that initial build is still going to be done in the cloud. We have then some kind of prescriptive process for turning the output of the algorithm into a recommendation to the client or a set of alerts that might go out to medical personnel or to family members. So we have a fairly complicated tech stack and set of processes around these. What do you see as the most challenging technical areas in bringing this solution together?

Hugh: The most challenging and where we are fortunate to have the world's best expert, Dr. Ramesh, it's kind of one of those things. Most challenging and most rewarding part is what he called building personal data model. So far all the discussion we have had is about predicting prevention. These are the two magic words in your IoT world when it comes to airplane engines and factories and Intel, all the machines, smart factories. How do you predict prevent bad things from happening early so things keep running?

So Dr. Ramesh, his vision is that the most valuable machine on this planet is human body, why are we not applying all these same technologies to personal health to predict and prevent things? So actually, he has a book just two months ago come out “Pattern Mining: Explainable AI”, data models are something. So that's kind of his sort of our Bible for this whole thing. It's more challenging, and it's more rewarding, and also comforting because he is the global expert in this field. And there are dozens of international labs and PhD is really under his umbrella. And he loves this project. And we are really fortunate.

He's playing the role of CTO. And he says, I'm there, and he attends our calls and guides us. But of course, every call with him is like one month of work for our team. But we are kind of becoming realistic and humble that we got to do so that's why we calling it okay, mobility index version one. There’ll probably be 10 versions before we are really fully satisfied. So we have made it enough where it's useful. But there's clearly a lot of challenges. The world has not solved this problem. The fall is the number one cause of death, is the number one cause of loss of independent living. Is 40% of the reason people get admitted to nursing home.

So everything to keep people at home are linked to predicting and preventing mobility. So we’re not taking it lightly, but that's where a lot of work is going. And you talked about edge and cloud. So one of the things our team does is a lot of these models happen in the cloud. But then we work a lot to simplify, condense those models so we can run them in the edge also. Our team has been able to do a lot of these things. We thought we couldn't do it. But we were able to train, build the model, then push it down to the edge.

Erik: I think that's a great vision. Because if we look forward into the next two, three, five years, there's a lot of work done in building chips designed for ML on the edge and other infrastructure, other new models. So, I think things that might be quite challenging today just a few years into the future, they're going to be quite feasible to do on the edge. So, good to plan around that. What's the timeline here? So you're raising a fund right now, you have R&D going already for a couple of years now, when do you think you'll have a product in the market?

Hugh: That's still, I think, a year away. What we are thinking is that end of the year, we'll be able to demonstrate two use cases. One is this predicting decline. How do robot predicts a decline? We are building a small living lab like a dummy apartment in which the robot moves, and then we can assimilate or have some elders come and demonstrate that we are able to understand these mobility patterns, predict early decline. And then how do you sort of guide them into maybe some positive behavior change excised? Very precise, depending on why their mobility is going down.

And the second is what you were talking about like this mirror and peloton type exercise where rehab exercises. But in our case, what we want to demonstrate is the precision that in older age, after you're come back from hip surgery, you have to do this very soft, subtle exercise, not too fast, not too slow, the movement has to be precise. If you feel your breathing high, you should be guided to slow down. So we are really adding like peloton on steroids for elderly.

Because it can be just go, go, go, let's go. We want to monitor their heart rate, so make sure they're wearing wearable, ask them to read, give us the reading. We can connect also. But we felt sometimes it could be a simple conversation, hey, slow down. There are some medical questions about pain detection. If it's too much pain at a certain level, you want them to stop, monitor their activity, find out what is the best sweet spot time when they really can do these exercises.

So for us, it's like just all the time keeping an older adult in their late 70s who's come back from a hip surgery in mind when we are designing these recovery exercise, and the conversations and understanding empathy. So, in fact, empathy is a big part of the whole product to really show that not just say why are you not exercising or it's time you miss three exercise? Well, they may be valid reason.

Erik: And we didn't really get into the interface. But it sounds like the interface here is going to be voice, which I guess is another set of challenges?

Hugh: Yeah, a big part is voice. And the subtle difference between lot of the Alexa and other toys is they are reactive, you talk to them. In our case, the robot is the one initiating conversations based on the personal model, based on the pattern changes, based on the activities of daily life. I think that's a big part of the challenge that you're earlier mentioned. Having a data model, having being able to predict these patterns, and then driving conversation, not just how are you? Just not some hollow conversations.

Erik: Well, then there's somebody else here that I should introduce you to. There's a company I was talking to just last week that's building AI process algorithms for right now focus more on automotive OEMs for customer service management, but really finely tuned algorithms for different purposes. And I think they have a very interesting and clear vision on how this will actually work. Because in the end, you can't build a generic algorithm that solves every problem that somebody might have. You have to somehow build a set of different algorithms that will solve different issues.

But from the user experience, it feels like I'm talking to this one robot, but in the back end you have maybe a more complicated architecture and algorithms designed for different requests or processes. But Chinese company, of course, could have a good perspective, and platform for building this, so maybe be an opportunity for a partnership there.

Hugh: Oh, very good. Sure. No, I expect we'll end up working with a lot of companies because we are using Microsoft Cognitive Services for something, Chinese companies for some of the local translations and for many things. Then you're right, all kinds of architecture hidden. But the front there's a very small conversation which makes sense, has the caring empathy in it, and the older person wants to engage.

Erik: Hugh, thanks for walking us through the process of pivoting your business and setting up this new entity or business here around the topic of aging in China. Anything that we haven't covered that you think would be important to discuss today?

Hugh: Basically, what happened is that as we started doing this thing, then there are some other products we didn't get into it. We are really becoming like this age tech incubator. We got some other small products in the pipeline, even things like a garbage bin. Because in aging, the nutrition and diet is such a critical part. And you can't expect somebody in their 80s to be fiddling with an app and telling you what they ate.

So we kind of want to use the garbage bin as a proxy. What are they throwing: fruit, vegetable noodles, that kind of stuff and have some understanding of really what might be happening in their diet. And so things like this, we have about six, seven of these little products. And the goal is that eventually, they all use this one geriatric data platform where your holistic care data is available. And just imagine for food is totally different people who need to know that is the diet for this person or not. For mobility is different physiotherapist. For mental, it's totally different people. But it's the same person.

So going back to your digital twin, to do this digital twin, we expect there'll be all kinds of companies. But currently, the fundamental problem in this is that for every product, there is a different tab, there is a different data platform, there's a different alarm system. And I visited some very fancy high end these kind of aging home with 35 sensors each with its own story, somebody got red lights, somebody got alarm, somebody got abs, somebody got cloud. It's just not going to work from aging person perspective, especially if it's an older person living alone.

So what our vision is that really create this platform, which is open for other products to add in. And even our own, some of the products we want to create should use the same ecosystem of platform app and those kind of things.

So in that way, I'll look at our company pretty much next 10 years doing nothing else. That's pretty much our roadmap. And we'll be looking to partner with all kinds of companies who are into aging technologies, and sometimes licensing technology, sometime using their technology. So that's how I see this happening.

Erik: So this is certainly a market that you can focus on for the next few decades and not get tired. I'm thinking, my wife would probably be interested in knowing how many bubbles [inaudible 42:54] her dad goes through on a typical month?

Hugh: That's a good point. These kinds of things can be any if nothing else and that's the hardest part probably. I talked about data model. But I really think after all said and done, it's going to be the relationship between the robot and the older adult. Is it lovable, caring, empathetic enough because it's very easy older adult to unplug the thing with this annoying, nagging thing.

So there's going to be lots of interesting design challenges all which are never easy. So that's why we feel it's going to need series of rounds of investments and trials and all kinds of things. But, hey, this is the market for next 20-30 years in China, which I think is going to be the largest market.

In terms of GDP, right now, it's about 7-8% GDP goes into aging care and all the services. By 2050, it's expected to be around 25% of China's GDP. And that's because based on the current pattern, and what we feel is the only way it can be come down is if this healthcare turns into preventive health and self-care. So self-care is our driving word that drives us. How can you let older people that give them dignity, quality of life independence?

Erik: It'll be really interesting to see how this works out. Because you have, on the one hand, the very hard tech side of building an algorithm that actually makes accurate predictions, and on the other side, this very soft side of how do you communicate to somebody around very sensitive topics in a way where they actually will want to listen to you and not turn this solution off because it's just a nagging presence in their life.

Hugh: Absolutely. So we are in touch with a lot of top design firms, even some university in New York. University actually, their whole design class is working with us on some of these challenges. So the teacher and some of their loved it, they say, okay, we have this fall semester. Whatever this semester is started in September, the whole class is working on some of these challenges and very excited. And so that's what I think will take, top minds, top design school professors, students to come up with all the interesting solutions for these very tough, challenging questions.

Erik: Well, Hugh, I wish you all the success in the world. And like I said, keep me updated as you move along this, happy to see where I can help.

Hugh: Okay, great. After the holiday, I'll be back Shanghai, we’ll set a time and catch up [inaudible 45:41].

Erik: Hugh, enjoy the beach in the meantime.

Hugh: Okay, thank you. Enjoy the very good session, perfect questions. As always, Erik, you're a great host in all kinds of events, so including I always enjoy attending your event and same as well.

Erik: Thanks for tuning in to another edition of the IoT spotlight podcast. If you find these conversations valuable, please leave us a comment and a five-star review. And if you'd like to share your company's story or recommend a speaker, please email us at team@IoTone.com. Finally, if you have an IoT research, strategy, or training initiative that you'd like to discuss, you can email me directly at erik.walenza@IoTone.com. Thank you.

Contact us

Let's talk!
* Required
* Required
* Required
* Invalid email address
By submitting this form, you agree that IoT ONE may contact you with insights and marketing messaging.
No thanks, I don't want to receive any marketing emails from IoT ONE.
Submit

Thank you for your message!
We will contact you soon.