Meet the AI Product Manager Transforming Healthcare

<p>hello everyone this is Cole from the American Journal of healthc care strategy and I'm joined this Saturday afternoon by kov please go ahead and introduce yourself and your role to us hey Cole thanks for having me uh appreciate the opportunity to chat more about Ai and health today yeah I'm KF swarup I am currently the product manager for scout at iterative health and Scout is a AI assisted colonoscopy algorithm so it is meant for colonoscopists in real time to help them detect polyps and so we uh use computer vision and uh AI algorithms to essentially help them detect more Pops increase their detection rates and hopefully have a higher quality of colonoscopies um across um providers [Music] a very cool product and organization and I really appreciate you spending this time on the weekend to join us can you tell us about what your role is as a product manager it's a role that people in healthcare administration might not be as familiar with as you are in Industry so can you explain what that is and what your day-to-day looks like yeah absolutely I'll start a little bit more with generally what a product manager does and then I'll go more specifically into some of the unique nuances with healthcare because Healthcare is of course pretty special as a space but yeah in general I think product management it's an interesting function because it varies so much from company to company and even within a company actually and I'll get into this a bit more it varies quite a fair bit just from product to product right depending on the market the type of um technology or service that you're providing so there's elements of it that are a bit of an art and then there's elements that are more of a science um but in a nutshell I would say uh a product manager's role is primarily to help you develop uh and deliver business impact so you're trying to provide the business value by helping leverage the resources of your team and allocating them or prioritizing them so that you make sure that you're solving first of all identifying and then solving the most important customer problems that deliver the most value and impact to them right so that's like the top line I would say of what a product manager does and you can you could imagine there's multiple parts of this so you want to make sure that it's valuable to the user so a lot of my time is spent being the voice of the customer internally understanding their needs um their pain points what does their incentives look like uh knowing those at the back of my hand and then there's the viability part right and so if you look at Marty Kagan who's part of like the Silicon Valley product group he's written a bunch of books about this but there's like value and then viability and the viability part is is more of the business part right so that's where you look at the competitive landscape the monetization legal and Regulatory constraints of your market and in healthcare there's of course also privacy and reimbursement all sorts of other challenges on the business viability side so you're almost like wearing two hats one is like the strategy hat what is the vision what is the strategy for this product and then there's like the execution hat which you can't do one without the other and be a successful product manager so I would say like on a general scale that's kind of what my day-to-day looks like and where I spend most of my time how did you get these skills I know that you started with a a bachelor from the Georgia Tech right yep how did you develop the skills in that management aspect because I don't think you have an NBA right no I don't a lot of it came from on the job and watching a lot of really successful Executives leaders and folks who are a bit more on the commercial side so my my background was in Material Science and Engineering so I was spending a lot of my time looking at polymers and Plastics and I got into 3D printing right out of school and spent a lot of time developing 3D printed additive manufactured products that go into your mouth so oral health for Orthodontics for dent for Dental products that maybe like you're Invisalign and stuff but slowly I realized like hey through a series of events during covid and restructuring I realized that I had a time to look back and see what are the parts that I really enjoy the most and a lot of that came from watching the sales team the marketing team really go out there and figure out what customers were doing and so the sales team is really strong in in my last startup and so they were kind of helping figure out what what we were building in many parts along with with product and I wanted to get closer to that through shadowing and and then some uh great managers who were willing to take a bet on a young young eager person like myself I was able to fortunately uh jump into the role and kind of get into the deep end of the pool and also just Shadow and learn some of those other non-technical aspects as you can imagine of course and and you did do well in school zumma klady and Georgia Tech is a very well-known institution as well you're no slouch academically I'm sure you could easily get into a top 10 MBA if that was the the path that you needed to take but I like how you didn't necessarily do that right away you went and and did this kind of learning on the job and that's kind what I wanted to ask as well is how did you make that transition into the AI space was that intentional or was that just an opportunity that arose yeah that's a great question I I actually have that question come up a couple of times from folks and actually no there was no intended pivot into AI what actually ended up happening was sides spent several years working on like I said oral health product development and then eventually product management for oral health and I felt looking back at what makes me passionate about making an impact I thought back and my my mother had gone through inflammatory bowel disease with crohn's disease for over a decade and it took her a long time to get diagnosed several years actually seeing multiple Specialists and it happened to be that there was this young startup that was just raising their series B based out of Cambridge and they were starting to hire product managers for solving problems in the inflammatory bowel disease space so they were figuring out how to use technology to identify the best candidates to get access to these new drug trials right clinical trials for example biologics or other kinds of Therapeutics that you can apply to Crohn's disease for example an alterative colitis so that really clicked when I saw that and and I really went after the mission and the vision of the company because it just resonated so much with a problem that affected my mother and the rest of my family and I've seen how difficult it is in in other parts of the world where you don't have access to a lot of the same kind of healthcare that you do here issue I've worked in population Health it's quite an undertaking I want to ask though it's really interesting is so you get a bachelor's in Material Science and you work up to that product manager senior product manager in Material Science and though the product manager role is still the same it's in a very different field it's in field where you don't have that formal education in did you have to prepare yourself for this AI how did you do that how much knowledge do you have now on AI what does that look like from that role yeah no that's a good part of the the learning curve that I think um was very steep I I will not trigger Cod it I think it was not an easy transition to go from uh my previous role which was much more Healthcare adjacent AI was not a component of the 3D printing technology that we were building so I had to learn first I had to learn a lot about the clinical space of IBD even though I had some consumer level information I had to learn about that and then on the AI front I think this was specific to computer vision so I had to First understand just machine learning and then go deeper into deep learning which is which is what the the specific kind of technology that we use for the computer vision algorithms that we build and that came through a number of different ways I'm fortunate ly based out in the Bay Area Silicon Valley where there's where I'm surrounded by Machine learning experts I have friends who are working in self-driving cars autonomous drone delivery all of which are applying computer vision in their day-to-day lives and so talking with them and then also of course consuming a lot of content on YouTube taking some courses online there's some great content that is really simplified down to any lay man can start to understand yeah just a lot of newsletters and YouTube as well as talking to friends was kind of how I got over that initial hump I would say in terms of where I am today I would still say I'm I'm in the beginner to intermediate level understanding U there's a huge Spectrum from where you start off just playing around with Chad BT Alexa Siri right and all the way down to the folks who are actually developing Chad GPT right and so along that Spectrum I still think I have a lot of room to understand how we can best use this technology but I think it's at least at the level where I feel like I can make good decisions alongside my machine learning team and technologists when we work side by side you're also surrounded by these experts as well so your knowledge isn't just thousand hours of your own exploration you're also getting all of this information from these experts in the field where do you think we're going with AI one I always question this myself should I go deep into this and invest in learning Ai and remarket myself as somebody interested in AI or is this just going to fade away in a couple years what do you really thinking and what do you believe is going to happen five years down the road yeah it's an exciting time great question I think you'll see some with any new technology that gets into the spotlight there's always going to be folks who are more on the optimistic side and some folks who are a bit more skeptical I'm somewhere in the like cautiously optimistic side because I live in this technology dayto day so I see the Promise and excitement of some of the advancements that have happened in the last uh two three years even though Ai and machine learning has been around for a few decades and we've had um major inflection points not just in the past two years but for the last decade or so I would say I think if you look at where we are today like generative Ai and large language models are obviously stealing a lot of the spotlight but I think we are in a truly like in a Renaissance of artificial intelligence there are other types of AI like computer vision like I mentioned natural language processing that aren't necessarily getting as much attention because maybe they're less consumer facing right coal and KF don't have access to these day-to-day things on their mobile every day right and they're not as creative they're much more Tech heavy so you don't get to see and use that right like my family and friends can't all get access to everything but they do to chat gbt but I think really there is there's a lot of opportunity especially in healthcare right because it's a $4 trillion dollar industry the costs are rising year overy year in terms of how much we're spending but the outcomes and the life expectancy of at least in America it's not really improving at the rate at which we're spending more and more and it's largely human driven right it's very manual it's all service High skill labor and if you look at some of the thinking from some of the prominent Venture capitalists who are investing in this space they some of them even estimate half of the industry is going to be AI driven because it solves two major problems with Healthcare one which is access a lot of folks aren't able to access the same standard of care even if Rural America or rural parts of other countries have healthc care may not be at the same bar or quality so AI can potentially solve that the second is cost right if you can make a nurse practitioner or a doctor or any specialist more productive the cost can start to decrease over time um especially if you see the cost of AI development and the cost of purchasing AI goes down as computational power gets cheaper and and more commoditized right so I think you you have to look at it from each stakeholder in the healthcare ecosystem as patients right because I guess that's all of us are or will be patients at some point I think that's where it's really exciting to see you're going to get more attention from your doctor if they're less bothered by having to multitask and take notes there's over 40 AI companies that are working on ambient scribe technology right so these folks are going to have uh the ability to let your doctor focus on you during your session and then when you walk away you'll have an AI doctor in your pocket that you can just ask follow-up questions to rather than waiting for that specialist to respond through through email or or just having to Google something and then sifting through thousands of different websites to find the right answer how does your company fit into this what is I think your company is more like you said on the kind of um computer vision angle that consumers don't always have access to what do you guys do and how would a product like yours benefit healthc care yeah no U tying it back to iterative I think is a it's a good analogy to see what happens in two of the other stakeholders right so we talked about patients there're still providers and then Pharma or Life Sciences companies so we have two two distinct products the first product is the one I'll talk about that I work on which is Scout that the role of the technology there is to allow doctors or gastrologists to essentially detect more polyps more polyps means you are essentially catching something that may eventually turn into an adenoma or pre-cancerous lesions and so the computer vision technology allows the the gastron neurologist even later in their day where they may be a little more tired right on a Thursday afternoon post lunch slump is a thing there's data showing that uh detection rates drop off towards the afternoon and evening and so there are other parts of the country where there are less experienced younger doctors who can now everybody can perform at the same level no matter what time of day and no matter how many years of experience they have everybody will be as good as the best gastroenterologist in the world if they all adopt AI right so that's kind of the vision with this provider facing value is you can then detect cancer earlier more consistently standardize these disparities that exist in outcomes so that's with Scout and that's our provider facing computer vision algorithm that that runs for p detection and then the second stakeholder which is uh focusing on our other product and that's for more of a Pharma and Life Sciences facing product this is what works on IBD drug trials that's called clinical trial optimization and there we essentially use Ai and computer vision again compounded that with services so we have a tech enabled service that helps these Pharma companies essentially recruit patients by using our algorithm and services to identify the best candidates and then allowing those those sites to be supercharged in terms of helping them get these patients into the trial right so there there is a provider facing side where we Lia directly with the research site that is trying to recruit patients we work with the research coordinator the principal investigator and then we also work with the life sciences company to give them the findings and insights with our with our patient funnel right so that's kind of how again similar there there's elements of computer vision but then there's also clinical data that we ingest from health records that we're hoping eventually we can use other types of natural language processing or large language models to then quickly sift through and give you the whole picture of let's say KV and say hey based on everything KV is now uh a good candidate for this trial and you can do that now in the matter of minutes instead of having to manually Screen through all the data and and that's where the algorithms come in is they can automatically proactively identify the patient before they even come in for their followup wow very two very impressive products with Scout which I have a few questions about that that I'm you know going to be putting on the spot a little bit so one of the concerns we hear sometimes with AI Technologies and healthc Care is I'm a fan of the technologies that kind of augment like we're talking about now where you still have a a fully trained physician but now he's assisted in being better similar to like Radiologists some people worry that you're going to have all these doctors out of a job now what how does that work does it really not work that way does it need still need a trained physician what do that like people are always worried about jobs going away but I think if you look at the Radiology space I was just on a panel by Microsoft at Microsoft and one of the keynote speakers there panelists mentioned folks were putting out all sorts of statement pieces about Radiologists not existing right and then they were were predicting this would happen 10 years ago or five years ago and look where we are today and there's still in fact actually a shortage of radiologist because a lot of folks as the panelist mentioned are were afraid to get into Radiology because they thought it was doomsday for them so I think there is still a very key role for Physicians to Pivot from doing some of the more manual tasks the more not necessarily grunt work but the things where machines and Technology can be better at right like looking at pixels extracting insights from pixels are things that a human can have variable quality in right and so maybe there are certain tasks that we should let the technology take on and augment us whereas The Physician can then focus on the other pieces because a radiologist still has to connect the dots between for example if you looked at a chest x-ray what are the other components like the family history the maybe there's a a spine Rel at ated injury that could have happened so maybe you need to think about other Specialists consult them and so there's a lot of other I would say creative part or not necessarily just purely creative but all these nonimmediate obvious things that that a physician or human is probably better suited at cognitively all right so I think there's still going to be this symbiotic relationship where I think there will be some jobs that I think for example scheduling back office claims processing revenue cycle management a lot lot of these in payer side provider side those jobs may very well cut down in the number of roles however there will be other roles that will open up just with any other new technology that you will now need for example people to operate drones that wasn't a role that existed 20 years ago so I think that's kind of how I I think about it I wanted to ask about that as well two questions I guess the first one is how competitive was it to get your role not necessarily from a senior product manager perspective but from an AI perspective are we going to have this huge amount of AI jobs that are needed is an AI certification or or reading a book on it or whatever a good investment I know that some people have been worried with the rise in computer science graduates it seems like there's so many computer science graduates out there that there's going to be some layoffs and some challenges in getting a job do you think we're going to see the same thing in the AI space or is it going to grow so much that is a good investment career-wise yeah there's I guess two parts of that question right one which is what type of uh skill set do you need to acquire and is it a worthy investment to acquire those skill sets how do you do that and then two which is more around how much supply and demand is there in AI uh I would say for the first part I think yes it is absolutely a great investment to start with just something right and start learning as fast as you are interested in in learning and I think those Investments that I've made they're not in the order of thousands of dollars right some of them are free right KH Academy YouTube absolutely free content Harvard Amazon a lot of these folks are putting out free content and then there's some I used Andrew's course for example to get an introductory course and then that he has a series of specializations but I would encourage folks who are listening to this to think about what are your superpowers what are you good at first of all and what do you really enjoy doing and if those point to you getting more on the technical side of AI then sure but you know I for one I'm I'm not someone who really enjoys programming 40 hours a week or I enjoy learning about it but I don't enjoy doing that and so I think I felt like my sweet spot was getting somewhat proficient but I think focusing more on other aspects so I think having that Elementary understanding then allows you to say okay from here now I understand the basics where do I want to go do I want to go on the clinical side right like Physicians and nurses they could go into the industry and they can then pivot into care operations clinical operations we have a number of folks on our team at iterative who came from the clinical side and and are now working in a technology digital Health company right so that's like the the first half I think the second half of your question about supply and demand I think yes it's highly competitive I'm not going to lie when I look at when we're interviewing I I I've been helping iterative recruit for other product managers and there has been of course a lot of big Tech and even in the startups there have been a lot of layoffs in the past two years that everyone's seen so there was this massive influx of of uh folks who are available for roles but I think over time there will probably be a higher Supply because the number of startups is just growing like I said there's 40 ambient scribe startups that are working on just documenting from voice the Physicians notes right so those 40 probably didn't exist three years ago so I think the supply is getting there but their bar for hiring is also not lowering right so if you look at their qualifications they want someone who has some experience in developing technology so if you're in the clinical role they do have opportunities for you to Pivot uh if you can get some cursory level uh technology knowledge under your belt right then you will be able to transition probably into a clinical advisory or a clinical data strategy type of role where you need folks who have understanding of how drugs get made right compounds like the bio the Pharma and biotech side and then you also need folks who understand how care gets delivered in Health Systems because folks like me will come in right from a a totally different background we're we're trying to figure out uh the technology side but we need some advisors and we need experts so that's where some of these clinical audience that you have they may be able to still in a competitive environment find their space within the AI digital health or health Tech ecosystem very good advice I think for all of us looking to Pivot or looking to to gain an extra skill set previously when new technologies would come out they were very domination focused we would want Scout in every hospital in the country or every hospital in the world with AI I've heard a few people say it's not that way what is your plan for scout is this something that you guys expect to see in every single Hospital in the country or is is that not what you guys are looking for what's the plan in terms of market share and Market domination with that yeah no that's a great question I think from a like I said product manager's role is to deler business impact so yes absolutely I would love for a scout to be in every endoscopic Suite attached to every every endoscopic Tower there's about 20,000 or so endoscopic Towers in the US probably like 80 to 80 to 100 thousand in the world we intend for scout to be able to raise the standard of Care by allowing anyone who's performing a colonoscopy right for now it's just U focusing on colonoscopies maybe in the future also endoscopies upper procedures but for now at least for screening and surveillance uh coloral cancer procedures we would love every gastroenterologist to say hey I think that there's an opportunity no matter how good my detection rate is already because a lot of them are doing a stellar job many of them are there is an opportunity for me to use Scout to make sure that I am consistently delivering that top-notch care that I give on a Monday morning on a Friday evening right and so that's kind of our vision is that yes we will one have the distribution to to give this access to everybody and many of whom are general surgeons right who are performing colonoscopies they're not Specialists and they can benefit a lot more potentially from Scout but two there is an element of the market that is not quite ready yet right and so we have to face the reality where um it's a fee for service system we need help from payers to reimburse AI technology who's going to pay at the end of the day is the question that keeps coming up in many of these AI panels or webinars that I attend and so we need CMS we need private payers to understand that there is evidence that is growing that shows that hey your cancer incidents will go down as detection rate goes up right for every 1% increase in detection there's a three % decrease in interal colal cancer so that's huge for one of the top three cancers in the United States so we need reimbursement that's the second piece that Scout is hoping eventually we can get there and we need some of the um physician societies to raise the bar for standard right now there's only a a 25% detection rate that is required so that means one in four procedures they need to identify at least one Pollock that's generally achieved by most Physicians but there are folks who are at Double of that so we need physician societies to say hey with AI we can raise the bar to 30 or 35% right detection rates and that's now the minimum because with AI that is where you should be and so that will then cause this Market to then start to say hey all right we actually now need AI to achieve those detection rates on a consistent basis and that's how you get standard of care ultimately in five or seven years hopefully sooner for us one of the challenges that I faced when trying to implement something like this in kind of unrelated area is that when you have consumers who switch plans every three years on average detecting a cancer 10 years in advance isn't always something that insurance companies feel like investing in that heavily are you overcoming that challenge by working with the government more by focusing on those physician associations how are you convincing these insurance companies that this investment maybe detecting cancer a few years earlier is actually going to be worthwhile for yeah it's a very difficult piece of evidence to hand in a sheet of paper to a paay and it takes millions of dollars thousands of patients and many years of interval cancer studies do you need these randomized control trials to be run to be able to generate that evidence we're still fairly in the early days of this technology so we haven't yet been able to have those solid discussions because what they're looking for we're still working on in partnership with with the societies right so the societies have given the AI companies uh like ourselves metronic has a a competing product and a few others they've given us the the vote of confidence and say hey we want to embrace your technology and we're working on generating some of that uh evidence however the government with CMS does give some of these so-called temporary temporary codes billing codes that allow you to Leap Frog or basically supercharge some of the reimbursement in the early days of any breakthrough type of technology so we're hoping that the that the government can at least help provide that extra incentive to say hey it's kind of like a subsidy or a credit for electric vehicles but for pops right you need that sort of extra help in the early days so that's kind of what we're hoping they'll give to the entire category we don't care just about Scout we want this entire category of AI to be able to be accessible by every patient across the world right because they all deserve the highest quality of care and that way after you do that then these insurance companies once you have that proof you have that data they'll be more likely to be supportive financially right and then eventually you'd have the CPT codes the billing codes would would come in uh longer term but hopefully some of these temporary credits that that the CMS provides for some they've provided them for some of the AI devices in the past but it's not it's not as common place as we would love so a few things you mentioned that you go to these conferences you go to watch these panels these YouTube videos all these different resources are there any that stick out that have been especially impactful so my audience can and myself really can gain some of this knowledge on AI yeah I think there's plenty of content in fact there's probably an overload of content these days where it's almost like you get into this initial period of paralysis where you're like oh my God where do I go who do I trust what level of depth can I even stomach at the start I would say think about how you prefer to consume content some folks are very visual some folks love listening right while you're washing the dishes you're taking a walk with your dog go to that specific method that works best for you and certain channels like I love andreon Horowitz has their own podcast that's a venture capital firm that has a significant investment in the digital health and health Tech space they have a podcast that's just generally on the technology space some of which are Health focused and then they have a specific Health focused one which is called raising Health that's a great one then if you love reading some folks are better off with newsletters right so I subscribe to health Tech nerds it's a great Community they even have happy hours and socials where you can meet Physicians you can meet Founders you can meet VCS and so Health Tech nerds is a great Community it's I think it's $150 if I remember correctly for a year of uh the newsletter and and you get access to all these events Fierce Healthcare Rock Health some of these are are also really good content and then lastly I would say if you're more of a social media person Twitter or ex As We Know now that is where I usually start or end my day with like just a 10-minute quick uh pulse check on what's going on in the latest world of large language models people will post uh a lot of Technology focused advancements but then also try to pair that on the tech side with uh some of the more clinical uh Health focused uh content that you can get from some of the other newsletters right so it's kind of a balance of of getting both of these to to kind of round out your knowledge very good suggestions we're definitely going to link those in the the YouTube channel I pulled up but htech nerds here and I'm going to book the mark that page there so excellent advice and thank you so much for sharing your time with us and even devoting a little bit over time I do really appreciate it hopefully as your company continues to grow we can have you back on to discuss maybe some of your research findings and hopefully it'll become standardized practice within a couple years here so hopefully we'll uh be able to Havey back on to discuss that no I appreciate that it's been a great conversation thanks for having me on and yeah would love to be back again in a couple years when hopefully AI is everywhere around us</p>