00:00Co-founder and CEO Ankit Jain says healthcare AI today is a world of co-pilots. The next few years
00:06will be about agents. His company is using AI agents to automate everything from insurance approvals to patient follow-ups
00:13and he joins us here on set. Ankit, thanks so much for your time. It's an industry that is really
00:21ripe for change because it's been stuck in the past seemingly for so long. What can you improve in terms
00:28of efficiencies in healthcare?
00:30Well, first, thanks for having me. You know, healthcare is one of those industries that affects every single one of
00:35us and we all want it to get better. We want it to get better for ourselves, for our loved
00:39ones. And it's terrible that when we're going through these healthcare journeys, we have to advocate for ourselves. We have
00:47to connect the dots between our doctors and our insurance companies and the support programs. And the opportunity for technology
00:53here is to connect those dots so that we can focus on healing ourselves.
00:58There's this argument, and you see some of this in Capitol Hill with lawmakers trying to push back on allowing
01:03AI to be used in things like prior authorization. That you essentially just arm insurers with a tool to more
01:09quickly deny claims. And then maybe you have providers who are also using AI to try to fight the insurers.
01:14Does this just turn into an AI arms race between the two different parties at the end of the day?
01:56I mean, I mean, I mean, but I mean, to Danny's point, these insurance companies are trying their hardest not
02:00to pay out, right? That's why the American public is so dissatisfied with health insurance.
02:07I had a motorcycle accident years ago, and it took me multiple years to resolve the insurance issues after a
02:14lengthy hospital stay. That's not because I was having trouble communicating with the insurance companies, because they were fighting as
02:22hard as they could not to pay for my treatment.
02:25You know, the reality is there's many patients that are going through the exact same journey you went through, and
02:30it's terrible that they have to go through this.
02:32On the other side, there's a set of folks that are overbilling, and AI is playing a big role in
02:39this as well. On the provider side, the companies that are giving the scribes that are enabling providers to not
02:45have to document something are now enabling upcoding and making sure that the hospitals and the doctors can get as
02:53much out of those insurance companies as well.
02:55So what we need to do is come together and make sure there's a good balance so that our insurance
02:59premiums don't go up because there's overpayment in the system as well.
03:02What about costs? Because the common intuition is that you can trim costs.
03:07Our producer, though, Lee, dug up this study from PwC basically showing that costs were expected to spike by 9
03:14% by 2027, and 70% of those surveyed said that it's use of AI documentation and billing tools as
03:22one of the top three drivers of costs.
03:25What about AI is adding to costs and not doing the opposite? It's just because compute is expensive right now?
03:30You know, and compute is getting increasingly expensive. As we think about AI models getting better, we're paying today not
03:38for outcomes, we're paying for machines to think. And we need to make that transition in the world of AI.
03:43Now, healthcare, surprisingly, has been thinking about this for a couple of decades.
03:47We've been thinking about paying for outcomes. There was this concept called relative value units where you pay for the
03:54complexity of work, you pay for the outcomes rather than just for activity. And the world of AI needs to
03:59transition from paying for thinking to paying for outcomes.
04:03Tell us about EVA. So EVA is one of our AI agents that is able to help patients through these
04:09journeys, whether it is checking the status of a prior auth, explaining to a doctor's office what needs to happen
04:15to get that prior auth approved, or to help a patient understand what to expect at their next infusion, or
04:22to check in on the patient to see how they're doing. It's our set of AI agents that help connect
04:27those dots between the different groups.
04:28But EVA is someone that I would deal with as a patient, someone that the doctor would deal with as,
04:35you know, the healthcare professional, or someone that the insurance company would deal with?
04:39All of the above. We've got EVA facing patients and helping the patient understand what to expect, or checking in
04:46on the patient, understanding their medication regimen.
04:49On the other side, EVA is helping doctors' offices get those prior authorizations approved.
04:55And finally, EVA is helping insurance companies communicate more effectively with the providers as well.
05:01For these large frontier models, it feels like all of a sudden, it doesn't feel like it has been all
05:06of a sudden that politics has been injected into them.
05:08With the White House telling Anthropik, your most advanced models can't be used abroad, so Anthropik says, okay, we're not
05:13going to offer them to anyone.
05:14As a business leader, how do you think about using these models if there is a risk that the switch
05:20can be turned off by the government now?
05:21You know, like everything we've seen in technology, AI is now becoming a key part of the supply chain, so
05:27you have to look at it from a supply chain risk perspective.
05:29Can you switch from model A to model B?
05:32How much control do you have, and how much control do your vendors have, downstream vendors have, and make sure
05:37you have that ability to switch models at the drop of a pin?
05:41Because some of these things are not in your control.
05:43We have, you have 98% of the payers that you work with, right?
05:50You have a massive chunk of the Fortune 500 companies that you're already working with.
05:57Does Infinitus become a de facto infrastructure layer in the sort of national healthcare model now?
06:04You know, when we first started the company, I thought it was ridiculous that the way data exchanges hands today
06:09in healthcare is over phone calls and faxes, and our hope is over the coming years, we replace many of
06:14those back office phone calls with digital connections with each other, so that data can exchange hands instantaneously, patients can
06:20get on therapy faster, the costs can come down.
06:23And that happens as we become part of that, the fabric of the healthcare system.
Comments