Artificial intelligence (AI) is already finding its way into ophthalmology practices. At the American Academy of Ophthalmology 2026 meeting, a panel of administrators and technology experts discussed how to put these tools to work without losing the human touch. The course, “Get Ready for AI: What’s Here and What’s Coming for Ophthalmology,” featured Julia Lee, JD; Alexandra Acabá Berrocal, DHA, FACMPE, MSW; Veronica L Plessinger, COMT, MBA; Brittney Wachter, COE, CPC, MBA, OCS; and Jeffrey Daigrepont.
This transcript has been edited for clarity.
Julia Lee: Hi everyone. My name is Julia Lee, and I had the immense pleasure of leading a great panel discussion on getting ready for AI: what's here and what's coming for ophthalmology. I'm joined at today's recap by some of these administrators and panel members. So I'm going to quickly turn over introductions to the others on this video conference.
Alexandra Acabá Berrocal: My name is Alex and I am the COO of Doctors Berico & Associates in San Juan, Puerto Rico, and it's a pleasure to be here.
Veronica Plessinger: Hi, I'm Veronica Plessinger. I am the director of clinical operations at Eyecare Associates in Youngstown, Ohio, and it was a great opportunity to be part of this panel today.
Julia Lee: And Jeffrey Daigrepont, one of my favorite industry experts in all things information technology in our eyecare world.
Jeffrey Daigrepont: Thanks, Julia. I was actually going to say I am not a practice administrator, but I'm very fortunate to have a great friend and colleague, Julia, who's invited me to her panel. So honored to be here and looking forward to a great session with the other members.
Julia Lee: Yeah. So we wanted to do this session because there has been so much about AI everywhere in popular media, in industries, across the spectrum of services, and certainly in ophthalmology. And in fact, the most recent issue of AAO's EyeNet was dedicated to AI advancements on the clinical side. So really we want to focus on what AI looks like in the practice management setting, how administrators currently use it, what tools are available, and differentiating that from vendor-driven solutions. Because I feel like every time I turn around, a vendor is either implementing AI services, augmenting, et cetera. It's kind of showing up in the platforms we already use, whether or not we actively choose to embrace AI; it's just showing up bundled in with existing services. And then there's obviously vendor solutions that are brand new—everything from call center management with AI agents, processing appointment requests, and other phone calls to ambient AI scribes that electronic health record platforms are pushing out.
So in the midst of all that, I really wanted to tap administrators like Veronica and Alex who are using AI in day-to-day practice because what we anticipate is that it's going to become another tool, and administrators need to grow their proficiency. Just as most of us can't imagine running our practices without being proficient in Excel or some of these other tools that we use, AI is going to be similar. And then Jeffrey is here just to talk about the overall context of these trends and where he's seeing certain pain points, where he's seeing things going.
Case studies and real life examples were a really important part of this panel discussion approach. I'm going to turn it over to Alex and have her share some of the things she's done in her practice to make her more efficient and her practice more productive.
Alexandra Acabá Berrocal: Julia, it's wonderful, everything you've said. I'm a big believer that AI will get to us [administrators] first, even faster than the clinical side. AI on the clinical side is very expensive, and I believe that we will be the leaders in driving AI into our practices just as we have with the lean methodology. This is just a part of it: a different adoption. So in my practice, I truly use it to manage. For example, a big problem I was seeing was all of the credentials and the documents that I need to have for all of my staff. I have approximately 15 to 20 staff members right now, so I needed help. I don't have a full human resources (HR) person. So what I did was simply build a tracker into my AI platform that completely tracks all of my credentials so that I don't have to. This removes a mental load from my plate, and this is the way that I'm using it currently, and it’s also available for my staff to access it and all of the institutional knowledge that we have in the practice. We no longer have to repeat ourselves because, at the end of the day, we just want to be more efficient and leaner.
Julia Lee: And have your staff members been open to these tools to help manage their HR issues and questions and, as you said, tracking where they are if they have licensing or credentialing coming up?
Alexandra Acabá Berrocal: They have, in the sense that at the end of the day, it's quicker and more accessible than I am. And so I think this is where it's going to start. We're going to see it in little bite-sized pieces in our practices. We're all using it, Julia; like you said, our electronic medical records are already using it. It's definitely a learning process, and it's a lot of teaching, a lot of understanding the benefits of the technology, but also understanding the risks. I think that's the beauty of this panel, and having Jeffrey as well, is understanding not only the beauties of the technology but also how you need guardrails around it to be able to implement it successfully.
Julia Lee: Yeah. Thank you so much. Veronica, I know you've been doing some really cool things in your practice with AI solutions, so I would love to hear from you next.
Veronica Plessinger: Thank you, Julia. Yeah, we have faced the same problem that I think everybody else has faced. You've got all these staffing challenges, and the answer only seems to be, "Well, we need to hire more people." But I wanted to take the standpoint of challenging us to make sure that we are being as efficient and optimizing our clinical operations as best we possibly can before we consider increasing our overhead, which is already a tremendous burden on our practices. So the way I've utilized doing that is—my goal was to create the most efficient, knowledgeable clinical teams that we can, from our surgery coordinators to our scribes, to our patient educators across the board. And that was done by utilizing past presentations that we would've set in our conference room and given via slide deck to educate these folks on policies, procedures, histories, standard operating procedures, and turning them into an interactive training website that is available in web format and in app.
This also includes a chatbot that I've trained to go back to our policies and procedures within the practice and these modules we created. They're just not quoting something out there from this infinite black hole that's the internet. And so they can ask it things as simple as, “What do I do for a contact lens exam? How do I perform an OCT [optical coherence tomography]? What questions do I ask a cataract patient without having to stand for 15 or 20 minutes waiting for a fellow tech to come out of the room or a preceptor?” And I know that it is going to respond with knowledgeable, excellent information that my physicians are going to be happy with, that my billers and coders are going to be happy with, when they get that chart note back.
Julia Lee: So your resource is actually looking internally to your own materials that you've already signed off on. You know that the content is—you built it, so it's not just randomly going out into the net and pulling from who knows where.
Veronica Plessinger: Absolutely. So my standard operating procedures binder has been digitized and then uploaded utilizing AI to help convert the content into modules that the staff can open and utilize. So if they want to know about disinfection or whatever it may be, all they have to do is go to the module for standard operating procedures, and then it has converted all of that content for me. And then of course, it's AI, so then I have to go back and do a series of checks and balances to ensure all of that content came over accurately. But I like to say that this is my CliffsNotes version of ophthalmology for our team. Even have it mapped for the chatbot itself. We get this alphabet soup of acronyms and abbreviations and things that come over. So our scheduling team, medical records team, they have difficulty making sure that they schedule the patient in the appropriate place.
They can ask it, "What's a BRVO?" It's going to quote AAO iSmart content on BRVO [branch retinal vein occlusion], not what's out in the internet world. So they'll go, "Oh, this should probably see our retina doctor," and it helps guide more than just the workup technician on the floor.
Julia Lee: So you and Alex are both using readily available AI tools that are out there, whether it's ChatGPT or something of like kind, correct? I'm asking whether these are off-the-shelf AI solutions that you are incorporating.
Veronica Plessinger: So this is a website. I have no financial interest in this website, but it is called Lovable. It's a weird name, I know, I had to tell my CFO, "Hey, no nefarious activities here. This is really the website." And what it does is this company utilizes AI to help folks who aren't coders—I'm not a computer engineer—build out custom websites and apps for their practices or their personal social media, however they want to utilize this. And so it is really very, very inexpensive. I pay 25 bucks a month to have this content built out and to host this website for my staff, so I can go in and customize content. Now, there is a free version that you can utilize, so maybe you want to use this in a much more trimmed-down version for your practice. Maybe you do have a small practice. We've got over 150 employees, so that trimmed-down version didn't quite work for me, but there are free versions of that, just like there are with other things.
Julia Lee: Jeffrey, I'd love to hear from you about what you were seeing in terms of trends and how what the administrators on the panel have shared resonates with you since you've worked with so many different types of practices.
Jeffrey Daigrepont: Thanks, Julie. I actually have the exciting and fun task of answering what AI isn't and what AI is. So in 10 years when you watch this video again, I'm either going to sound extremely cringey, or I'm going to be a total genius, so we'll hopefully go with the latter. I think the tools that I've seen and even heard the panel members discuss here are very appropriate at this level, along the lines of what AI is. I think it is exactly what you heard today. They are tools that will support and enable and help us increase our capacity. What it isn't: It's not going to replace you. You have to have the subject matter expertise to feed it those prompts; otherwise you'll get slop, and we've all seen AI slop. You're seeing it on your social media, people post pictures that they create on AI and there's a hand with 8 fingers. I mean, it's not perfect.
Now, I do think it improves over time, but I seriously doubt it will ever replace the human element, especially when you get into empathy, compassion, recognizing a person who walks in who’s in pain, who’s uncomfortable, how do you show up for that? Those are things that human connection—at least I hope and pray that we'll never replace that element within our health practices, which deal with the human side of things. I don't see it eliminating all of our jobs. I certainly don't see it wiping out humanity. I think that's the extreme side of the fear of AI. So as Julia said, I would encourage everyone to embrace it. I would be very careful not to rush out and buy the next great thing based on a demo.
By the way, we've been here before. Ten years ago, how many EHR vendors all rushed in? At the height, there were over 800 EHR vendors. Now we're down to just a handful. So I think we're going to see some thinning of the herd similar to what we've seen with EHR, which means you have to ask yourself one question: Is the AI solution I'm buying today a feature or function, or is it a legitimate company and product? And you just heard 2 examples where there was enablement of features and functions that are operating today. You don't have to go purchase licenses, you don't have to buy equipment, you're not having to buy a server. Most of this technology you can subscribe to. In fact, my prediction—and this may be the cringey part—I think it's just going to inherently start to appear in our products and tools. I don't know if you're going to have to go out and purchase another platform. Nobody wants to do that.
So that's my prediction; whether it's cringy or comes true, we'll see. I guess I'll bookmark this and come back and watch it in 10 years. We'll see what happens.
Julia Lee: I'll definitely do the AI panel reunion in 10 years, but thank you so much for sharing your insights. I'm hoping that everyone who has an opportunity to listen in takes away some inspiration for how to responsibly incorporate AI into their practices. Thank you again.