Welcome to Real Talk on AI and I care with your host Dr. Scott Morris. And I'm co host Dr. Rayhan Ahmed in each episode we discuss current news add in a little AI education and debate innovative AI tools and topics that will change the industry.
Welcome back to Real Talk Rayhan Philby and I'm so interested here. I unfortunately was not able to go to Switch because of some family issues. But I am so looking forward to hearing about Switch which was put on by S. Oleksandrika at his next vote.
Tell us what you went. You were excited. He was text. So Rayhan was text, audience ran was texting me all day during going, dude, you should have been here.
This was crazy. So fill me in. What did we see? Well, you know, I didn't know much to expect.
I just walked in with an open mind. And you never know with these things, Scott. Is this going to be very commercial? It is a company that's putting on a very large, obviously, a huge, perhaps even the dominant player in I care.
And so I was curious. And I was pleasantly surprised. And in many cases blown away by what I saw. So just to think about what they talked about, well, it was called Switch for a reason.
They called it Switch. And I agree with this. We're really entering a paradigm shift where technology and innovation are really at an inflection point. And of course, that's driven by devices like smart losses, which they talked, of course, a ton about and AI.
And so you can imagine a future where, and we've talked about this a lot in the Aculonics, where you're coming in and you could not just get your eye exam done talking about your retina and your cornea and various issues with your eyes, but your whole body and giving a lot of data behind that. One of my key takeaways was obviously, there's smart glasses as a huge category where meta was Essela or Alexa, looks at a car involved and you have Google and there's rumors and I think very probably smart people at Apple working on the same thing. But it's not the guarantee that's going to be our AI companion. There could be pins, there can be AirPods with headphones, but I love the idea of glasses being the form factor of choice.
Just because it fits in such a privileged position, it's on your face. It can monitor your attention span. It can monitor what you're looking at in sort of your line of sight. So it makes a ton of sense that that in my mind, that is the pole position.
That's the leader for this world of an AI companion. The other really interesting thing about that is that I don't think the future is going to be everyone wearing the same types of glasses. That's kind of boring. I'm like, God, are we really going to be like this dystopian future?
We're all walking around with these black glasses. Well, but you have an iPhone, right? That's right. I have an AirPod and they're all the same.
There's something, and I think a company like Essela, or Exotic, has sort of gets it that there's a huge fashion component here, which I did not maybe I did not appreciate as much before. And I think that's going to be a huge component. They talk about a ton of other things. It was a really fantastic conference that they're working on a lot of the most interesting sort of weirdest, coolest idea I looked at was something called neuro refraction, but basically getting a getting your fraction by just looking at an image and monitoring your EEG as the image gets clear if it determine your refraction is based off that.
No way. Yeah. And so you think about applications in terms of people with kids or elderly people, or just like maybe that's just going to be the way it's done in the near future. So a lot of really interesting, cool technologies.
And I totally agree that the switch is upon us, right? That's something we talk about a lot. So yeah, I was, I really thought it was a great event. And there's a great thing about it.
Especially right on the heels of you doing that podcast with with Kizzer about, you know, are we at the table around? Are we at the table around them? And I think that was exactly right. And the main point behind this podcast was that there are very large companies that are looking at this.
And the big fish. And there's a huge role for I care providers to get part of this conversation because our patients will be using these devices and they need to know the op-down that can act up to an optometrics issues that are involved in their time. There are absolutely a ton of issues to be aware of. So yeah, I thought it was fantastic.
I hope they, you know, it seems like it may be an annual event. I'm not totally sure there. But look forward to going next year and it may be Scott. I think I'm just so, I am so sad I did not get to go to that.
But we have Carl Spear who was kind of like the main push of that at that meeting. He's going to be on real talk here and I think about three weeks. So I'm probably going to pick his brain a little bit more about Switch and where he sees some of that going, you know, he's got an inside track into maybe the future of what exotic is thinking and planning for. So I'm super interested in having a conversation with Carl.
We've been friends for. Geez, 30 years. And so I'm very interested to see what he has to say about that. So now I didn't get to go to the second half of the week, which was Vision Expo, East in Orlando.
And you know, it was very interesting and I've been going to Vision Expo East for 27 years I think. I think that's what just a vision expo, right? This is a first year. Yeah, it's just Vision Expo.
It's just Vision Expo now. And I don't know, you know, I was, I was excited about some of the technology and disappointed about the crowds who wanted to learn about the technology. And so I think it was a little different. I mean, the exhibit hall was definitely smaller than it's been in a while.
There was a lot of discussion about operational AI at the meeting because there was a lot of phone systems, a lot of coverage and benefit eligibility engines. There were a lot of training engines, you know, how we educate people. And I think it was, there was obviously diagnostic AI, right? I mean, that's a lot exotic.
I had a big booth. TopCon had a big booth. This was there. In fact, Ultras 1, you know, innovation award.
There was really cool, interesting technologies that we've talked about before. But I saw, and I'm not the coin, yes, or thing, but I think this is the first time I saw a switch about what AI was going to do more looking at it for an operational standpoint than from a clinical decision making standpoint. And so, you know, I, as I walked around to everyone of the booths, I think I hit every single booth at some point during the two days I was there. I would say that a lot of them had AI agents and platforms running in the background.
But none of it was as much as I was hoping to see. So I kind of feel like operational AI might be the next big discussion for the next year. But I feel like we have a ways to go before it's going to make a serious difference in how we operate our businesses. Yeah.
So, like the administrative component AI, which we've always talked about as a low-hanging fruit because there's so many bottlenecks and pain points where it can help. So it's so many ways to ease the friction. Yeah. So, I think anybody's really got it down yet.
But there was definitely a major, you know, I would say one out of every two and a half booths that were there were about operational technology. Yeah. First time I've seen it be like that. Which, yeah.
Kind of says maybe there's a paradigm shift coming in, focusing on revenue generation, decreasing friction, improving efficiency, improving the patient and the provider-provider team experience. I saw more of that than I've seen at any meeting in quite some time. Are you finding among your colleagues that they are taking up solutions and actually taking the plunge into trying out these various technologies or everyone's still doing a lot of watchful waiting? Well, you know, and actually I gave a talk with Selena McGee and Brianna Ru, Brianna's been on the show before.
I gave a talk on Friday afternoon and that was exactly our subject. Is, are we taking the plunge? That wasn't the name of the talk, but that's kind of what went. And I think that the camp fault, I think there's three camps, Rayon.
I think there's the people just buying technology AI embedded because it says it's AI embedded even though it might not be who are just buying it because it's a cool looking toy. And then they're disappointed and frustrated because that really wasn't the problem that they needed to solve first. Then I think there's the people who, and I think a big chunk of it's the wait and watch going, okay, I know I have this problem. I've identified my problem, but I haven't found the right technology to fix it because it's an add-on, right?
And I think that this was one of the comments I made a couple times during the seminar. And I made it during CCO2 is I think we're at that intersection between, in the past, innovation has always been a value or a time ad, I shouldn't say value ad, it's always been a time ad. It was we're going to get this thing, we're going to have to do something else to get the results we want. So it's time adding to the system.
And I think that we're just starting to see that inversion of maybe now we're going to have time subtract. And I think there's a lot of people out there going, show me where it's really going to save me time, energy, and money, and I'll buy it. But you got to show me. And I'm not sure that all parts of the industry have really started figuring out that that's what people want to purchase is they want to purchase time subtracting solutions.
We're still in the value ad, let's just add this onto this EHR, we're going to add this onto this EHR. And we're going to go to this new portal, we're going to go to that new portal. And you got to remember that username and that password. I still think we're at that inversion.
And I think that's, I'm not trying to be pessimistic and I think vision X is always a great meeting. It has been for 20 years. I just saw a change for the first time in that the people were asking the people in the exhibit hall were asking different questions. Then I've heard them ask before people are asking, how is it going to make my life better and faster?
And that hasn't always been the questions. Sometimes it's about does it give me better patient care is it a cool tool versus is this a tool that actually works. Yeah. Maybe some sophistication there, you know, and going to the next level of of how the ads actually going to improve.
It makes me excited, Rayhan, for what askers is going to look like, what what academy is going to look like is that will we see this movement away from this is a cool tool to do this is a tool that works. And I don't know, we're going to see, you know, I wouldn't say I was optimistic or pessimistic. I think it was a realistic approach that people are starting to ask some of the right questions. Yeah.
I think so. I think it's definitely shows a maturity in the AI application of society. But speaking of new tools, I know for a fireside chat or for the next element, our our knowledge by tonight, you're going to be telling us how to build an agent, which is, you know, super exciting to me because I'm not going to build an agent. Yeah, I'm not sure I can do in two or three minutes, but I have been building just an entire series of agents the last couple of weeks.
It's been like, like, all I could do this, I could go search it or I could Gemini it, or I can just build an agent, which takes me about 20 minutes longer to actually not just figure out this question, but figuring a bunch of questions. And you and I had that conversation. I think it was you and I had that conversation a couple weeks ago. Is that I think we're going to have in the future a team of humans that work in our practice and a team of agents that work in our practice.
And those humans and agents are going to work together to get us the right answers. But but that means that we have to have a team of agents, right? So I think let me kind of start with maybe a little bit about why an agent. Well, I think we need to think about agencies.
The primary goal when I'm building one of these or when one of my colleagues is building one of these is how do we eliminate brain drain? You know, I some people call it cognitive friction. I think it's brain drain. Like how many times do I sit there and go instead of searching through I okay, here's a very example.
My my staff manager, my office manager, the other day was we were getting ready to hire any tech and she's got we need to update our manual and this is change and that's changed. We do something and I said, oh my gosh, how about we just create an agent and operational agent that does that for us and she's like, what do you mean? And I said, well, why don't we give it to our current, I don't know, our manual is probably 56, 57 pages long. I said, why don't we go in there and we're going to say.
Why don't we take a look at all the new laws in the state of Colorado in terms of staff employment, look at all the challenges we've had in terms of write ups. We're going to throw it all in there and say agent, I want you to make me a new manual. It's up to date with all of the corporate HR stuff, the topping in the state of Colorado and all type of stuff. And I'll tell you what, I did it and it was like it printed out.
Now it was 53 pages. All the updated stuff and it did in about 60 seconds. And I was like, okay, well, that's kind of like a Gemini search. If I just put the right stuff in, but then I said, okay, now what I want you to do is I want you to develop a training set based on that manual.
And that agent built an entirely training session for my new staff about how to do that. And all I did was say, do it. I didn't have to program it or do it ever. So I was like, I almost hired and I don't think it's perfect yet.
But I hired an HR management agent that I had to build. But now we're starting to go, how else could we use this agent? Same thing I started saying, okay, so one of my one of my student docs was like, what's the algorithm for looking at a patient with red? I what's your algorithm?
And I said, well, this is my algorithm. But and I'm pretty good at dry. I have a treat drive for 30 years. I've spent thousands of hours on the podium.
I've got hundreds of articles on dry eye. And I thought, you know what? Why do I'm sitting there trying to figure out an algorithm? I literally went in and I dropped everyone of my presentations.
I dropped everyone of my papers in and then I said, and I want you to go research the web to find out I want you to build me a algorithm that helps figure out dry eye. And I'll tell you what, it was pretty good. It was perfect. But how long would it have taken me to do that to sit down and develop all that?
Even though I would consider myself somewhat expert in it, I've done it a lot. It would have taken me, see, 30, 40 hours. Yeah. Well, with this agent, it did it all in a few minutes.
And I keep, now we had another patient today, a young lady who came in, she was very young, horrible dry eye, blah, blah, blah. And I said, okay, I pulled up the agent. I said, okay, now I have a 12 year old with completely tear deficient dry eye with horrible staining, blah, blah, blah, using the protocol that we've developed, tell me what the problem is. And at the end, the agent said, you should probably look at some blood work.
You should look at what a testosterone level is and you should look at what a vitamin D3 level is. And I was like, yeah, you're right, I should. But I hadn't really thought about it, right? I mean, the vitamin D thing living up here, she's a really fair skin blonde.
That's a vitamin D deficiency is a really big issue at this time of the year. But so this agent helped do it. And so now what I'm doing is we're going back and we built an agent to look at almost like our CFO to look at our current KPIs. And now it's going back and looking at the last seven years worth of data to tell us where we could improve our stuff where we can improve.
Is it a revenue, where's our revenue per hour revenue per minute problem? And it's actually running in the background right now. So I'll be very interesting maybe on our next podcast for me to tell you what it says after it looks seven years of data to say, here's what we need to fix things. How long would it take me to do that?
No time because I never freaking do it. It would just be too hard to look at all the specifics. But and I'm sure the agent won't be perfect. But guess what?
Neither would the CFO hire you. I'd have to take the CFO and train and go, right, that was good. But what about this? So I think there's a lot of us.
So how do you build it? Right? So what's the steps? And you know, I think in the past, I used to think of building an agent like I need to code for this.
But the reality, it's not really that way anymore. Now I have to say, let me give a starting point. Let's see what your outcome is. Let's curate that a little bit and go, this part was good.
This part was bad. How do we fix the bad parts and how do we amplify the good parts? And so the way we have to do that is very simple as we take and look and say, okay, let's take the data that I have. And you know, for the audience right now, I'm writing an article that will hopefully be in next week's journal called, what does a minute cost?
We look at revenue per minute, not revenue per hour, but revenue per minute. And I realize that our, we have little to no data in I care. Our data really pretty much stinks. Is that as I sit there and try to pull resources from the last 20 years of journals.
The data is just bad. And so I said, well, let me use my own data. And I plugged in 25 years worth of data when I was developing this financial agent and said, okay, let me look at. What the revenue per patient is and how that's affected by what my revenue per frame is and my revenue per I care per lens pair is and how that's impacted by how many hours were in the clinic, you know, what my revenue per hour is.
And so I don't know what the answer is going to be right around, but we're going to find out. Right. So kind of to collect your proprietary data and that might involve like, you know, your financial data or your HR manual or your service protocols for our med spa. Right now my med spa person is developing all these protocols for how we do different things in our spa.
And then you take it and look at now look at my standard operating procedures. Now, maybe you have a standard operating procedure, maybe you don't, maybe you need to go develop one and then plug it in. But you take all of this data, right? We need data and you and I always talk about you say data is, you know, the new oil.
I completely agree. And right now I think that we all have this immense amount of data in our practice that we may be not can access because it's stuck in some EHR. But we have lots of places we can grab that data. Then we have to develop a system.
Oh, I hate to use the word prompt, but let's say personality, right? Is I want to give the agent a personality and say, okay, so you're the lead financial analyst in our practice and we're in a high end practice. So I want you to help identify where are places that we're missing out on revenue opportunities based on the last 20 years worth of data, comparatively to what national standards that you can find. So I'm going to give it a personality and say, I want you to not be nice.
I don't want you to tell me, oh, you're doing great in this subject. I want you to just get down into it and say, you're messing up in this particular area. Now, I will tell you, I think if you were going to ask me my gut instinct to what this thing my agent is going to report to you tomorrow, my agent's going to report to me tomorrow morning round. It's going to tell me that we are spending too much money on our frames.
Number one, because we've chosen not to use certain vendors. And I think it's going to tell me that our lab bill is too high, even though our remakes are low, comparatively to what we should be paying. Simply because we have to be using certain labs for certain vision and insurance plans. That's why I think it's going to tell me I'll let you know tomorrow.
So you kind of have your brain, which is the knowledge and then you have your personality like, don't be nice to me. Don't don't tell me what I want to hear. Tell me what I need to hear. And then you're going to say, OK, so now what I want to do is connect this agent to internal chats that I've had.
So we went back and we we use, you know, we talk about this round as I use Fireflies a lot. So every one of our staff meeting, every one of our financial meetings, I got Fireflies on. So what I'm doing is taking all of those audio files and or conversion to text about anything that I do with financial and I say, OK, I want you to integrate what we're saying. And then I sort of add some suggestions here to let folks know and you've grab your pen.
Do it alone. I actually use corrections and then I do those slidesно Please response to just what I do here, and anytime that I actually want me to register, We're going to look at what your EOBs were. We pumped in about 100 EOBs, and we pumped in a bunch of our lab bills and said, I want you to look through this and find where the problem is. Now, how long would it take me to do that?
Zero time because I just wouldn't. It's too much. Right? And once it's done with this, we want to build an agent.
That's my next agent I want to build right now. I want to build an agent to go in and look at profitability for all of my managed vision care plans. Not an agent I think I'll build a cell, by the way. But I want to build an agent that we'll look in and go digging into all of the EOBs, all of the lab bills, all the reimbursement, look at our contracts and say, which plans should I dump and which plans should I keep?
So how long would it take my billing person to do that? She's been working on it for almost a year. And I still don't have the answers I want. And she's really bright and been with me for 20 years and gets it.
So I look at we're going to build agents. And no agent's going to be perfect on the first go-around, just like no staff member you're ever going to hire, no human staff member, wherever going to hire is going to be perfect in the first place. But these agents are going to get smarter and better. And they're going to, I believe, if you ask them and curate the right stuff, saying, this is the feedback I want that they're going to give us feedback that then we as humans make better decisions.
Exactly. And the next step with the snology basin hand is then acting on their own and autonomously and doing these tasks maybe had an interval in making appropriate changes based off what your goals that you've communicated to are. Yeah, exactly right. I mean, we're going away from LOMs where you're feeding them information to creating these agents that are standalone that you're right.
I mean, maybe all these years I've been taking our frame sales and our EOBs and going, yeah, they're in a file cabinet. Now instead, I'm like, no, we need to save them and give the agent access to that library of data to help it come up with better answer. And I agree with you. I hadn't really thought that's a great point, Rayon, is that maybe I need to say, and I want you to do this every 90 days.
Right. I didn't do that. That's a great idea for the audience. This is how you get better, right?
Rayon and I are like bouncing ideas off each other. They go, hey, he take my agent, make it. Do this every 90 days. And maybe we're going to see this trends.
We have an agent, and that's another thing I got to set them as. I think about nine months ago, something changed within a specific vision discount company in terms of the way we were being reimbursed. It didn't tell us. But both my billing person and me within a week of each other said something changed.
I can't put my finger on it, but something changed. And I asked her, did you get a notification from this particular company? And she said, no, I said, neither did I. But you and I both know something changed.
And interested enough, then I started talking with one of my friends who lives in a different state, not to put him on the spot. And he called me like a week later, and he goes, have you been watching this company or reimbursement? I said, dammit. This is a third person that's told me this, in less than a week, something changed, right?
But I've been trying to figure it out for a month now, and I can't figure it out. Maybe when we use this agent, the agent will figure it out. Well, let us know. I think I gave an number of updates for us for the next podcast, Scott.
So let's see what the agents, if they do their homework, they'll come back with an answer. You probably will tell my health frustrated or mad, I'm next week about whether that was a good or bad thing. Maybe agents aren't going to reduce my stress. Maybe they're going to increase my stress.
That's a whole another conversation about how I actually change in our brains and causing stress. But for next time. For next time. I hope you learned something today, hopefully.
So please take advantage. Go to the meetings. Switch sounds like it was super interesting. I'm so sad I missed it.
Vision Expo, I think I got out of this. Are we seeing a transition into people thinking about how operational AI will make their light better? And last one, at least, is will we have a team of humans and virtual synthetics that work together to provide us answers that maybe we just thought on an intrinsic gut value this was happening, whereas now we'll actually have agents and people work together to give us answers of how to build maybe not just a better practice, but how to provide better care. That's a fantastic way to end.
Thank you everyone for listening. Of course, as always, check us out at AI and Icare.com. You can email us at Scott. That's SCOT.
One T at AI and Icare. AI or Rehan at AI and Icare. Thanks so much and we'll talk to you next time. You've been listening to Real Talk, an AI and Icare.
You're weekly podcast to keep you informed about AI technologies revolutionizing Icare.