[00:00:01] Welcome to the Chalk Talk Jim Podcast, where we explore insights into healthcare that help uncover new opportunities for growth and success. I'm your host, Jim Jordan.

[00:00:16] Welcome back to the podcast where we explore groundbreaking innovations in technology and healthcare. Today's guest is Akshay Monga. He's a technology strategist with over a decade of experience solving complex problems across all sectors.

[00:00:31] From pioneering advancements in automation in major global firms to leading innovation in healthcare at Highmark Health and Amgen.

[00:00:41] Akshay's journey is nothing short of inspiring. In this episode, he shares his insights on how data, AI, and cutting edge technology are shaping the healthcare and pharmaceutical industries.

[00:00:51] We dive into topics like precision medicine and overcoming data silos and the future of AI in drug discovery and clinical trials.

[00:00:59] If you're curious about how technology is transforming healthcare, this is the episode you don't want to miss.

[00:01:06] So Akshay, tell me a little bit more about yourself.

[00:01:10] Yeah. Hey everyone. I am Akshay Monga and I have like over decades experience in solving complex problems, especially in the tech sector.

[00:01:19] So I started my journey in Dubai right after my engineering at this big German conglomerate called Thiessenkrupp.

[00:01:26] And back then, this is like a decade ago, one of my pet projects was how to get robots to talk to elevators and use an elevator.

[00:01:36] So I've been tinkering with the tech sector ever since I graduated from my undergrad.

[00:01:42] And I believe like my life is serendipitous.

[00:01:46] And as I was looking at my B school options, I had a chance meeting with the admissions director at Johns Hopkins.

[00:01:53] And that brought me to US to do my MBA on a scholarship.

[00:01:58] And it is at Hopkins. While I was studying there, I got to learn about the healthcare sector because Hopkins is synonymous with healthcare.

[00:02:06] So you get to talk to the brightest minds, exchange ideas, learn about healthcare.

[00:02:11] And it was like a great honor and an experience.

[00:02:13] While I was doing my B school, I got to do my internship at Amgen.

[00:02:20] And that brought me into really into using technology in the healthcare space.

[00:02:25] And that's where I was like, hey, my passion for technology, healthcare really needs the tech to come in and bring about better outcomes for everyone.

[00:02:34] And it's like a perfect blend.

[00:02:36] That sort of even gives me personal value as I look at day to day.

[00:02:40] I have to ask because I saw it on your, you did a small consulting opportunity or internship at NASA Langley.

[00:02:48] What was that like?

[00:02:50] Oh, that is pretty interesting.

[00:02:52] And that's the advantage of doing B school.

[00:02:54] You get these chance opportunities and you just jump on it.

[00:02:58] They had this technology which they wanted to commercialize.

[00:03:01] And as typical B school students, you can look at the evaluation, do the DCFs and okay, put a dollar value to it.

[00:03:09] Do the market assessment, figure out where to focus, who to sell this technology to.

[00:03:15] How do you license that technology?

[00:03:17] Set that pattern end to end for them.

[00:03:19] And it was one of the great experiences that you can take out of working at a B school.

[00:03:24] So many people say that healthcare is like rocket science.

[00:03:28] And so I just thought I had to mention that it definitely is.

[00:03:31] So you've done the manufacturing side, but you also did the insurance provider side.

[00:03:36] That's true.

[00:03:37] As I graduated out of Hopkins, I got my first role at Highmark Health.

[00:03:41] So Highmark Health for folks who don't know is a blended healthcare system, which means they have a payer and a provider.

[00:03:47] So peer is an insurance company, provider is a hospital system.

[00:03:52] So within Highmark Health, I was working in the strategy team and we looked at everything from like a M&A, which is mergers and acquisition companies to acquire in the insurance area, all the way down to tech partnerships, whether it's specifically for insurance use case.

[00:04:06] And back then NLP was a big thing where I could read EHR records.

[00:04:11] And again, I'm dating myself when I say it.

[00:04:13] Now no one talks about NLP.

[00:04:15] It's all about Gen AI, Generative AI, LLMs, large language models.

[00:04:18] That was the role that I was playing at Highmark Health.

[00:04:22] And that sort of like has brought me to pharma, but I made a switch during COVID.

[00:04:27] And right now I lead the technology strategy and innovations in the R&D space, research and development space at a pharma company.

[00:04:36] So give me a little contrast.

[00:04:38] It's very unusual to get someone that's got both the insurance background and the company background.

[00:04:45] And from my experience inside hospitals, so it was a vertically integrated systems is they tend to not to have the data that we have, say, in the pharma and the medical device side.

[00:04:55] We can we know down to the half penny what's going on in our organizations.

[00:05:46] And that's the concept of data silos.

[00:05:59] If you want technology to impact healthcare, you need access to data.

[00:06:04] So it's like a catch-22 problem that I've seen it go across these different domains.

[00:06:09] And maybe that's why you had your big tech companies and you've got Amazon trying to do something with one medical.

[00:06:15] You had the JP Morgan Alliance, Berkshire, Amazon trying to disrupt and that sort of started and then stopped.

[00:06:24] I think the problem with data is something that everyone is getting to see firsthand.

[00:06:30] And that's when you start to backpedal and be like, OK, how do I deal with it?

[00:06:34] So you found your way after Highmark into so you did the Amgen twice, right?

[00:06:38] You did it early before Highmark where you were more operations.

[00:06:42] And now you're into technology strategy and innovation.

[00:06:46] And without getting into Amgen stuff, how has your view changed and what should just give us an opinion of what you're seeing in the landscape and things that are coming?

[00:06:57] Yeah.

[00:06:57] So on the landscape, what I've looked at is everyone now has access to the cutting edge technology.

[00:07:04] So you've got LLMs or large language models come out of OpenAI, Google, Meta, Anthropic.

[00:07:10] The landscape is out there.

[00:07:12] So the differentiating factor is going to come from the underlying data, which is going to be unique to each and every form.

[00:07:19] So as an engineer, you love math, you love equations.

[00:07:22] So this is something that I've created for myself, which is data plus AI equals solutions to business problem.

[00:07:30] So when you look at the solution to business problem part, that should always be the prime focus.

[00:07:37] Why are we doing this?

[00:07:38] There's got to be a reason for us to spend money in doing anything.

[00:07:43] The AI part, that's where the tech firms are innovating.

[00:07:46] They're giving you the GPD 3, 3.5, 4, 4.0.

[00:07:51] So they are innovating.

[00:07:52] They're pushing that boundary along.

[00:07:54] So data is the part that's under the control of every firm.

[00:07:58] So how do you format it for optimal results?

[00:08:02] Because you can't just throw any data at an AI and be like, okay, now you're going to solve all of my corporate problems.

[00:08:08] It has to be formatted in a very specific way.

[00:08:12] And that we can delve into how it impacts the pharma landscape at large.

[00:08:18] And just for the listeners to understand the landscape.

[00:08:23] So you've got the healthcare system in the US is a group of peas, I call it.

[00:08:27] So you've got patient in the center.

[00:08:29] You've got peer insurance companies.

[00:08:32] You've got providers.

[00:08:33] This is hospitals.

[00:08:35] And providers have physicians, doctors.

[00:08:37] You have pharma who manufacture drugs, find the drugs, commercialize them.

[00:08:42] You have pharmacy benefit managers, BBMs, who are figuring out the value and the dollar value of those drugs, managing the payments.

[00:08:50] And then you have a pharmacy where you dispense the drugs if you need them as a patient.

[00:08:55] So in this landscape of peas, let's dive specifically into pharma.

[00:08:59] So pharma, you could think of it having as three major pillars.

[00:09:04] So one of it is research and development.

[00:09:06] That's where they find the drugs and get it all the way through the FDA approval process.

[00:09:12] So where you can market it.

[00:09:14] You've got the manufacturing side of it where you look at large scale commercial manufacturing once the drug has been approved.

[00:09:21] And then you've got the commercialization aspect, which takes care of the marketing, the sales, dealing with BBMs and all of that.

[00:09:27] So if you double click into the R&D, the research and dev space, you've got things like target identification and drug design.

[00:09:36] And this is the space where a lot of innovation is happening now with the generative AI front.

[00:09:41] The reason why that's happening is because all of the low hanging fruit, the drug molecules that had to be found, have been found.

[00:09:51] Now it's the tough stuff.

[00:09:52] Now it's the stuff like at the top of the tree, think of a metaphorical tree and you're trying to climb at the highest branch trying to figure out, hey, what drug molecule I can find.

[00:10:02] And think of one of the good examples, you've got AlphaFold by Google, which can predict all sorts of protein molecules.

[00:10:08] And think of this example, like a cancer tumor cell.

[00:10:12] Could you create a molecule that could dock on one end with this cancer cell?

[00:10:17] Because it's got protein expressions on top of it.

[00:10:20] On the other end of this molecule, could you dock with the T cell, which is the antibodies in your body, and then use that antibody to fight cancer by itself?

[00:10:30] So these are tough problems to solve.

[00:10:33] And this is one of the areas where generative AI is coming in and trying to help it out.

[00:10:39] You said something on AI that might be a good pause here that I've heard anyone say it this way.

[00:10:45] Is AI can be a tool that is a product, right?

[00:10:50] Like I create a new product or service with improved features.

[00:10:52] It can be a process, right?

[00:10:55] So it improves production delivery or discovery.

[00:10:57] And then it can change a business model.

[00:10:59] Like we can take out a piece of the model or we can do something different.

[00:11:04] And I think that those were three fine lines that sort of are very gray, right?

[00:11:08] Because you can sit there and be working on a certain project and realize that protein interaction that you were just describing, albeit it had no applicability to project A.

[00:11:20] It's the process discovery of that interaction that you might use later on project B, right?

[00:11:27] So this is where these information architectures that are talking about are mind boggling.

[00:11:33] And you can see that simple example, how it makes sense.

[00:11:36] But then when you try to organize it as an organization, it's just absolutely daunting.

[00:11:41] It is.

[00:11:42] And as I said, the way the complexity has increased, we are trying to look at when you're looking at designing the drug.

[00:11:49] And this is where we are so much closer to, we've always talked about, oh yeah, one day we're going to have these drugs that are uniquely created for a person, uniquely created for you.

[00:11:59] So we are moving in that direction, but it's super complex.

[00:12:02] If you could imagine, you've got your electronic health record.

[00:12:05] You've got your biomarker data.

[00:12:08] So it could be a genetic data, things of that sort.

[00:12:10] You've got lab results that come into it, specimens.

[00:12:13] You've got imaging, MRIs, CT scans, x-rays.

[00:12:17] Now, how do you take all of this data, feed it into a model to give an accurate output for you?

[00:12:25] It is a very complex problem to solve, but I think we're moving in that direction.

[00:12:30] And can I ask just for the audience that, so when you figure that model out and you're trying to put it out into the healthcare system as a process, would we call that precision medicine model?

[00:12:39] It is in the direction of precision medicine.

[00:12:42] And one of the really good examples that you'll see come out right now, so there are a few drugs that are commercialized across the famous names that we have out there in pharma, is they attack really specific gene mutations for cancer.

[00:12:57] So think KRAS G12, you've got C, you've got AB, so you've got a lot of these specific mutations out there.

[00:13:05] And once you create these really specific drugs, the overall outcome that has been seen in peer-reviewed papers that are out there in the public domain, the overall outcome is much better and there's a much better response to treatment because this was really targeted towards a specific molecule and all essence.

[00:13:24] It's really targeted.

[00:13:25] So it is, I think, one of, as we call it, the future is here. We've always talked about it.

[00:13:30] So this is one of the great things of using the tech and trying to get us there to find these molecules because they're really tough to find.

[00:13:40] Now, in other areas where you could look at is rare disease.

[00:13:44] So I think as you've rightly called out, we've got these tools to help us generate AI.

[00:13:48] Could we point them at rare disease where no treatment exists?

[00:13:53] And that's also because the data, underlying data, is also very low and hence the term rare, right?

[00:13:59] If it was a lot of data over there, it would not be that rare.

[00:14:02] So there is a huge impact to people's lives and that is real.

[00:14:07] Anytime I see any drug that has like a pediatric outcome, like it pulls on your heartstrings, right?

[00:14:12] Because these are kids who need treatment.

[00:14:15] This is exactly where we are.

[00:14:16] You could use that tech to fasten and hasten finding these new drugs.

[00:14:20] And then along the process, as you call out, it could help in manufacturing, commercialization, and along the way so that while you move from drug discovery to market,

[00:14:29] that journey can become way faster than what it is right now.

[00:14:32] But you struck a thought with me as you were talking is that as we're mining these data sets for what we call the bigger projects,

[00:14:41] we may discover a solution to a rare disease that we can now commercialize because our discovery came focusing on a multi-billion dollar market.

[00:14:54] Whereas if you started out saying we're going to go after this rare disease, you could never justify the development costs.

[00:15:00] It's not, it's a loop, right?

[00:15:02] And then I was talking to a precision medicine person last week that said our model has ABCD, whatever that model would be.

[00:15:08] And after we've run it for a while, we learned that there's a piece, there's something going on between A and B, that now we can give that data back to Pharma.

[00:15:18] And now they can understand the market size and the biomarkers needed to prove whether you, so this is very interesting.

[00:15:27] So it's all, as you describe it here, it's morphing together.

[00:15:30] And a lot of that is coming from data that Pharma has that the rest of the industry may not have.

[00:15:36] Yeah.

[00:15:37] And I think, I just also want to call out like a lot of these times the data is actually sitting at the hospital in terms of EHR records.

[00:15:44] And then we also, and rightly so, we have these firewalls where what data you can access and can't access.

[00:15:50] So that's right to get to that model where you're talking about, hey, this specific cell can help with this disease.

[00:15:57] It is still at that point in silico, so it's only a theory at that point.

[00:16:02] And to prove it out, you do need the conversation with these hospitals or perhaps there are these middlemen who have the data who could help you out and then prove out that market size as you're calling out.

[00:16:14] Hey, there are these people with this specific disease in this area and then you find out if there's a market, there isn't a market for that.

[00:16:22] So it is a really complex problem there.

[00:16:25] And this could be one of those things where we try to figure out how do we make sense of it?

[00:16:30] How do we remove the complexity?

[00:16:32] If we should remove the complexities and other things.

[00:16:34] There's a lot of public policy that automatically overlaps every time we're talking about the healthcare industry at large.

[00:16:41] Wow. As I heard that, and this is a more general question because I know you focus on technology and drug development at the same time.

[00:16:48] When you describe the model you just described, there is an opportunity for integrated delivery networks and large hospital systems with data to actually share in the development benefits of the pharmaceutical industry to get more profits into their side of the business.

[00:17:06] I've never thought of that. That's way out there. But do you see that as a potential in the future?

[00:17:11] If you're talking about the finance and the numbers, the way you've described it, yes, that's there.

[00:17:16] And I think that could be a potential item in the future.

[00:17:19] I think it's the way the systems are set up. So every time you go to a hospital, no one is thinking about the profits.

[00:17:25] You're always thinking about what's the best outcome for patients.

[00:17:29] And I think that's what a lot of the folks that I work with, it's always about, hey, how do we help patients on a day to day basis?

[00:17:36] And I think that's the key that I would like to join and work on is they have that critical data.

[00:17:43] Now we can look at privacy, we can look at HIPAA, how do we get the data authorization, help out a farmer with that data set, and then create that virtual cycle that you're talking about.

[00:17:53] Yeah, that's where the core and key data is, and how do you make it work together.

[00:17:59] And so I don't look like an accountant when I say, I'll just say that when I was with Johnson & Johnson, they did some large, expensive clinical trials,

[00:18:09] reimbursement ahead of launching some new expensive technologies.

[00:18:12] And it wasn't something that they necessarily needed to do.

[00:18:17] But what they recognized that they didn't do it and get that code proactively, that they could unintentionally financially harm a lot of the small regional hospitals in the process when they're faced with doing the right thing for the patient.

[00:18:31] So my statement on that was that our hospitals are operating at sub 4% profit margins, and they're starting to get rid of product and service.

[00:18:46] And so I think that's the best benefit that they can bring to the community.

[00:18:51] So I can envision an opportunity here to throw some cash in that way so these systems can deliver to the patient, which is a lot of, most of them are nonprofit is that's their mission.

[00:19:02] I think that I love it.

[00:19:03] That is so true because, and this is where everyone along the system can make use of this new technology of Gen AI,

[00:19:10] because you have access to this great AI.

[00:19:15] Now, whoever has the data, you could attach it to it and see what business problems that we can solve.

[00:19:21] Speaking of that, in the hospital space, with the generative AI, there is an out-of-the-box feature that comes with it,

[00:19:29] which can help summarize papers, patents, clinical trial data, deep research data.

[00:19:35] So folks in hospitals, like if they are spending 100 hours a month looking at the new articles,

[00:19:42] new treatments in their area, and doctors rightly do, it can help them summarize things.

[00:19:46] So out of that 100, if I can make it that doctors are now spending 80 hours a month, out of the back, 20% efficiency.

[00:19:53] Boom. Right there.

[00:19:55] And this summarization goes across the space.

[00:19:57] So you could use it in an insurance company.

[00:20:00] You could use it in pharma.

[00:20:01] Same researchers looking at drug molecules, same research, different papers, different focus.

[00:20:07] All of a sudden, if I can give everyone like 20% productive efficiency, can you think about it?

[00:20:12] Like the US healthcare market is sitting at what, $4.5 trillion.

[00:20:18] And if I say, hey, if I can make it 20% more efficient, like that's hundreds of billions of dollars of efficiency right there.

[00:20:25] And I think that's your point when we say more efficient in other industries, people think of losing jobs, but this is an industry where we can't find enough people.

[00:20:32] Yeah, this is not about jobs.

[00:20:34] It's more, hey, how can I get the latest drug, the best drug out there faster?

[00:20:39] How do I treat a patient so that their outcomes are much better?

[00:20:43] They have things like early detection and prevention.

[00:20:45] How do we get there?

[00:20:46] How do we get to an improved diagnostic every time a patient has to go through it?

[00:20:52] How do we ensure patient safety?

[00:20:54] Do you've got drug adverse events?

[00:20:55] Can you preempt them?

[00:20:57] So there are, these are the ways I'm talking about the efficiency, like using a computer to say, hey,

[00:21:02] person A with this profile, when they get this drug, just be careful of these adverse events.

[00:21:07] You're preempting it so the hospital is aware, the entire system, the care team is aware that this can happen.

[00:21:13] So they're not caught off guard and they're not prepared for it.

[00:21:16] So you're already efficient right off the back, as opposed to on the back foot, something happens and then you're coming to the ER.

[00:21:22] That's the efficiency that I'm talking about.

[00:21:24] Yeah. So what are the challenges in getting to this future that you talk about?

[00:21:30] Oh, that is a brilliant question.

[00:21:32] And if I were to answer it, one of it is data.

[00:21:35] I think you've seen the theme, like I work on data strategy a whole bunch and that silo-ness of data and then the ability to access data and work on with a global team.

[00:21:46] And then each country, EU has its own set and regulations, Japan, China, Asia Pacific, India, Singapore.

[00:21:53] So every country has its own requirements on how you can handle data.

[00:21:59] If you can get the data, how the privacy compliance must be handled.

[00:22:04] So that there is complexity there.

[00:22:06] The second and the big challenge in this space is clinical trials.

[00:22:11] And to make it really safe for all of us, and maybe I'll just explain clinical trials for our listeners out here.

[00:22:17] So when you get a drug, you get them through phase one, two and three.

[00:22:22] And that's when you launch in the market.

[00:22:24] So phase one is all about toxicity.

[00:22:26] If you put it in human, yeah, it's not toxic enough.

[00:22:29] They can handle it.

[00:22:31] Moves to phase two, a larger population.

[00:22:34] You introduce a placebo.

[00:22:36] You maybe look at different dosages to say, okay, 100 milligrams is not enough.

[00:22:42] Maybe 140 or 200 is the optimal dosage.

[00:22:45] So you figure out your dosage there in phase two.

[00:22:48] Phase three is the larger scale study where you prove out to the FDA in the US that yes,

[00:22:53] this drug is better than what's out there or nothing exists in the space.

[00:22:58] So now I'm treating something that's new and unique.

[00:23:00] And that's when you get the FDA approval.

[00:23:02] And then you have a phase four, which is the post-marketing to continue to prove to the FDA,

[00:23:07] hey, this drug is working as intended and as proven.

[00:23:11] And the adverse events can get not looked at better.

[00:23:14] So getting patients identified for these clinical trials is a big gap.

[00:23:19] So if you look at any study out there for pharma, the biggest gap comes from that patient identification.

[00:23:25] You're not able to get enough patients for clinical trials,

[00:23:29] especially as drugs become more and more specific.

[00:23:32] You're looking at someone with a very specific profile that it can treat.

[00:23:36] So there are now these companies out there who are also trying to innovate,

[00:23:40] and that's why you've got the whole ecosystem that works together.

[00:23:43] So you look at patients' EHR record, and EHR is electronic health record.

[00:23:48] Again, with access, with appropriate permissions, you can look at an EHR record and then compare it to the inclusion and exclusion criteria of a clinical trial and match them.

[00:23:58] And that, I think, is an area where there's a big gap and hopefully we are able to solve it.

[00:24:05] And this is why we would need a neutral third party because you're right, as you pointed out in a couple of examples earlier about profits and things of that.

[00:24:14] You don't want folks to think, okay, this is happening in a pharma, so there's a profit motive, or if it's being done in a hospital, there's a profit motive.

[00:24:22] Maybe a neutral third party says, hey, I take your data and then I try to get you the best outcome, and that's what everyone wants.

[00:24:28] So getting that patient identification piece, if we can get it better, then you can have more clinical trials, more drugs coming out there which are having good efficacy, good outcomes for patients,

[00:24:41] and I think it's a really good virtuous cycle that we can create.

[00:24:45] I'm looking for some data here from a project I did in 2015.

[00:24:51] It was a capstone at CMU, and I'm not going to find it, but the crux of it, I just want to validate something you said.

[00:25:00] It was something like a third of all clinical trials, meaning pharmaceutical companies have said, I want to fund this trial.

[00:25:08] I've got the money to fund this trial.

[00:25:11] Didn't get complete because they couldn't get the recruitment and they ended up aborting it, and I think that's an excellent point that you make.

[00:25:19] I'm not going to find it, but it had nothing to do with the drug failing and everything to do with just not being able to find the patients.

[00:25:27] Yeah, and rightly, FDA does require a specific number, but these clinical trials, and now the FDA is also talking about things like diversity,

[00:25:38] making sure that the patient population that you're using is diverse to represent what is US.

[00:25:45] So that adds another layer of complexity to it.

[00:25:49] So you're going to have a lot of discussions on it, what's an appropriate size.

[00:25:53] The other thing where we talk about generative AI and we talk about FDA, they would also need to move lockstep with the industry.

[00:26:02] So can I use Gen AI to predict a molecule?

[00:26:06] Is that allowed?

[00:26:07] Things like that.

[00:26:07] So you need the government and the rules and regulations to move along in lockstep to make sure that the industry,

[00:26:14] like if I'm using Gen AI to match patients with clinical trials, is that allowed?

[00:26:19] Do I need an FDA approval?

[00:26:21] Will it be considered by the FDA as a SAMD or software as a medical device, which needs an approval?

[00:26:28] So you can't just willingly throw something out there.

[00:26:30] So there is policy that will always flow in whenever we are talking about latest technology being introduced.

[00:26:39] And that's another space policy that there needs to be taught.

[00:26:42] And a lot of leaders in the space need to come together with government officials to make sure that the ecosystem that we have is protected and enhanced.

[00:26:52] And I think the policy people always say that technology will always be faster than policy and law that supports it.

[00:27:00] So I think of self-driving cars many years ago when they were first debating, they had this multi-year pause because someone got hit by a car.

[00:27:08] Being from Pittsburgh, you know exactly the story.

[00:27:11] And it was who's responsible.

[00:27:13] And it took them years to sort through it.

[00:27:15] And at the end of the day, it appeared that it was the person that was responsible.

[00:27:19] But the question is, so who pays for that?

[00:27:22] Is it the insurance company?

[00:27:24] How does this get sorted?

[00:27:26] And so these are multiple challenges that we all face on a daily basis.

[00:27:32] Now, you've given us a story of a lot of change here, but I'd love to go a little deeper on.

[00:27:37] Tell us a little bit of a time where you had to adapt or shift strategy quickly as you've made all these changes in your business.

[00:27:45] And how did you do it?

[00:27:46] And how did you adapt?

[00:27:48] And how did you sometimes during rapid change, how do we keep our mental well-being?

[00:27:52] Oh, that is a very good question.

[00:27:54] There's a perfect example for this.

[00:27:56] So think year 2022.

[00:27:59] Let's go back in time machine.

[00:28:00] So this was December.

[00:28:02] We've created our tech strategy for the year.

[00:28:04] We did the leadership presentation, hi-fives all around, celebratory dinner.

[00:28:09] All of that is done.

[00:28:10] I come back from California to Pittsburgh.

[00:28:12] And in pharma, some pharma companies in the US, they have a concept of winter shutdown.

[00:28:17] So you get a week and a half, the week of Christmas, New Year, shut down, everyone's off.

[00:28:21] So we do that.

[00:28:23] And then OpenAI releases and becomes famous with their chat GPT.

[00:28:27] And boom, come January, we had to pivot.

[00:28:31] We had to incorporate LLMs, understand how do we prioritize this technology?

[00:28:36] Because now we know the use case.

[00:28:38] It can be used in R&D, commercial, manufacturing.

[00:28:40] Where do we go?

[00:28:41] And within each of these segments, how do we prioritize?

[00:28:44] Which one will give most value?

[00:28:47] How do we test it out?

[00:28:48] So you want short tests with low dollar value.

[00:28:52] Because if you spend millions of dollars every time, you're going to be a highly inefficient company.

[00:28:57] So you want short tests and then prove it out and then scale it.

[00:29:03] So with a low value test, you can get a value outcome.

[00:29:06] Oh yeah, this is going to give me a 1 million, 2 million, whatever dollar value outcome.

[00:29:10] So create a prioritization matrix.

[00:29:12] Okay, everything is going to get ranked sat this way.

[00:29:15] These are the values you're going to use.

[00:29:16] It was one of those times where you read a lot, read industry-specific news publications, go to tech conferences.

[00:29:24] Like I learned a lot.

[00:29:25] And all these tech conferences are out there, public.

[00:29:27] Microsoft has one.

[00:29:28] Google has one.

[00:29:29] Amazon has one.

[00:29:30] OpenAI, NVIDIA, all of these big guys, they have their tech conferences.

[00:29:35] So you go there, you learn a whole bunch about how they're using this technology in the healthcare industry

[00:29:39] with specific streams on life sciences, hospitals, insurance.

[00:29:44] They'll cover it all.

[00:29:45] So I would say, read a whole bunch, read a lot.

[00:29:50] And then you have to be really adaptable.

[00:29:52] Like you get thrown with new data every day.

[00:29:55] The innovation is really fast.

[00:29:57] And you can't just say, oh yeah, I'm happy with the status quo.

[00:30:00] It is changing and you've got to change with it.

[00:30:02] So that sort of brings me to the point of, hey, how do you deal with all of this change?

[00:30:07] The mental health aspect is really true.

[00:30:09] One of the things that I try to limit myself is like weekends are for myself.

[00:30:13] So I try not to look at my work during the weekend.

[00:30:17] I shut off.

[00:30:18] Luckily, we have the summer shutdown, winter shutdown in my company.

[00:30:21] So complete shut off.

[00:30:22] Travel with family.

[00:30:23] Completely cut off from work.

[00:30:25] And then exercise.

[00:30:27] So I make it a point to go to the gym at least three to four times a week.

[00:30:32] I have a toddler and a newborn.

[00:30:36] And then with that, there's all these complications.

[00:30:38] And hey, time flies.

[00:30:39] And all of a sudden, oh yeah, this week I didn't go to gym even once.

[00:30:42] But yeah, just spending time with family is a brilliant way.

[00:30:46] Especially babies.

[00:30:47] You lift them.

[00:30:48] They smile at you.

[00:30:49] You forget about work and all of the other things.

[00:30:51] So yeah.

[00:30:52] Good for you.

[00:30:53] So what resources do you follow to keep current?

[00:30:57] Listeners always love to hear new sources of information.

[00:31:01] Where do you personally go?

[00:31:03] Yeah.

[00:31:03] As a B school, we have to have Wall Street Journal.

[00:31:07] You get it along with your degree.

[00:31:09] So that's done.

[00:31:10] The tech conference is us calling out.

[00:31:12] All of these companies have newsletters and information that flows out periodically.

[00:31:18] So just get on that subscriber list and you will get it.

[00:31:22] I also attend a whole bunch of research and development lectures that I get access

[00:31:26] through my company.

[00:31:27] And these are scientists at various universities.

[00:31:30] And I listen to them, how they are trying to use technology.

[00:31:35] And that's where you could hear where the gaps are technology-wise.

[00:31:39] And then you could be like, okay, how would I solve this gap?

[00:31:41] How would I use what we have today access to in terms of technology to solve this space?

[00:31:47] And then again, ideate, run a small pilot or a use case and then expand.

[00:31:52] So don't go for the entire pizza.

[00:31:54] Eat it in bite-sized chunks so that you can have the whole thing.

[00:31:59] So as you reflect upon your experience, what's the biggest lesson you've learned thus far?

[00:32:07] I think one of the biggest lessons that I've learned is in healthcare industry, and this is where I really invite all listeners to work in, like the impact is real.

[00:32:17] You're really changing someone's life.

[00:32:19] You're impacting them.

[00:32:21] In case, let's say you take a case of a rare disease where there's no treatment.

[00:32:24] And even if you had a small part to play in making sure that the data strategy was in place so that the treatment could go on, and now it's in the market.

[00:32:34] Think of the folks that you've impacted who had no treatment, and now all of a sudden have a treatment can have a better quality of life in all the sense.

[00:32:41] One of the lessons is work in an area that you find valuable, that brings you happiness, and I think healthcare is one of those places where you can do that and have a good impact on everyone.

[00:32:58] And what do you see as the biggest opportunity or threat in the next five years to healthcare in general?

[00:33:04] On the opportunity side, I think it's going to be how I look at it is in terms of efficiency.

[00:33:10] So we talked about the U.S. healthcare market.

[00:33:14] You're sitting at $4.5 trillion, $400 billion specifically on the drug-related costs.

[00:33:20] Can we gain efficiency by automation?

[00:33:23] So you have this AI.

[00:33:24] Could you automate document processing, document generation, like the stuff that should be automated, none of the critical stuff.

[00:33:31] And what you do is instead of creating a document, you review a document.

[00:33:36] You're more of a reviewer as a human as opposed to the creator, and then you reduce your time.

[00:33:40] So if I start to gain efficiency at, let's say, 3%, not that high, you automatically have already offset the inflation rate.

[00:33:49] So your spend as a country will not go up next year because your efficiency has matched the inflation rate and you balance yourself out.

[00:33:57] That, I feel, is the opportunity here that we have enough and more ways where we could start to gain efficiency by using the latest technology.

[00:34:06] And in terms of threat, I would say in the healthcare side, the threat would be lack of policy or lack of unified policy would be one of the biggest threats.

[00:34:16] Because if you create an ecosystem where it's from the patent side all the way through, can I use it?

[00:34:21] What sort of AI can I use in the drug, the clinical trial process?

[00:34:26] Where can I use it?

[00:34:27] In what format?

[00:34:28] If it's clear, if there is guidance on it, I think it'll help the entire industry.

[00:34:32] So I think that's a threat which could go anywhere depending on how a court case is settled.

[00:34:37] And so there are some, as I call it, waves that could be made from any new information that's out there.

[00:34:42] Great. What else would you like to share with the audience?

[00:34:45] I think one of the things that I want to call out is work in the healthcare industry.

[00:34:49] There's a lot of value there in terms of impact to people's lives.

[00:34:54] And this is a lot of good, great stuff that can still happen, still needs to happen.

[00:34:58] I'm all about, hey, work in the healthcare industry.

[00:35:01] And the other sort of disclaimer that I want to put out for the listeners is like all of the thoughts and opinions are mine and not of the companies that I'm working for.

[00:35:09] And this is not in any sort of an investment advice.

[00:35:14] So this is just a conversation with Jim.

[00:35:16] Very good.

[00:35:17] Thank you very much for being a guest here.

[00:35:19] It's been very enlightening.

[00:35:20] Thanks.

[00:35:23] Thanks for tuning into the Chalk Talk Jim podcast.

[00:35:26] For resources, show notes, and ways to get in touch, visit us at chalktalkgym.com.