Ram Yalamanchili (00:01.294)

Hey, Dr. Fox, how are you?

James Fox (00:02.742)

Good thanks so much for for giving me the opportunity to have a little chat here

Ram Yalamanchili (00:08.726)

Yeah, I'm excited. So I think to begin, I'd love to hear your story. Tell us more about yourself, your practice, obviously your research program.

James Fox (00:22.388)

Yeah, I trained, initially I trained my residency with the University of Missouri, then I trained with Ike Ahmed up at the University of Toronto. And at that time, really that fellowship gave me a real interest in research and real interest on doing things on the cutting edge. And so that's what attracted me to research and always behind any research site that

that has any chance of being good is a large number of people who are putting forth good work, whether that be research coordinators or even quite frankly, front desk, the technicians and so forth. So I joined ICONiCare out here in Grand Junction, Colorado, approximately nine years ago. And we started a research site, started off with one full-time study coordinator that really wasn't full-time, but that's what we told.

The first sponsor, yes, this person's full time, you know, and I think everyone who's done research that started it from the ground up can relate to that kind of thing. And now we've progressed to doing over 25, maybe 30 studies. I don't know. I don't count them. I didn't count them in preparation for us hanging out today, but we've done a fair number of studies. We have three and a half study coordinators that are full time and that's legit. We actually have that many now. So

So yeah, that's kind of where I've come and where I am right now in the process of our research program.

Ram Yalamanchili (01:57.848)

You know, I know we spoke about this, but why did you get into research? What's like the, I mean, you have a pretty strong, mean, pretty busy clinical practice from what we know and what we've seen, but I'm curious what was like the original thought around why get into research?

James Fox (02:14.73)

Yeah, I think ultimately there's very few doctors that wouldn't feel like we went into this to help patients, to help people. And I think we oftentimes start that way in direct patient care. And then I think over time that evolves to how we think about where we can contribute to our part of improving patient care.

And so I've always loved doing new things. I've always loved taking on new surgeries, taking on challenges in that regard. And I'm always excited. I've always been excited to take on, whether it be a new surgery or a new procedure, a new device or new medication, I've always been excited to try it. and so for, myself to not only have the opportunity to try these things earlier, but be part of the group of individuals who

bring, allow a product to be fully evaluated to make sure that it is an appropriate thing for all of our patients to have it available to them. That was like a major driver. So pretty much just wanting to be on the cutting edge and contributing to patient care in a more than just a one-on-one.

Ram Yalamanchili (03:32.046)

Cool, cool, I see. And from, you know, when you said you had to start with one coordinator, did you have prior experience before that? Were you doing any research in your residency or maybe part of?

James Fox (03:44.214)

Yeah, in my fellowship, research was very emphasized, both on investigator-initiated-like trials, large multi-clinic trials, and also FDA studies and registration studies. It was up in Canada, so there were studies through that as well. So, yes, it was big time.

Ram Yalamanchili (04:11.864)

So I think that's an interesting vantage point. So what you're also telling me is you've been seeing research or at least working within the research framework for quite some time. Outside of that nine years, you've been at ICON, right, through your practice. So maybe from your own words, how would you describe things evolving, if any, which you can share? Like, what was it like when you first got into it?

today, right? And I'm talking pre tilde, not like after we started opting, but prior to that, how's it been in your perspective?

James Fox (04:43.786)

Yeah.

James Fox (04:50.196)

Yeah, I mean, I haven't considered that question before you asking me, but I actually think if we're looking at what we're researching and what we're looking into, there's been massive changes, like incredible improvements. If we want to look at processes that go behind that research, there's actually been very little change. And that necessarily isn't a bad thing. I mean, if it ain't broke, don't fix it.

And I think that research has worked. Research is something that's a discipline in monotony, oftentimes. I mean, there's the very intriguing, exciting part of research, and there's the monotony side of research. And you can't have one without the other. so the behind the scenes stuff, the binders, the regulatory documents, the keeping up with things, there's very little changes, even going back

Now, how long has it been? good 12 years or plus from when I was training in my fellowship. Those changes do I did not seen significant changes in the behind the scenes component of research much.

Ram Yalamanchili (06:03.362)

Got it, okay. And right now, given that you've got a very active research program, what would you say is sort of like, how do you describe your research program from a volume perspective? And what percentage of your active clinical patients are usually in research? Have you looked at such numbers just to see where things are when compared to your clinical practice?

James Fox (06:31.22)

At any given time, we've found ourselves to be in anywhere from eight to 15 studies at any given time. I don't know what the percentage of patients is that's a little hard. have a co-management network and we have a constant filling up of different patients and sending patients back to referring providers. So that'd be a little difficult. I will say that strategy that we've had is to make sure that every

Every main service line that we offer to patients has a research study within it that patients have the option to participate in. And so oftentimes we do fairly well in enrolling. We are one of those practices that's fortunate enough to just have an extreme stream of patients coming our direction.

So the percentage of patients who are actually in research compared to the number that we actually see is relatively small. But the amount that we're actually seeing from a research side, I think is fairly sizable.

Ram Yalamanchili (07:39.5)

I see. That's interesting. what you're also saying is, from a patient's perspective, if they're coming to your clinic or your practice, every service line has some option outside of just standard of care. I'm assuming that's an awesome thing, right? From a patient perspective, it could be.

James Fox (07:57.93)

Yeah, yeah, most certainly it is. I I find that oftentimes patients do want to contribute. I think also there are times where studies are fitting a niche where the only thing a patient kind of has available to them may be a study. So there's a wide range of why a study might be beneficial to patients. And our tact is that if they are within the inclusion exclusion criteria, that it's presented to them. I mean, we are not, we are not a site.

and I don't think there's many of them around, we are not a site where we're trying to GM research down people's throats. And at the same time, for patients to have that opportunity, think shows great value. I think it's valued in our community. Yeah, so it's.

Ram Yalamanchili (08:43.244)

Yeah, totally. That makes sense. So given that amount of extensive experience, I'm very curious to hear about what are the challenges you've seen. Clearly, I think we sort of touched on staffing and such. if you were to just describe all the different challenges which you might have encountered or are encountering,

Knowing what you know, would you tell a younger yourself or somebody else who's trying to start in research.

James Fox (09:14.646)

Yeah, yeah, I think that even people are more experienced than me and have a bigger program than I have would definitely start relating to, I think the number one issue is scalability. so also there's ebbs and flows in research where maybe you might have 10 studies actively recruiting at one moment, and then you have three that are actively recruiting and you're kind of closing up some studies. And so this is the thing where staffing is really critical.

Research is built on trust. so we as surgeons who are not in research, we're trusting that the FDA and clinicians that are involved in research are doing things properly so that then when we use products, we can have confidence in those products. The FDA trusts in a way that they regulate and they monitor as well that the sponsors and the sites and so forth are putting forth good data and running the studies well.

I think ultimately us as principal investigators, we are trusting our staff to run the study appropriately. And there's only so much resources on a fundamental level when you're at the site level, there's only so much resources that one person can do. So I think probably the biggest jump is to go from one full-time coordinator to the second full-time coordinator. But when you go from one to two,

you have one training, the second one. Okay. And when you go from two to three, you're busy enough that that training sometimes gets watered down compared to when the first person got trained to the second spot. And so that scalability of staffing, because I, for those who are in research, they know that your trustworthiness of your study coordinators is probably the most important thing.

along with them being really having a good attention to detail and making sure they're doing things the right way.

James Fox (11:18.134)

Trust is a massive part of this. And so the bigger your circle is of trust, and the more humans that are involved, the more human error can take place. And so I think scalability is the number one challenge I have for research. I'm fortunate enough to have the volume of patients and the interest in the community for research. But

And the ebbs and flows make that scalability necessary. Not just, not just, okay, now we're ready to have a third study coordinator. Well, sometimes as far as the budget's concerned, it would be good if we add two and then it would be good if we had four. And so you hire three and on the whole, that's a good idea, but you know that from a budgetary standpoint, it'd be great if you could furlough one for a few months until you need one again, but that's not how it works. I mean, you've got to retain your staff. And so,

that scalability is, I think, the number one challenge we have in research.

Ram Yalamanchili (12:18.402)

What has your methodology been in terms of hiring, training, just evaluating staffing? How do you build this, your current practice, would say, particularly your research program? Have you noticed some things which work? And have there been challenges which you ran into which clearly did not work, for example?

James Fox (12:42.582)

Yeah, so we've hired people who have ophthalmology experience. We've hired people who don't have ophthalmology experience, but have extensive clinical research backgrounds. I think that it's potentially easier to train people who understand ophthalmology to train people on research than it is to train people who are actually really good at research on how to do ophthalmology. One of our research...

search coordinators right now did fit that mold of they came from research that was outside of ophthalmology and they've joined our our group and it took growing pains and it was great. I mean he did he really has done a great job at developing into somebody who understands ophthalmology well, but I think certainly if you're starting off on research, think my my first my main research coordinator was the lead technician.

of our clinic. And so somebody I already trusted, right, very important. Somebody already trusted and knew they had the skill of meticulous attention to detail. And

Ram Yalamanchili (13:50.21)

So you took somebody who you already trusted, and also has a ton of experience in ophthalmology, and essentially trained that lead coordinator to be, I guess, more aware of how research is performed.

James Fox (13:53.13)

Mm-hmm. Yep.

James Fox (14:03.392)

trained and grew together. know, mean, since we started, did not, we did not take this program over from an experienced investigator. We started from the ground up. And so there were a lot of things we listened to mentors and to people who would help us in the process, most certainly. But at the end of the day, her and I were in a room together oftentimes figuring things out for ourselves. So.

Ram Yalamanchili (14:31.693)

Got it.

James Fox (14:31.816)

So yeah, that's a little bit of a tough place to start with, but also we know the nuts and bolts it took to get here. so I think that allows us to really have a firm grasp on how to take things to next levels, because we know the foundation of the levels that we've created.

Ram Yalamanchili (14:51.692)

Right. And having met your staff and also your lead coordinator, I think one of very interesting thing I've noticed is when I first met you and your staff, you're very open about the challenges you've had. I think there were some active challenges around maybe some staff. There was some staff churn. There were some things which you would have said we could have done more efficiently.

Even with such a trusted and, I guess my point is that even with a really strong program and a staffing model, which you already have, it was very clear that she was very much inundated with work. And there's just a lot of overhead which was being spent on essentially like, tasks which otherwise would have been spent on maybe patient care or new program, new studies, that sort of thing.

Can you give us maybe your version of what that looked like right before we started working together, which again, I'd like to touch on as well. But I kind of want to give a sense of what was it like on that day when we met and sort of describe your practice and what it was like.

James Fox (16:01.802)

Yeah, yeah, we had several, sorry to interrupt, but yeah, we had several studies that were rapid fire two to three month to six month studies where the inclusion exclusion criteria was fairly generous in relation to how many patients would have access to these studies, which led to a large amount of recruitment in a very short time with overlapping studies that were similar.

maybe not in the areas that they were being studied within, but in their intensity and in their ramp up. so at some point people only have so many resources that they can deal with. think all of us can relate to that. Like there's no one who's listening or whatever to talk about this stuff that doesn't understand that. And

Ram Yalamanchili (16:56.652)

I thought that was you.

James Fox (16:58.216)

Yeah, yeah, that's new stuff, right? You've never been tapped out, right? So, I mean, when we were dealing with our secondary, our secondary research coordinator having approximately 20 hours of overtime a week and them saying things like, I don't want this overtime. When can we get a solution?

Ram Yalamanchili (17:02.062)

you

James Fox (17:22.484)

knowing we needed an additional staff member right then and there, but if we hired somebody new, we'd have to train them and actually slow down our processes. You then as an investigator start wondering, well, should I not be presenting these options to patients? And that's a disservice to patients when you come from a perspective of giving patients opportunities to participate in these studies. And so also you want to...

make sure that you are contributing to studies well to help these products be evaluated fully. So we just were at, quite frankly, a breaking point psychologically. We hadn't yet hit a breaking point physically because my team was spending way too much time.

compared to what they should carving into their personal lives and their personal dedication to the research program and also to the patients within it and also to the data and the making sure things were done well. I, know, when you, when you go through that with a program, you never, and you actually care about your team, you never want to see your team in that position. And so,

And that just deals with that scalability, man. I mean, you never want to turn down a study because you don't have the staffing for it. You want to turn down studies if you don't have the patient population for it, but you never want to turn down a study if you don't have staffing for it. You never want to turn down or actually neglect to discuss anything with somebody, give them the opportunity to participate in the study because you know your staff's overwhelmed. And so...

At times it seems to be that there's a zero sum game between those where really that scalability is the part that puts a ceiling on what you're able to do or not able to do. And that's where we were when, hey, when we met, that is definitely where we were in that hearing my concerns, somebody, mutual acquaintance of ours, colleague of ours, you know, heard my concerns and

Ram Yalamanchili (19:25.516)

Yeah. You know, I

James Fox (19:42.794)

That's definitely why you heard those concerns very early on is because that's how we that's how we had the opportunity to connect with.

Ram Yalamanchili (19:48.226)

Yeah. You know, just a quick detour, right? It's very interesting hearing your perspective of where you are. And because I walked in that first time and fortunate enough to sort of like personally come in and meet you and your staff on that day, right? I have a very interesting perspective, I think. you see that. So the way I remember it is we spoke on the phone and I immediately was, I was kind of pitching you, hey, there's this whole thing called

AI teammates which we're building and we want to build this for clinical research. And I was also very open to you saying we're not there where, you know, everything we do is like 100%, right? We want, we're building right now, but we want to work with somebody who's like excited about this, wants to make a difference and can take it somewhere. And even though what you're describing right now seems like, man, I've got a lot going on. I shouldn't probably be doing something completely out of there. I don't remember you like giving me that vibe. You're very much like,

dude, this is super cool. Like, yeah, I'd love to like learn more about it. Let's, let's, let's like unwrap sort of like, you know, what's happening. And so if you recall, I came in and we spent like the whole evening, working, talking to your staff with you, taking a tour of the facility research program. And I saw this board where you had like, I don't know, 15 studies, a whole bunch of them active, a whole bunch of them recruiting and,

James Fox (21:10.774)

Yeah. Yeah. Yeah. Yeah.

Ram Yalamanchili (21:14.83)

I was just looking around, like, wait, but I only see two people in the like, you know, how are you guys doing all this work? And, and I remember your lead coordinator saying, yeah, like we're just like, you know, we're all in on this thing. do, we take care of all this work and we do it well. And it was very clear that they're, they're capable. They're very, very capable in taking on a lot of, a lot of work and doing this. But at the same time, I think, like you said,

James Fox (21:35.806)

Mm, right on. Yeah.

Ram Yalamanchili (21:42.284)

It wasn't like, you know, this is fine for the rest of our lives or like, you know, future, right? We do have to do something about this and so that, and then we started talking about it. So I want, maybe like, I'd love to hear your perspective on what it was like to sort of hear what Tilda was building. And you knew this was early. We were, we were doing essentially a collaboration of sorts where we both were agreeing on, you know, we're going to keep building. We're going to keep making it better.

And at some point, AI will get there where things are materially impactful than where it is today from an accuracy quality. And our argue over the last four or five months of doing all this, I think we're starting to see quite a bit of that turning out to be true. So what I'd love to understand is how was your initial reaction? Why did you say, why did you even try this out? Because given that you knew we were early, like late last year, right?

James Fox (22:18.592)

Yeah.

James Fox (22:39.348)

Yeah, I mean, the reason I'm in research and, you know, is innovation. So there's a lot of different perspectives that I think people have on AI. think there's people who think that it will not pan out to what it could be and it'll fade away or it will not be a part of our lives. And if it does, it may be a part of our lives where we can ask it.

silly questions and it'll respond to us in a way that's slightly better than what Google can do. Or if I ask it to paint a picture, it'll paint a picture a little bit. That picture may have six fingers on each hand, but it'll do its best. I think there's then the people who are all in AI and that's all they see and that's all they do. I think I come from a situation where I really truly believe that

with the things I've looked into AI, that AI is going to make a big difference in our world. that it, I also am very aware that AI requires training and that it requires frameworks that are made by the people who are designing how it's trained and what information and data that it can be trained on. And I saw that our problem, though it may not immediately in one second be solved by what you were offering,

It would.

Working with Tilda would allow us to see what AI was capable of within research on a very practical level on helping us with our immediate problems that may not immediately lead to solutions. But seeing the long game with it too, and knowing that these problems that we had, if we put our heads in the sand and keep doing things the way we always did them,

James Fox (24:34.378)

it would be suboptimal. I think we'd still be able to get by. I think we'd be able to do things the way they always have been done. if AI has the potential to change this industry behind the scenes the way we kind of discussed that possibly it could, that could be a real benefit sooner rather than later to what we were trying to do and what challenges we were facing. Yeah.

Ram Yalamanchili (25:01.102)

So, you know, going into sort of, it's been what maybe we've started working with your team since November. So it's been about three, four months right now.

James Fox (25:13.568)

Yeah, not long. feels like longer because we've done, I mean, we've done a lot, but yeah, not very long.

Ram Yalamanchili (25:18.567)

Hey, we also have a great time working with each other. think sometimes that counts.

James Fox (25:21.685)

Yeah, that's funny.

Time flies when you're having fun and sometimes when you're doing a lot, time doesn't fly as fast even if you're having fun doing it. Yeah.

Ram Yalamanchili (25:29.474)

Yeah. It's super fun to work with you and your team. You guys have such great ideas, and it's collaborative. It's awesome. And most importantly, you understand the process, which is also important, the process of development of something new and innovative, and how that goes. So what I was going to ask you is, looking at the, let's call it before and current,

James Fox (25:33.332)

Yeah? Yeah.

Ram Yalamanchili (25:56.31)

What's your feedback? How have things panned out so far? I I always say we're in the second innings of the AI revolution, right? So we're not there there, but we're certainly not at the starting point. There's several things which are already happening and we do a good job with. But I'd love to get your thoughts on what is the material impact? What does your staff think? What do you think?

And then I'd like to follow up with what is the feedback? Where are things where we fall short, or as an industry, or ourselves, we need to focus on and get to a better place? So both current impact and areas where we can potentially be better at.

James Fox (26:37.866)

Yeah, I think that, like I laughed when you said that we've been working together for three months. I mean, that feels three and a half months, whatever. That feels pretty absurd with how we've, what we've been able to accomplish and how we've been able to test this. And I guess that can speak to the fact that we've seen real tangible benefits in such a short period of time. And I do think that there's a lot more benefits that we can achieve over time.

and we can go into specifics of that as, is necessary. I think we're, I think the problem we were having on scalability, and I just got a close up little email there about us, another study, suppose. So, but the problem we were having with scalability, that was addressed fairly quickly. It did take some extra time for our staff to learn the system and the process. And it took more work on the front end.

then it would have taken to just keep doing things the way we were. But as we've worked well and it has led to savings of time and it's been able to lead to, and I know you had shown me some of the data was interesting. It's been leading to faster query response times. It's been leading to a lower number of manual queries that are necessary.

inputting of things is cut down on time. And so there's been real time savings. And that doesn't mean time cutting corners. It means literally things that AI on a very basic level can do for us that otherwise we're spending resources, valuable resources of our finite staff doing. So that's been the benefit now. I think one of the things that could be

Or a challenge that I did not recognize at first is how quickly some studies can like get going, you know, and start startup, you know, and how quickly you need to kind of pivot from contract to startup of the study. And it takes a little bit of time to line everything up from a, a, this case, we're working with you from an AI assistive research program.

James Fox (29:01.27)

to help us get everything in line. And I don't think that's a drag on the basis of what you are or what the sponsors are. It's just that there's not a habit yet formed of how these two groups, including ourselves, those three groups really, can make that process more efficient. And so that's, think, something that will be important in the future.

You don't want to hold up on recruitment because not everything's in line. You also don't want to get ahead of using a system that is AI driven. You don't want to start doing it without that and then play catch up later. That really takes the, it creates redundancy of how you're going to do things right from the start. And so this is the learning that I've found we've we've done. And I'm curious, what are things that you have found?

in learning in how we've worked together and what challenges you've seen and also what things you've been encouraged by.

Ram Yalamanchili (30:10.978)

Yeah, I think there's been several, frankly. think over the past three, four months, I think we've learned quite a bit. We've been very active and busy on the R &D and engineering side to address the concerns. frankly, there's some very good ideas which came out of our collaboration on how we can address this. So what you just spoke about is study startup. There is no reason study startups should not be measured in hours to a day or two.

rather than weeks. I can totally relate to what you're saying. It has not been a huge focus when we first started because we had quite a few things on our plate. And that is something we are actively working on. And we have made tremendous progress. So this is going to be cutting edge in terms of where we are going to be in a matter of weeks right now. So in the next few releases, we'll be at a place where things will be instantaneous, right? Like from a study start up perspective, setting everything up, to go. So that learning experience was definitely

Unique in the sense that you have a pretty prolific research program, and that is sort of unique, right? Like it's not everybody where you get a study, a CTA is signed, and you're ready to go. You know, once you're initiated, you're ready to go the next day. And that's the personality and the essentially like the bar you hold yourself to. Your is very, very like different and unique compared to everything else I've seen in the industry, which is wonderful, right? I think.

James Fox (31:27.99)

That is true.

Ram Yalamanchili (31:35.682)

The world would be a better place if we had more people and more programs like how you're describing it.

James Fox (31:41.472)

I'm sure we have many programs like that to be fair for anybody who's watching this. Somebody's watching this right now and they're thinking, I mean, none of us surgeons are competitive. None of us surgeons are competitive. I'm sure there's someone watching this right now and being like, man, I gotta be better than this guy. But no, I mean, I think we should hold ourselves to high standards,

Ram Yalamanchili (31:50.092)

Yeah, I'm gonna go.

Ram Yalamanchili (31:58.54)

You

Ram Yalamanchili (32:02.594)

Yeah. I mean, I'm saying relatively speaking, sure. But you you asked me what is, what was unique. I would say that was unique for me. And then we started realizing, you know what? We, we, got to, we got to get there, you know, we're going to get there quick to, meet with that sort of expectation. And that means going back to my engineering team and going back to product and sort of, you know, look at our backlog and see what, what is it that we need to do to kind of like get there. Right. And so that was super interesting.

James Fox (32:07.104)

Sure. Yes, that's fair. That's fair.

Ram Yalamanchili (32:32.206)

Another set of things I think which are really fascinating is ophthalmology as a research like vertical has some very interesting problems. Like for example, you and I spoke about how IE criteria matching for certain types of studies can be much better. As in like you can derive insight from a recruitment perspective, which could potentially save a lot of time, can help the patient be matched to that trial. I think there's a bit of that on the patient ID perspective.

we've also learned a lot about how your staff's current procedures work. So that's, that's every, every practice has something unique in that sense, like the workflow itself. And so we, we like learned about how reg documents get done in your, in your practice, how finances get managed, how stipends get managed, how is patient communication managed. And these are all sort of nuance, right? There's small changes, which we have to be okay with. And we have to train our AI teammates to say, well, Dr. Fox.

this is their SOP and this is the style in which they work. So instead of being like, this is what we do and this is how you got to go, allowing some amount of trainability in the process upfront, that was another learning for us. And so we had to build some tooling so that we can allow that, not only for you, but for anyone else who we're working with, working with multiple sites. So I think we've learned.

Nuances about ophthalmology for sure. We've learned nuances about site practices and investigator preferences. And then just from the fact that because your program is really prolific, we've learned a whole bunch about where you hold yourself from a standards perspective. And I think all that kind of goes back into our day to day because we love working with this sort of a practice because there's just so much to learn and so much to improve on.

I'm sure once we build it, then everybody benefits, right? So it's a good way.

James Fox (34:25.75)

Yeah, 100%. Yeah, I think you mentioned something like we have worked together. And I think right now AI isn't such that it can figure everything out on its own. And so having that responsiveness on you guys' side, we're responsive to things that we need to be better. And then...

Your guys have been very responsive to figure out how that is, not just immediately, but sometimes it requires some conversation like, why is this? What are the things about it? You know, and it's helped us look at our own processes as well. Some of the questions you guys are having, for us helps us look at things in a different way. so AI is not good enough right now to take out collaboration with humans and

It is the collaboration with one human to another, but make no mistake. You're going to need to, if you're, if you're growing practice or if you're growing a research program or if you're a growing company, I mean, this is a business thing. You're going to have to be training somebody at some point, the more you grow, you're going to be training a lot of somebodies. And so, um, right now we're working together. We're training one another and you're

in control of like training the AI to learn. We're not going to have to train AI again. We'll be modifying how we train it. But if someone moves along from our practice, you know, they have another opportunity or they go elsewhere or retire or what have you, we will always have that training that we put into AI behind us. And we'll be growing off of

I fully do not anticipate artificial intelligence to replace our research team anytime soon. And soon I can use hand gestures to say not anytime soon. But what I can do is I can keep our core group small and manageable and more nimble and more able to take on challenges, have less bureaucracy that can occur when you get larger and larger systems.

James Fox (36:44.01)

I haven't experienced that in research, I certainly have experienced that to some degree as a practice grows. You get into that at some point or another, it just is how it is. And so I can keep my group of trusted people small while training a partner that will not go away, whose resume grows by the inputs we put in it. And so while maintaining a staff and

quite frankly, being able to retain a staff better who is not so stressed out, who they have somebody to fill in, to chip in on certain things and take things off their plate. So this is the, this I think is the thing that I would say for people who are either questioning how AI could help, doubtful that it might be able to, or they don't want to put the time in. think recognizing that you're going to have to take time training somebody.

No matter what, until artificial intelligence is fully aware and what is it? AGI? does that stand for? I'm not an expert. What does that stand for? Yeah, artificial general intelligence. Once it can kind of be better than humans from the start without training. mean, boy, we're into something completely different. But until then it is the collaboration of humans who are working with the AI to get it done. So has it been work to work with you guys?

Ram Yalamanchili (37:50.167)

Gentlemen, yeah.

James Fox (38:11.05)

Yes, it has. Would have it been work for us to be trying to train somebody from scratch? Yes, it would. In fact, the work that AI is doing for us has allowed us to train another individual because we have time to do so. So we had three study coordinators. We started adding a part-time study coordinator. That's one of the study coordinators that was full-time.

It just didn't work out and we needed to part ways. And we couldn't think of adding another study coordinator because boy, that would take too much time training. So literally the work that AI has taken from our coordinators has helped been able to go towards towards training of a new study coordinator. You need room to breathe in order to train. And that has done that.

Ram Yalamanchili (39:03.212)

Yeah, mean, that's really well said. It's essentially the few things you're mentioning as challenges, right? You want to get to scale, but then your staff is overloaded. And you're trying to solve for these problems. And the better way to perhaps do it is you have your AI teammates. doing part of your workflow. And

you know, from our perspective at Tilda, we have been through the research program ourselves. I myself have built a biotech company. I've run clinical research programs as a sponsor. so having that trust, I think, and building a team which deeply understands research, and then coming to you and saying, you know, we're not just talking technology. We understand the concept and the importance around process for research itself. you know, you know, tying it all together from...

core competency in engineering and product building, along with core competency in delivering research as a service, right? And from our team. I think those need to be there together. Our idea is that, you you can build trust if you are just one or the other. think, especially in this domain, if you're talking about taking a sort of like a big bet, right? Where AI is going and you're building something really innovative, you kind of have to have both perspectives, both from a clinician or research perspective.

as well as from the technology perspective. And yeah, I'm really glad to hear what you're talking about. I think that's exactly what we have seen multiple times, not just with your practice, but many other collaborations we've had or we're working with right now. They very quickly tell us, this is great because I'm getting a breather. For the first time, I'm seeing my staff actually saying, if things are going to continue working in this direction, I think we'll be in a better place. At least I'll be in a better place, personally, which is great.

that gives you a lot of other benefits. Your staff is happier, you have better retention, you have better trust. You start to build a much more productive team actually from that perspective. And I think one of the things I've realized building my previous biotech company was that because research methods have been evolving, as in there's new discovery mechanisms, there's new products, there's new vectors in terms of what type of research you can perform.

Ram Yalamanchili (41:17.602)

But the methods haven't changed. Like you go to a CRO, it's the same exact methodology. You go to a site, same exact methodology. So you do end up running into these pretty big bottlenecks, essentially like large brick walls where innovators on the biotech side will come and say, well, I have this amazing sort of product innovation which I'm bringing into the market. And then you hit this like brick wall because there is no way to accelerate past that bottleneck. You have to go through what everybody else does.

I personally think that's really not the right place for us to be as a society. I think there's so much innovation waiting to happen in medicine. And there's only so many great individuals who are dedicated and would love to do great research, sort of like you and your team. And we should empower them in whatever way we can, right? So you kind of have to give them the right tooling and bring them there.

So yeah, it's definitely like AI plus a trusted team delivering it all together, working with companies or practices like yourself, collaborating, fine tuning it, and then bringing it into value. So the last.

James Fox (42:27.03)

Let me say one quick thing on that, if you don't mind. So it's not like research has been run wrongly. We've found out how it can be run well. And I think you, when, mean, for me, I wasn't thinking about AI as being a solution to our challenge because I had the brick walls around us, you know? And it's not through lack of care or lack of trying.

Ram Yalamanchili (42:30.402)

Yeah, please.

James Fox (42:54.55)

I knew that the solution to the fact that the brick walls were there were to further reinforce those brick walls or to build on top of those brick walls. But when we have the opportunity, and this goes in medicine, this goes in anything, when we have the opportunity to move those brick walls, and I think the second we think, wow, we've improved enough, we don't have these brick walls anymore, we'll run into another set of brick walls. And it may be that you run into it for years.

and then, and then blow through those brick walls eventually. But I think right now we have an opportunity to move through some of the brick walls that we have rightfully stayed within, but may not be necessary to stay within and still produce quality research. That's one thing. Another thing I wanted to say is that oftentimes people will call me Dr. Fox. expert. I'm the expert. I'm the whatever. And this is

Certainly the, this is certainly the case, right? This is for sure the case. But I mean, our study coordinators are experts, man. I mean, they're experts in their domain in a way that I'm not even an expert in what they do to the level that they are. And quite frankly, if you have a great study coordinator, they better know more about, they better be getting to the place where they know more and then they need to excel in a way where they know more than you. They should. That's, that's what you should be.

developing and growing and inspiring people to do and have people that are capable of that. But you do not want to bog down experts with things that are not necessary for their level of expertise. And so we would bring in temps, we had brought in temps to do some of these lower level tasks as it were.

But then that trust factor attempt doesn't necessarily know what they're putting in or doesn't, you know, may not recognize that, that doesn't sit quite right, or that doesn't make much sense. And so to add skill to that level, that lower level of things, and by the way, that skill has not always been uniform in our experience just yet. And we're three months, three and a half months in. So,

James Fox (45:11.104)

But we've seen that skill increase over time. And we know that that as we partner together, that AI model is going to train better and better and better. And we are going to get the point where we have high quality, know, temps working for us, quite frankly, and allowing our experts to be experts.

Ram Yalamanchili (45:32.334)

Yeah, absolutely. I personally think, at least in the domain of AI, there's only one way to go. It's going to be more and more intelligent. That is just the trend. And we have not hit a brick wall just yet in terms of where that ceiling is. So it's exciting. I think things will get materially better in the near future. And this is all going in just one direction, in my opinion. So I think one of the last areas I want to discuss is

Let's just say everything you and I are envisioning comes together or comes true, right? Like in terms of where AI can go, we have AI teammates which are not only doing, in your case, we're doing data regulatory finance. Maybe they do more. Maybe they do more workflows within each of those categories. Where do you see things going? Maybe start with...

your practice, but I also would love to understand what do you think will happen to the space because you've been in research in ophthalmology specifically for the past 12 years and you've seen where it goes, you've seen the pace and you've seen how innovation happens in the industry, but how do you envision things will change? Like is the next 10 years going to be about the same, will be different? Like and what's your thoughts on that? Paint a very broad picture.

James Fox (46:32.437)

Yeah.

James Fox (46:46.56)

Boy, very broad picture. I mean, I'm not an expert at AI like you are. know, don't know. know. Yeah. From a clinician standpoint, I, like I said, I don't have the experience you have. I've just, I'll tell you, I've used chat GPT since rather early and it's a hell of a lot better at helping revise some emails of mine that might come across a little aggressive. So I'm making fun, kind of mocking myself in this and

Ram Yalamanchili (46:51.342)

I'm asking you with the camera, if AI were there, what does that mean from a clinician's perspective or a researcher's perspective?

James Fox (47:16.534)

And so now it can be, it's become a real help. It's not a cheat code. You still have to be yourself. But I think that where AI has the potential to go is kind of absurd to what is capable. to, I mean, it's fair. hear you, Ram, to ask me what I think is going to happen over 10 years of AI. But I think...

AI six years from now will be able to tell you where it's going to go over the next four years better than any human being will be able to. So I.

Ram Yalamanchili (47:47.438)

That is super, super very interesting comment you made. It's like coming from the inside level, right?

James Fox (47:51.862)

Yeah. Yeah. I mean, that sentence right there, that's a, that sentence right there. Yeah, exactly. Like where does a caveman think that it's good that we're going to be in society in tens of thousands of years? I mean that in 10 years, things could be a whole lot different, but I think my goal biting off bite-sized chunks, I mean, quite frankly in our entire thing, I mean, I think in our whole industry in, in what

devices are available in what, basically, I mean, I don't know. I can't even, it's, it's going to expand far more than research of course, but to take up a bite size chunk of this for what I see it doing over the next six months, nine months, one year for us, because I think it's hard to project out where AI is going to be, with how nearly limitless things appear. could be at this time.

I think that where I see it is functioning on the level of one and half to two study coordinators. And it will function as more study coordinators, the more busy we are, and it will function as less study coordinators, the less busy we are. So it will help us with the scalability. I'll be able to keep my group tight and small and nimble in learning how to do things better on the areas that we're experts in.

And I think it's going to first do all our menial work and then it's going to continue to grow. I don't think it's going to replace my team. And I think there's fear. think part of why AI is not looked into is the fear of what might come. When will I Dr. James Fox be replaced by AI? And it's coming, whether we like it or not, or our fears are there or not. And so

figuring out a way where it can work in conjunction with the human beings that are the experts in your program. That's probably what I'm most looking forward to is we've been three months. I'm looking forward to seeing how our human to human collaboration will create a stronger AI partner that will allow our employees who are research coordinators to...

James Fox (50:11.562)

be more self-actualized in their ability to do things that they quite frankly have not been had time to do because they've been bogged down by other things. So many different places. mean, one thing about our, there are some study centers that do data mining and look for patients. We don't, we literally look for patients in the room that we are doctor to patient. I'm looking at the criteria of a research study and knowing that and then,

and then having a conversation with the patient in that moment. We just don't have time to assign to our study coordinators, even though it's an effective tool to do that, we got work to do. We have things that we need to do. And so it's just gonna, I think our functionality and our scope of what we're gonna be able to do as a research program is gonna extend far more than what it has already. And I do think we're a fairly successful research site.

Ram Yalamanchili (50:55.662)

I want to add something.

James Fox (51:10.856)

And for us to see that we could grow ideologically as quickly as we potentially could over the next six to nine months, I mean, that's exciting. So 10 years, I have no clue, dude, but six to nine months, I think I have a pretty good idea as to where I want to see things going.

Ram Yalamanchili (51:20.45)

Yeah.

this was great.

Ram Yalamanchili (51:31.436)

Yeah. And something you pointed out is interesting, right? Which is from your perspective, your team can grow, you're taking on more work. But at the same time, I think if there were more programs like yours enabled, I also think it'll have a really big net positive effect on medicine. At least that's my goal with building Tilda. And another way to put it is when we look at your

James Fox (51:55.478)

100 %

Ram Yalamanchili (52:01.326)

program metrics, how efficient things are, how you're consistently one of the top performers on many of these studies. And we're talking about from a quality perspective, consistency perspective, patient satisfaction, like various metrics, right? I would love to have like the world have, you know,

500,000, 10,000 of such like research programs because I do believe there will be that amount of volume of medicine or innovation coming down the pipe. It's only, to me it's only a matter of time if we are, you know, we're building foundation models which can answer core biology, basic science questions and so on and so forth. You know, you are expanding the scope of how many innovations are gonna come down through the pipeline. And ultimately that means having more prolific programs like yourself being replicated.

being assisted by teammates like ours, like AI teammates, will be a huge net positive for the industry, right? And frankly, everybody, for all of us. So I think it's, I'm really excited. think what we're showing in metrics at the moment, it's early, but like I said, it's second innings. We're not playing the game. We are in the game for sure. And yeah.

super excited to see where things go and how we can basically bring what you're talking about to life.

James Fox (53:25.792)

Yeah, innovation is a passion of mine and I looked to this as a, I looked first, our first contact was due to a need that we had and it morphed very quickly into seeing how this could be beneficial to more than just ICON iCare, Dr. James Fox and the research coordinators within and other investigators within and so forth. I mean, quite frankly, honestly, I do consider this a privilege to

work at this level of where things are. I really truly do. I think that I don't particularly care that I'm remembered for anything. I don't really care. But I do think it's a privilege to be in things so soon and to help be part of a solution that could make such a big difference over time.

If it's just you and I, Ram, over the course of our lives, that's a shame because exactly that, we need to make this have a larger scope of impact. But it is a privilege to be a part of the early start on things. And I look forward to more and more people adopting, because more and more people adopting, quite frankly, will train the AI models.

Ram Yalamanchili (54:51.118)

Thank

James Fox (54:51.306)

better so that my AI model works better. So I do look forward to adoption of this on a altruistic, like I'm excited for what it can do and on a selfish level as well, I suppose, in the fact that that will train AI models even faster.

Ram Yalamanchili (55:09.966)

Yeah, no, absolutely. Like the said, it takes a bullet. So we were certainly getting there. Great. Well, thanks for your time.