Host: Health care accounts for nearly 20% of the American economy and while Silicon Valley promises will make medicine cheaper and faster. The reality on the ground threatens to deliver the exact opposite. Now just this morning, pharma giant AstraZeneca reaching a deal worth as much as $1.86 billion with machine learning biotech Insilico. Now AstraZeneca tapping the company for AI-driven drug discovery for difficult-to-treat diseases in the short term. Tools that spot health risks early are actually driving spending higher, but as VC shifts from experimental pilots to full operational autonomy, can algorithms truly make health care deflationary or will legacy billing incentives simply absorb the productivity gains? Well joining us live to break down the economics of automated medicine is Amo Madon, founder and CEO of Brady and Graph. Great to have you back. Thank you so much for joining me.
Amo Madon: Thanks for having me.
Host: Well, we all know that we're paying attention to artificial intelligence, the capacity of what it can do, as well as the investment themes and capex spending. But when it comes to healthcare, what's actually happening right now and the risks and the opportunities?
Amo Madon: We're still very early in what AI is going to do to transform healthcare, and I would say we're at the very start where most of the investments have fallen in two categories. One is investments that are focused on operational tasks inside the payer landscape or the provider, the health system landscape, and that includes things like the front office, a lot of billing, administrative-related tasks, and so on. And the other is what I say is experimental AI spend, which inside larger organizations is, hey, let's arm everybody with the right AI tools and kind of see what happens. So I said those are really the two things we're seeing and we can talk about why those are kind of driving up costs in the short term, but also what we need to change for the costs to come down long term, which I think is a really exciting discussion.
Host: So let's expand on that, Emil, because we're all paying attention to what AI can do, especially when we're talking about healthcare and how it can speed up progress when we're thinking about drugs as well as diseases. So what are you bullish on and what are you not so bullish on?
Amo Madon: Yes, look, I'm very bullish on the 10-year horizon for sure, right? And I think just to use one of those areas you laid out, Remi, is care delivery. And there's a number of places in Medicare, in commercial employer-sponsored health insurance, where people living with a chronic condition, living with, say, back pain or diabetes or heart disease, might only see their doctor or specialist once every 6 months, once a year, and we know that these AI systems can play an incredible role in providing people support in between those visits, guiding their care, and ultimately translating that to better management of the condition and reducing the total cost of care. So those are the kinds of opportunities that are, I'll say, the mid-stage of the journey, which are the ones that I'm excited for us to kind of see in the market, right? And that's, and then there's a long, long stage where machines take on some of what we call the increasingly autonomous tasks in clinical decision-making, diagnosis, treatment, and so on, and I think that's a bit of a longer-term play and we can talk about what we need for that as well.
Host: So I do want to ask you very quickly about the regulatory guardrails, whether regulation is good or not so great for this. But there are a lot of moving parts and of course we continue to look at inflation. So how does AI and healthcare actually affect inflation?
Amo Madon: Yes, I think we need to create swim lanes where companies can move fast and can bring these ideas to life, but do it safely, right? And I think right now it's a bit of a broken landscape. And so the risk is we end up where we ended up with telemedicine, where we had technology, but lots of different kind of benchmarks across states, across regions, and it took a long time until COVID happened in 2021 for there to be general broad guidelines at the national level. So I think we have to figure out, and Utah is on one end of the spectrum, there are other states on the other spectrum that are either encouraging or discouraging kind of the autonomous opportunities here, but how do we create a safe swim lane. And you know, again, these AI systems are kind of like toddlers learning how to walk, so we do need to understand that and respect that, right, but also create ways for companies can learn and iterate fast. Create open benchmarks, in the AI world we call them evals. How do we create kind of benchmarks and evals, not just data set that was collected for training from human providers, but also actually designed for use cases. So those are all things we have to get dialed in if we're going to move faster here.
Host: And less than 60 seconds here, but we are counting down to year-end and when we're thinking about trends that we're focused on for 2027, what are you paying attention to in this space?
Amo Madon: Yes, I'm really paying attention to what's happening at the regulatory, but also at the startups, the tech companies, innovators of how we see the emergence of what I'm calling hybrid care. So AI and doctors working together in sort of daily care delivery settings where AI takes some of that, takes some of the burden for the lowest acuity, the simplest, kind of the most basic of clinical care. I think that would be really interesting to watch as a consumer of healthcare.
Host: Well, Almo, it was always great talking to you. Thank you so much for joining us and thank you so much for all of your insights.
Amo Madon: Thank you. Thank you.
