Using Audio Technology to Reduce Administrative Burden in Healthcare – Punit Singh Soni, Co-Founder of Suki

Shiv Gaglani: Hi, I'm Shiv Gaglani. As Raise the Line continues to focus on the healthcare provider burnout crisis, we turn our attention today to an exciting approach to tackling a mundane problem: paperwork. Surveys show documentation is the number one pain-point for overworked, burned-out physicians, with doctors reporting that they spend twice as much time on paperwork as they do seeing patients. 

Today's guest, Punit Singh Soni, is the founder and CEO of a company called Suki, that offers AI-powered voice solutions for healthcare, resulting in major reductions in the burden of documentation. Before starting Suki, he was a product manager at Google, and Chief Product Officer of Flipkart. Punit, thanks for taking the time to be with us today.

Punit Singh Soni: I'm happy to be here. Thank you for having me.

Shiv Gaglani: Of course. We like to start off with a pretty basic question to level-set our audience. How did you get interested in technology, and then ultimately healthcare?

Punit Singh Soni: I'm an engineer by training and I started at Google relatively early in my career, and was in the search team in the early days of the company, and then one thing led to the other, and I ended up actually running mobile apps product management there for some time, which was basically an induction into the consumer end of the business. I did a bunch of different things, did mobile gaming and social networking. When we acquired Motorola, I was asked to go run software there. And then, before I started Suki, I ended up going to India where I was the Chief Product Officer of India's largest e-commerce company. I guess it's only logical that I'm doing healthcare now.

I got interested in starting Suki because, when I finished my Flipkart gig, I wanted to do something where time would be a superpower. And what that means is that if you put in enough time and resources, it would really lead to something tremendous. And certain areas more than others actually demand that kind of time and resources, and healthcare is definitely one of them. 

When you start looking at healthcare, then one of the biggest public health crises in the country that very few people talked about is this issue of clinical burnout. We started Suki with the idea of using voice as a tech trend that could actually help impact clinical burnout. There’s a lot more to that. But that's at least the high-level know-how why we started the company.

Shiv Gaglani: Let's get into that, because voice is very exciting ever since obviously Alexa, and Siri, and Cortana, all popularized voice-based applications. One of our favorite thought leaders in the space is Scott Galloway, who talks about voice being the actual metaverse—and Air Pods being a very valuable company in that space.

Nuance and Microsoft: They're going into voice-based AI for healthcare applications. Ellipsis, we had Mainul Mondal, who you may know, is doing some voice applications for psychiatry. I would love to hear, what's the unique proposition of Suki, and how you differentiate? And maybe talk more about the killer application of documentation.

Punit Singh Soni: I think that it is important to reiterate what is the mission of the company. Suki's mission is to make the healthcare tech assistive and invisible, so that you can allow doctors to do what they're good at, which is focusing on clinical work. The product that we started off by building was a voice-based digital assistant called Suki. And the first skill of Suki was clinical documentation.

But the ambition of Suki is much more widespread. These are just different workflows that can make things better. At the end of the day, voice is today where mobile was in 2005/2006. When we were starting to look at mobile in Google and think about the idea of building mobile apps and Android and all of that, there was this sense that in a few years, mobile would become the predominant form factor that everybody would use.

It would be the more predominant user interface of technology. I believe that that's what's going to happen to voice. Why is this going to become the next generation of user interaction model for technology? And in places like consumer tech where you're seeing Alexa and Cortana and Siri and Google Assistant kind of do what it's doing, you're going to see, what that's doing is it's creating market permission. It's enabling people to learn that it's possible to use voice, to actually do it in a lot of use cases.

Now, the interesting part of this is that, as interesting as consumer voice experiences are, there’s basically three or four things we do with our voice assistant: Play a song, check out the weather, check time, set our timers, things like that.

In more constrained enterprise settings, you can do a lot more. Because the number of workflows that you have to do are very limited. and you have do them repeatedly. The ontology that you're dealing with is much more fixed. Therefore, the chances of accuracy are much higher. And especially in healthcare, doctors are actually trained in medical school to speak in a much more structured, thoughtful fashion, which actually lends itself very well for voice experiences.

The summary is, why is the next user interaction model of technology? We believe that consumer tech has just created market permission and accessibility for people to consider voice to be used widely. Healthcare tech is absolutely primed, because of the various vagaries of healthcare, to be the perfect area where voice can actually have a huge dent.

And then finally, as you see, Nuance and M*Modal, who have been basically the last generation's players playing in this: Voice Tech has now improved. There are deep learning models, intent extractors, that understand what you're saying, if you say it flexibly, also. All of that can be actually used in a way to build the next generation of voice experiences. 

The last generation was dictation. The next generation and the current generation is command and control. But there's another generation that will come after that, which will be ambient. And the idea is, there are going to be very large, significant, companies built as every generation of this voice experience comes to the fore. Suki is just basically one of the latest ones in this space.

Shiv Gaglani: That's fantastic. Can you talk to us a bit about the three to four specific applications? I mean, that's what you started a Suki with, but to kind of paint a real-world picture about how it's being used by clinicians in the healthcare system. Maybe give us some of those exact actual examples, and maybe you can categorize them based on the initial dictation, versus command control, versus what you're doing, what features you're working on that will eventually enable the ambient flow.

Punit Singh Soni: Yeah, absolutely. Happy to do that. I think, look, Shiv, the key part of building voice experiences is you need to build what I call the MDP: a Minimum Delightful Product. What is the smallest product you can build that can actually complete a workflow end to end in a way that's consistent and repeatable? Voice is prone to its vagaries. You say, what's my day like? It should understand that, basically, you're asking for your schedule. You say it in a slightly slurry way, there was some noise in the background, the mic was not close to you and suddenly it's not going to understand. There is a lot of complexity in basically making sure the voice experience is consistent. Okay. 

Now, when we started building this Suki assistant, the first feature we focused on was clinical documentation. The reason is that about 30 to 40 percent of all time taken in the administrative work the doctors do is taking notes. The idea is that you can say things like, "Hey, create a new clinical note for Jane Roberts. Insert my normal review system change. This particular aspect of this, insert a surgical consent, prescribe this particular medicine. Suki, stop. I'm done." And push it automatically to the EMR. It seems very logical. It seems very obvious. Instead of typing, instead of dictating, you can just, say, talk to it as if it's a clinical partner. It will generate the note based on what it learned from your prior patient encounters. That's the first feature, first skill. 

Another skill that we build after that was, as you were starting to do the note it would extract these codes—called ICD-10 codes, for example—and put them into the right place in the EMR. These codes associated with the problems are how doctors get reimbursed. Now you have actually really done a much better capture of exactly what the doctor had captured, and put it in as data, so that the loop of revenue can be created.

Then the logical area to go into was show-me commands. Hey, what medications is this patient taking? Can you tell me if this patient has any allergies? Bring me a summary of their past medical history. We just launched that a few weeks ago, where that became a third skill to look at. 

But now you have a fourth skill that will come later on where we start doing orders. Prescribe this medicine, prescribe a lab order, prescribe a flu shot. Then you can go over to inbox management. All those pings that doctors get, saying here, can you update my medication or send me another batch, etc... It an all be automated using voice.

If you start looking at all of these—documentation, coding, order entry, inbox management, show-me commands, clinical decision support, clinical decision optimization—every single one of these Shiv, is a $1-10 billion industry of its own. Now, you're talking about a world where you can take these amazingly complicated workflows, and automate them simply using voice. 

What Suki has a chance of becoming, is basically the next generation of an assistive platform, on which a lot of healthcare interactions sit, inherently making healthcare tech assistive and invisible, so that doctors can do what they do best. That hopefully gives you a summary of the ambition of the company and what kind of workflows are possible.

Shiv Gaglani: That's incredible and certainly, one reason we like having people like you on the podcast is because, there's that famous Wayne Gretzky quote: "You want to skate where the puck is headed, not where it is right now." The people who are listening to this—many of them are currently medical students—by the time they're practicing residency or post, they should be using these technologies in their workflows. 

Going into that, how do you see Suki growing? Is it a B-to-C product like Flipkart e-commerce, or is it B-to-B? Are you working with the health systems directly? Tell us a bit more about where you are as a company, and then the growth path ahead.

Punit Singh Soni: Absolutely. The interesting things Shiv is that, one other thing that surprised me when I got into this space: Obviously, when you spend your time building mobile apps and games and social networking products or e-commerce products or new phones like Mot- X and Moto-G and others, then you kind of get used to this idea that you're always in the center of whatever is the cool thing of that particular time. 

You get into healthcare tech, and it looks like this staid, somewhat boring, slow-moving field. It surprises me because, in my opinion, this is the sexiest area to be in. The biggest technology company ever built is going to be in healthcare. Okay, I don't know when that will be. Will it be in five years or 10 years or 15 years? But the idea that a sector is 20 percent of US GDP, 5-7 percent of every country in the world's GDP, you need it when you're born, you need it, when you die, you need it all throughout.

The idea that there's not a technology company of the size of Google or Amazon or Netflix or Microsoft in this space seems ridiculous. It's just a matter of time. For me, the first part is that if I was going to build a company, I might as well mark around in a space where you think the next biggest, large, huge technology revolution is going to happen. The idea that you can build an assistive platform that can actually start making a real, immediate impact on the users made a lot more sense to me. 

Then the question was, how do you get that technology once you build it in the hands of users? We probably took what I call one of the most difficult go-to markets that exist in business, which is to sell to health systems. The bad news is it's really hard. Sales cycles are painful. There are six levels of bureaucracy in every one of these things. You would be surprised to know that for every one doctor there are three administrative people in the US. 

The good news is that when you actually can add value, focus on the user and get on the other side of that sales cycle, you actually are really well-positioned to create a significant impact and then entrench in the system. 

We have focused our energies on basically, going after these super large health systems and creating serious value for them. We have made our product accessible to independent clinics and small, SMB doctor setups also. If they want to use the product, they can ping us. Two-week free trial, you like it? You start paying. But what happened as a part of doing this, is that we ended up actually building this world-class speech technology, which became a part of the Suki speech platform. 

Then other healthcare tech companies came to us and said, can we have the voice to our products using this platform? So now we have two businesses—a Suki assistant, and a Suki speech platform—both in service of making healthcare tech assistive and invisible. 

How does it grow? To be honest, you can do go to market, you can make sure that you actually sell to the right places. You can price it and all of that. But really a company's fortune is written in serving its users. Making sure that you have a very good sense of what doctors want, and making sure that we can give them technology that for a change, actually makes their life better is a very key part of the philosophy of this company. And if we do that right, the company has no boundaries in terms of growth.

Shiv Gaglani: Yeah, 100 percent, staying very focused on the net promoter score of the end-users. You've talked a lot about doctors, who are obviously the most right now the most influential clinical unit at the basis of a healthcare system. But obviously, there's more and more scope of practice: nurse practitioners, physician assistants. I assume these applications can work across, up and down the spectrum. Maybe even all the way to patients? Are patients using Suki for any reason? Education? 

I know one the patents Amazon got a lot of interest in was when they were able to patent being able to analyze people's voices and understanding if they're sick and then, maybe recommending different things for a cold that they may have via the e-commerce platform. And given your e-commerce roots, I'm curious, have you thought a lot about not just physician-focused use cases for Suki, but then going again, up and down the clinical spectrum all the way down to patients?

Punit Singh Soni: Yeah, very good question. I mean, the first focus of the company is clinicians. We usually say internally that every pixel in this company is in the service of the doctor. But when we say doctor we mean clinicians. You are talking about doctors, you are talking about MAs, PAs, you're talking about nurses. All of them are people that we intend to serve. 

When you get to a point where you have built a product that can serve them effectively, there are extensions to that. The idea that the doctor can create a clinical note in a thoughtful way can also be then used to allow the patient to have access to their medical data. The idea that a doctor can create a plan for a particular patient can also mean that plan information can be used to provide education to the user on how to actually take care of their own health.

There is a world in which I think Suki will be a very good product and have the right data to be used in patients settings. But not right now. Right now, our focus is clinicians. You have to remember in 2016, if you look at all these solutions that are used just for clinical documentation, which is one of the skills of Suki, then the net-spend was approximately $10 billion just in the US. 

The idea that we are spending all this money on just writing notes? You could probably use a chunk of that money and make core healthcare better. We have a very, very, very long way to go. And the key thing for a company with that kind of ambition is to be very focused on who the user is and make sure that they build product for that. And today it's clinicians. And it's in making sure that their administrative burden is decreased.

Shiv Gaglani: Absolutely. Let's talk about the form factor. Is there a mobile app that the physicians carry around with them? Is that the primary use? Or the waiting rooms or the examination rooms have built-in speakers that connect to the hospital system? Was it all of the above and it just magically syncs between all these?

Punit Singh Soni: Very good question. To be honest, there are two points to make here. We learned very quickly that adding any kind of capital investment and humans—what I call humans and hardware—into the system is a recipe for significant pain later on. If you basically build a product where you have to put a physical speaker in the room, deploy speakers, send it to a low-cost location, and have a human being then type notes into it, now you're dealing with large scatters of humans. And you're dealing with a large amount of inventory of hardware. And scaling that is not easy at all. It's not like a software solution. 

In fact, I somewhat mildly joke around saying that a human-based services business is a low-margin, difficult business to scale. Hardware business is a low-margin, difficult business scale. Healthcare is a low-margin, difficult business scale. Is like the holy hell of Trinity. But actually, Trinity of hell, you create healthcare, humans, and hardware in one group for a product. 

We have constrained ourselves by never using a human or using any custom hardware for our product. The second part of this is we have learned that it is important that you meet the doctors where they are. This is a constituency that if you try to tell them this is exactly how you want to do things, they're not going to get it. They have been doing clinical practice for like decades and they have a style and a way of doing it. What you really want is a product that can adjust to whatever their style is.

One way to do it is to make sure that it's available where they are. Suki is an IOS app. Suki is an Android app. Suki can be used on computers as a web app. Suki can be used in PCs, on the Windows application. All of these work with each other. All of these have similar commands. All of these are interoperable over time. 

Basically, you log into your laptop, you're sitting at lunch break, you click on the badge, you start talking to Suki. Then you walk between patients and you pick up your phone and you basically say who's my next patient, give me that information. Then you go over somewhere, and you sitting at a computer and you decide to click on a batch and just dictate something. All of it is just Suki. It doesn't matter where you actually access it.

We have seen a lot of doctors actually just freed up by the mobile. People using it oddly enough to check on their notes in baseball games. Using it, not that we recommend it, but some percentage of them use it while they're driving. Hey, who's my next patient? Create a new clinical note. Change this one thing about this, prescribe that, I'm done. Let's go to the next patient. You end up actually just meeting them where they are.

Shiv Gaglani: Yeah, that's wonderful. It goes back to your first comment about the commonality between building any great company, especially in tech, is user experience, and making sure that's taken care of. And then, to go to market, you have to be mindful that there's no go-to-market if there's no wonderful product behind there.

Punit Singh Soni: Yeah, if you don't have a core product, if you don't have users who believe in the product, if you don't know who your user is, then no amount of business savviness on how you're going to sell the product is going to save you, in the long term.

Shiv Gaglani: In long run. Yeah, exactly. I know we're coming up on time, but I just had a couple of other quick questions. The first is, as you may know, Osmosis is a health and education company. We like to fill in knowledge gaps, and I'm curious if you could snap your fingers and fill in a knowledge gap for anybody interested in healthcare, whether they're going to be on the clinical side or the administrative side, what is one thing you'd want them to know that would maybe help Suki grow or just in general, help make the healthcare system more efficient?

Punit Singh Soni: To be honest, what I would probably talk about is the idea of how do you intermesh and respect clinical and technical cultures? There is a lot of smart people in medicine. They have an amazing mission that they are working on. There's a lot of students who actually look at these blogs and think about building careers in medicine. The truth is that whatever you going to do in medicine is going to be interwoven with technology in the near future. 

Now, what has happened historically is, there is a good, I would say, hardy sprinkling of arrogance between both kinds of people. Clinical people kind of think of technology as IT. And what's the big deal? You have to understand our flow and the tech is just psyche. And you talk to technology people and they're like, the fact that they using all this crappy software means that they probably don't know what tech means. Who are these people? They don't understand tech at all. 

This mutual arrogance has created, basically, some of the mess we have today. Where you notice there are these technology systems that are so antiquated and difficult to use. And doctors who think that they actually know exactly how to solve the problem, but don't think of technology as a real tool but as just IT systems. The key barrier to building thoughtful technology that actually really enhance healthcare and tech is in having respect for both cultures.

There is a culture of clinical professionalism and there's a culture of technical professionalism. And there are two different kinds. One can be actually structured at some hierarchy, but with a very strong mission orientation. Another one can actually be very flat with this idea that not everything is fixed. You can try different things and you'll end up on the other side. And a great belief that tech can actually be the solution to most evils. 

Neither of them are completely right. Finding a way to balance that, and for all those who are listening to this, understanding and having mutual respect for the other side is super critical if you have would actually really take healthcare to the next level.

Shiv Gaglani: Absolutely. And that course or whatever experiential course could be developed for that would be useful not just for healthcare technology and innovations in any field, but clearly society. We could talk about mutual respect between all sorts of groups that are right now are warring.

Punit Singh Soni: You're exactly right. Shiv. There is a reason why the largest technology companies in the market today—in our generation—have not been able to crack healthcare. There is also a reason why the largest healthcare systems on the market have technology that is like 10 years behind the technology my mobile developers used to use when they were building games, 10 years ago.

I think the reason is that we all need to have a lot more significant understanding and know that the next generation of people working on this are going to be both clinical and technical in nature. To me, the barrier to building the greatest healthcare tech company ever built is basically, in the marriage of these two cultures. And the good news is that there are a lot of people who are interested in it. We are firmly on the way of doing that.

Shiv Gaglani: That's awesome. Yeah, I love that. And informatics itself is now a subspecialty—Clinical Informatics. People can go into it, which obviously, you probably recruit out of this feel all the time. My last question, and then an open-ended one: Obviously that's some good advice for anyone listening to this, who's that coming from the tech side, or the healthcare side, how to have that empathy and merge and do something meaningful as part of a company like Suki or starting their own, if they want to. What other advice would you give to an early stage, student who is interested in a career in healthcare technology?

Punit Singh Soni: I would say that learning the basics of what exactly is the core problem you're solving is important. Let's say you want to actually start a company or you actually want to join a company, or you want to join a system because you think that working as a part of their tech setup will actually allow you to provide the most impact you want to provide. 

I would really, ask yourself, who are you serving? And do you really want to serve them? And if you really want to serve them, then what are the two or three main problems they face? Having that kind of what I call almost consumer product management lens to your career allows you to clarify a lot of things. 

What happens a lot of times is people want to get into this space. They go and they get completely bogged down in this system in the bureaucracies of the current present. But if you actually have the orientation of what is it that your user needs and how you're going to solve that, there's a shot that you will make choices that will be much more satisfying for you. 

Whether that is a part of a company like Suki or other companies like us, whether it is a part of a modern future-facing healthcare system that you think has a subgroup that you can work with, or it is actually just simply a part of picking up the education stream that can allow you to learn a lot more of this, to me user-centricity is super important, not just in building companies, but also crafting careers.

Shiv Gaglani: Absolutely, really great advice for anyone interested in their career and making a real impact in the world. Is there anything else that you wish I would have asked that you want our audience to know about before we let you go?

Punit Singh Soni: No, you actually asked most of the things that we wanted to discuss. The primary thing I'll reiterate to everybody is clinical burnout is the biggest public health crisis that very few people talk about today. Not only does it actually hurt doctors, nurses, MAs, and PAs, but it also reduces the quality of data that we have on our platform, which inherently has huge impacts on healthcare innovation. 

It leads to degradation and destruction in the quality of patient care, which has healthcare negatives. And then, on top of it, all the money spent in tiding over this is inefficiency is just money that should be spent on healthcare. We spent so much money on systems that run healthcare that we spent probably less money on actual health.

For everybody who's listening, this is an area that's actually totally worth your time. In my opinion, some of the most seminal companies of the next generation will be built in this space, and they will probably be the seed to building some really thoughtful, real technology companies that can make healthcare tech better. Give it your attention. That's all I will say. Give it the attention and come up with solutions. This is a space that's worth your time.

Shiv Gaglani: I love it. Thanks so much for that. It's very motivating. Obviously, the reason we even call this podcast Raise the Line is we want to improve the healthcare system. One of the core reasons is we can train hundreds of thousands, millions more healthcare professionals, but it's a leaky bucket and people are burning out because of systemic issues like how Insurance works, and how they spend more time in front of a screen than in front of a patient.

Those are real public health problems, as you mentioned, and everyone thinks about it in a very deterministic or binary way. You either have someone practicing or not practicing. It's much more probabilistic because, in the weeks and months leading up to someone burning out or having enough moral injury to leave, you know that for those weeks and months they're dragging their feet. They aren't as effective clinicians as they hope to be. 

It isn't just, we lost a hundred thousand physicians this year to burnout or whatever it may be. It's hundreds of thousands have been ineffective because of these problems. I really wish you the best of luck with growing Suki. It's a really important problem you're solving, and hopefully many of our learners will not only contribute to what you're doing but also be end-users for the sake of all of us in healthcare.

Punit Singh Soni: Thank you, I appreciate your questions. And I think you have a very thoughtful point of view, and a great vantage view actually of all the issues that we're trying to deal with. I appreciate the opportunity.

Shiv Gaglani: Awesome. Thank you again Punit. Good luck with it.  And with that, I'm Shiv Gaglani. Thanks to our audience for listening to this show, and remember to do your part to flatten the curve and Raise the Line, we're all in this together. Take care.