When starting a company, most people like to think big. After all, Steve Jobs didn’t start Apple to make it easier for people to type.
But sometimes those big visions can swallow the small ones—and according to Harjinder Sandhu, the founder and CEO of Saykara, it’s the small visions that keep you going. “Your user doesn’t care about that big vision,” he says. “What they care about is, I have this problem today and how am I going to solve that problem?”
On this edition of UpTech Report, Harjinder discusses his journey to solving the problem of healthcare providers being overburdened with administrative tasks. His solution is an AI assistant that can alleviate those day by day, minute by minute hurdles that interrupts practitioners from delivering care.
More information: https://www.saykara.com/
Harjinder holds a doctoral degree (Ph.D.) and master’s degree (M.Sc.) in Computer Science from the University of Toronto and a bachelor’s degree (B.Sc.) in Computer Science from The University of British Columbia. He is a former Assistant Professor of Computer Science at York University and Visiting Scientist at IBM.
Saykara has developed a platform that uses artificial intelligence (AI) to interpret conversations between physicians and patients, transform salient content into clinical notes, orders, referrals and more, then populate both structured and narrative data directly to electronic health record systems.show more
Earlier in his career, Harjinder co-founded MedRemote, a healthcare-focused speech recognition company acquired in 2005 by Nuance. He then served for five years as Vice President and Chief Technologist of Nuance’s Healthcare R&D division. He also co-founded the patient engagement company, Twistle, and remains on its Board of Directors.show less
DISCLAIMER: Below is an AI generated transcript. There could be a few typos but it should be at least 90% accurate. Watch video or listen to the podcast for the full experience!
Harjinder Sandhu 0:00
It’s always fun and wonderful until you have users. And then once you start having users and you’re selling, it’s like, wait a minute, this is a lot of work.
Alexander Ferguson 0:14
I’m excited to continue our conversation Harjinder. In our first segment, we talked about your platform and AI assistant for physicians and how it’s changing the game for those who are having to well, they want to speed up the charting process and this five, six years in, I’d love to hear more about your story, your the journey that you’ve been on this path five, six years that, how did you get to where you are today? I’d love to know. And you can even start before the five, six years ago?
Harjinder Sandhu 0:43
Yeah, sure. So I’ve been around in the healthcare space in particular, around speech recognition and natural language processing for about 20 years, I started my career as a professor of computer science, co founded a company with a friend that applied speech wreck to the medical dictation space back then physicians used to dictate their clinical notes. And we started applying speech wreck to that we sold that company to nuance, which is the maker of dragons, the leading product for position dictations. And did that for a number of years and went on to co founder another company that’s doing patient engagement, automating just interaction patients, post pre and post facades of care. And then came back to what I really love around speech and an AI in in say, Kara. And so you know, the journey for say, Kara, as I said a little earlier was really centered around, physicians used to just dictate their narratives. And now the game has changed. It’s all about putting data into the medical record. And so there’s this overwhelmingly difficult problem of understanding conversations across any industry, this is a really difficult, difficult challenge. In healthcare, the advantage you have is that physicians are very, they they’re very repetitious. They repeat the same kind of thing. If you patient comes in for shoulder pain, they’re going to ask the same kinds of questions, they’re going to do the same kind of physical exams, they’re going to, by and large, come up with the same cup broad category of assessments and plans. And so once you can start recognizing those patterns of repetitions in the process of the conversation that physicians are having, you have a real shot at understanding what these conversations are about and applying technology to it.
Alexander Ferguson 2:33
Now, in our last previous conversation, you mentioned I asked if you knew something five, six years ago, what would you know, then, you mentioned this just the fact that health care system is so complex? Is that what you said or no, let’s dive into that.
Harjinder Sandhu 2:50
So there’s a the healthcare landscape is complex in a couple of different levels. One is just the IT landscape, you have these electronic health record systems that you have to figure out how to interact with, you have the fact that across many different specialties, of medicine, physician behaviors, markedly different it’s different kinds of problems, physicians, what they document can be very different. So if you’re doing a, let’s say, you have shoulder pain, and you go see a primary care physician, what that physician documents and cares about is different than if you go to see an orthopedic surgeon for that same shoulder pain. And so these kinds of complexities when you’re building an AI solution, ideally, what you want is one kind of thing that you can focus on initially. And what we found is, you know, as we’re building our solution, there’s just so many different variables, so many different kinds of, and then of course, physician behavior is very different. Some physicians like to use the product in one way. Others have completely different kinds of behaviors. And so you have to start adapting to all of this. At the same time when you’re kind of bootstrapping the company.
Alexander Ferguson 3:50
These two previous ventures have med remote and Twistle. Right, both co founded dos. You were co founder and CTO now going into founding another company and being CEO. Any change in your in your perspective and thoughts of being CEO versus CTO and even insights you could share to others who are CTOs now becoming CEOs?
Harjinder Sandhu 4:08
Yeah, being a CEO is hard. I will say it, you know, thinking about the overall overall business. So I had the luxury when I was a CTO of having a great, great friend who was just fabulous as a CEO for both Metro mode and Twistle. And I didn’t have to worry about a lot of the a lot of the business side of things. And of course, when you’re CEO, you can’t let anything drop, you have to think about it all. The biggest thing I think we did about a year ago, is and this came really to a bear when we started commercializing up until then you’re building the product. And you know, everything is great. And, you know, for any product company. It’s always fun and wonderful until you have users. And then once you start having users and you’re selling, it’s like wait a minute, this is a lot of work. And so about A year ago, we hired a president CEO to kind of take on the commercialization side of the company, the best decision we made, I made as a, as a CEO is to just take, take that half of what I was spending a lot of my time on. And particularly for us why was really important was that we haven’t had any point had a product where we could say, hey, the product is done, let’s just sell it now, and go out and just keep marketing and selling it. It’s never been at that stage, we are trying to rapidly evolve and build a product. And, you know, I liken it a lot to, you know, changing the tires on a car while you’re driving, right? You know, we’re we’re evolving this, we’re adding bigger, we’re upgrading the engine, we’re, you know, changing out everything in that car while it’s on the road now. And so you really need somebody that’s actually focused on the road and somebody else that’s kind of building out the the core parts of it at the same time.
Alexander Ferguson 5:54
Funding wise, you guys were were able to get some VC funding early on, what would you say is the biggest mistake one could make when seeking funding and acquiring that?
Harjinder Sandhu 6:07
I think timing is essential. Once you once you have VC funding, there’s there’s a certain clock that starts running, right, and you have to make sure that the product is at a point where you’re really ready to start scaling it or start getting to that next stage. In terms of, you know, once the funding runs out, are you going to hit the milestones for that next next step, and there so I think that’s something that, you know, you have to be be cognizant of, and if you can go further, before you get into that point where you need VC funding, you know, that’s, that’s always I think the best best point is take it as far as you can before you actually need that funding. So the biggest thing is that when you’re, when you’re doing a pitch, and this counts for whether you’re selling to your story to a VC, or whether you’re selling it to a customer that’s looking to adopt your solution, is that you really have to have both a big vision and a small vision in mind. So when we talk to health systems, or VCs, you know, they want to see that big picture, what’s that big vision that you’re trying to sell. In our case, it’s that big vision is we’re gonna have an AI system that can sit next to that physician, and not only document care, but it’s also going to be able to participate in the process care, it’s going to save you money, it’s going to improve your patient, the quality of care, and so forth. Now, that’s the big vision. The question is, how do you get there? And what do you do while you’re trying to get there, right? Because for as a startup, you can easily starve long before you ever get to that big vision. So you have to have that small vision as well. And the small vision is, what are we going to do today? What are we going to do in the product, forget about the big vision because your user today doesn’t care about that big vision, what they care about is, I have this problem today. And I’m going how am I going to solve that problem. And so you have to really be able to cohesively tie together, both that big vision and a small vision and say, hey, if I execute on this small vision, and just take baby steps along the way, I’m getting to this big vision, it’s not a big leap, it’s not a huge chiasm that I have to cross in order to go from here to there. Or if you buy that big vision is important. You know, there is a story right now there’s something that I can do that’s productive and useful today to get there. And I think all successful companies have had that. That combination, whether they articulated that clearly or not, they have that as part of their their pitch.
Alexander Ferguson 8:33
I really appreciate you moving us to the customer, because funding is one thing, but the customer is really what funds the eventual growth and profitability. Any lessons learned that you can share on getting those first initial customers on board that you can then able to grow with and really prove out the concept of the product and scale?
Harjinder Sandhu 8:53
Yeah, no. And that I think, may well be the hardest part in a startup you have your first customer, especially your first significant customer is really critical. And you have to convince them that you know, you’re for real, right. And when you have no customers before that, that’s really difficult. So it’s I don’t know what the if there’s any penisy, any kind of magic formula to this, in our case, because I had a long track record in healthcare, that I could point to, we were able to get conversations with a lot of a lot of different health health systems and medical groups, and you know, some of them were willing to try things and you know, we were upfront, hey, you know, we just need users that are going to try this. It’s going to be a little flaky the very first time you try it, and you know, there’s a lot of bugs and other things. But, you know, are you willing to try it? And in our case, because the problem we were solving for physicians was so big. I mean, physicians literally were, you know, spending hours up to 11 o’clock at night, just documenting their clinical encounters, that they were willing to try it and So we spent a lot of time with some early users there, the first 10 users we had before we actually started commercializing, we gave it gave the product to them for free and just said, you know, just use it, if you find it useful, keep using it. And by and large, we were able to get past the point where the platform was not stable initially. And we got it, you know, kind of worked our way through a lot of those issues and made it work.
Alexander Ferguson 10:25
Any stance or metrics you can share of where you guys are today with the usage of the platform and customers.
Harjinder Sandhu 10:33
So we have over 30, different health systems or medical groups using the product today. So it’s actually getting very rapid adoption. Interestingly, the scale, the uptake has increased since COVID. Started, which was surprising for the first two months since COVID. In March, in April, we had this dramatic slowdown, because all of a sudden, patients weren’t going to see their doctors and doctors were sitting at home wondering what they’re going to do with their time. But obviously, that didn’t last, starting around June or so we started to seeing all the physicians coming back for the most part and starting to see patients again, and then just a lot of pent up demand and physicians. You know, just saying, hey, you know, there’s got to be something better, I need to find a solution. And so I think there’s a general recognition across the industry that solutions exist to solve the problem, physician burnout. And so it’s been really busy for us in the last few months.
Alexander Ferguson 11:35
As your base of customers grow, you need the right team to be able to support that growth, any lessons learned or even common mistakes that you’ve seen people have made when it comes to hiring and building a team that you can advise around?
Harjinder Sandhu 11:51
Yeah, I think as with everything for the team, timing is everything hiring the wrong person, or too high level of person too early can be difficult, or too late can be a challenge as well. So again, I think, for us, finding somebody to head up commercialization at the point where we were starting to commercialize, you know, in retrospect, I should have done it a year earlier, before we started commercializing. You know, so that that is really important finding salespeople that come in at the right time. And the challenge for sales is that, you know, when you’re first starting to sell, you’re not going to be able to grow at the pace to motivate a really, really top notch salesperson. And so you have to be at the point where, you know, they can earn, you know, the level of commission and everything that they’re accustomed to, at the right time. And the same thing is true on the machine learning side. Timing is really important. When we started building our infrastructure, you know, and you’re starting to look for machine learning people, well, most machine learning, people just want data, they want data, and they want to be able to apply algorithms to that data. And when you’re building your overall architecture and infrastructure, there’s really not a lot of data to give them, you’re still building it and the data hasn’t started flowing until users start using your system. And so you have to find the right time to start plugging the right kind of skills into the, into the company.
Alexander Ferguson 13:17
So what’s that indicator for that hiring phases? It just before? You know you needed it, right, just right after a little wait a little bit longer? Where’s the
Harjinder Sandhu 13:27
the magic is for us? Because you also have to pair that with your funding, right? So if funding was never a consideration, you know, of course, you hire everybody right up front, and you have them sitting around until you need them. But that’s never an option. So I would say hire them when not hiring them would actually crash the company. And so if you wait any longer, of course, you’re dead. And if you do it too soon, too much sooner than that, then you’re probably wasting, wasting money like that.
Alexander Ferguson 14:02
How big is the team today? We have 27 people. Gotcha. And moving forward from here going into 2021. What hurdles and challenges do you see you’re going to need to overcome in order to keep rolling.
Harjinder Sandhu 14:17
So there’s a lot of things again, healthcare being what it is, it’s a constantly shifting landscape. COVID has introduced some new variables for us, a lot of the care has moved online. And so we’ve had to adapt to that already. As you look forward, of course, the pandemic, you know, fingers crossed is not going to last forever. And so there’s an open question of what is that mix of online versus in person care, it’s not going to go back to where it was. I mean, now that online care is here, it’s gonna, it’s here to stay. But the question is, what’s the mix and how do health systems and medical groups adapt to that over time, because a lot of this just caught people off guard and they just slap together whatever solutions they could. And over the long term, you’ll see plus firms are merging that do a much better job with online care. And so we have to adapt and, and start working with those platforms better. And then the other side of this for us is that the compensation structure for physicians is a shifting landscape as well, and the way physicians get compensated, changes what needs to get document and how they document care. And so we have to be able to adapt to those changes as well.
Alexander Ferguson 15:27
For you, personally, as a leader, do you have any favorite books, audio books, podcasts, that you would recommend that you found great insight and learnings from?
Harjinder Sandhu 15:37
I wish I had time to listen to podcasts and read books. And all honestly, I spend pretty much all my waking hours just thinking about how to get that next stage of the company. And I think, you know, there’s a lot I used to do a lot of reading before. But you know, since I started, say, Cara, and there comes a point in each company where you start being able to sit back and just think and we haven’t reached that and, and say Kara, so I would say you know, you’re in the trenches, fighting one battle after another. And you know, I’ve read a lot of materials and blogs that give you these words of wisdom and stuff. And, you know, I try to absorb, absorb those lessons, whether it’s from the likes of Bill Gates, or others like that luminaries that have, you know, have successfully built, built things. And you know, that share leadership principles. But ultimately, I think it really just comes down to being in the trenches and fighting, fighting that battle and day to day basis, getting your customer to be happy. That’s all that matters at the end of the day is if you have happy customers, you’re going to succeed if you don’t, it doesn’t matter what you’re building.
Alexander Ferguson 16:44
Last question for you, what kind of technology innovations do you predict we’ll see in the near term, the next year or so, and long term, 510 years from now.
Harjinder Sandhu 16:54
So I think the biggest thing that’s changing and this is really, really what excites me is this is absolutely the right time to be in doing conversational and natural language processing kinds of work. So there’s a lot of really exciting work being done. The likes of Google and Facebook, and you know, all the big tech companies are spearheading a lot of these, I would say seismic changes in the way natural language processing is done. And even within the last two years, if you were to look at the way it was done, even past two years ago, two way it what we’re able to do today, with some of these pre built models that already understand out of the gate, before we even use them, they understand English and how to how to understand some of the basics of other language or other languages as well. That’s really exciting. And that allows a company like ours to just take a lot of that core innovation that’s being done, and start leveraging that and building building on top of it and adding our own novel techniques and innovation on top of that. So what I see happening in the next year is that NLP is really about to undergo in terms of commercially, the kinds of seismic shifts that deep learning introduced, say, a decade ago or so. And that’s, that’s happening now, you’ll see that in the next year or two. And so over the next five years or so what I think is going to happen is that for since the dawn of computing, we’ve been dealing with computers, not on our own terms, but in terms of the interfaces that the computers define, right. So every piece of software you ever installed, you had to sit down at some point and learn what its interface was and how to interact with it. And what you’ll see as a shift as NLP gets better and better is that computers and software will do a better and better job of just understanding us on our terms. And the way that we talk to people the way we’re having this conversation right now. And it’ll mean that there’s a lesson less reliant on a reliance on a particular interface and more and more just natural, natural dialogue. Now that of course, is the the big kind of vision that movies and books have been painting for for the last century. But I think it’s actually getting getting close now and we’ll start seeing a lot of that over the next five years.
Alexander Ferguson 19:15
That concludes the audio version of this episode. To see the original and more visit our UpTech Report YouTube channel. If you know a tech company, we should interview you can nominate them at UpTech report.com. Or if you just prefer to listen, make sure you subscribe to this series on Apple podcasts, Spotify or your favorite podcasting app.