Eyeworld

84 | EYEWORLD | WINTER 2026 P RACTICE MANAGEMENT About the speakers Patti Barkey, COE Director of Surgical Sales Strategy, Practice Development Bausch Surgical Saint Johns, Florida Carrie Jacobs, COE Executive Vice President of Operations Chu Vision Institute Minneapolis, Minnesota Eric Rosenberg, DO Assistant Professor of Ophthalmology Westchester Medical Center Babylon, New York Ben Seals, MBA Chief Executive Officer Thomas Eye Group Atlanta, Georgia Gary Wortz, MD Commonwealth Eye Surgery Lexington, Kentucky A 2026 Annual Meeting session recap A session at the 2026 Annual Meeting— Aii: Adoption, Integration, Implemen- tation—covered the adoption of AI and other technologies in the prac- tice. This session, on the ASOA pro- gram, was a collaboration between the ASCRS Digital Clinical Committee and ASOA, featuring a panel of administrators and clinicians. The session began with speakers highlight- ing some AI terminology and language associat- ed with this technology and advancements. Eric Rosenberg, DO, noted that this is a whole other language, just like entering oph- thalmology to begin with. "It's a reality that's going to affect you in clinical workspaces and your private lives," he said. "It's the future of what we're doing." According to Gary Wortz, MD, the "most important" point is understanding that the LLM is a foundation that these models—ChatGPT, Claude, Gemini—are built on. "Prompting is sometimes the secret sauce to getting a good answer," he added. "It's the context you're giving the LLM or the AI. You're giving it instructions on how you want it to answer." Dr. Wortz noted that the better inputs you give the AI, the better outputs you're going to get. Carrie Jacobs, COE, commented on the difference between an agent and a bot. "I think an agent is working for you, and a bot, you're making it work," she said. "The agent is assist- ing you, and you're telling the bot what it needs to do." Patti Barkey, COE, stressed the importance of being specific when you're prompting and asking questions of the AI. "You [need] to think when you're asking the questions; you have to be specific or you end up having to start all over again," she said. Guiding the conversation, Ben Seals, MBA, pointed to the fact that this technology is a tool—it's not a business strategy, but it's import- ant for practices. "We have a plan for our busi- ness, and AI is a tool we're going to use, but AI isn't the singular strategy," Mr. Seals said. "This technology is going to be a disruptive force, and it's going to change things." His practice has broken down AI by looking at 5 pillars: 1. Pre-appointment: Scheduling, outreach, eligibility check 2. Pre-check-in: Patient education, intake forms, reminders 3. Encounter pre-doctor: Triage, data capture, pre-visit summaries 4. Doctor interaction and documentation 5. Post-interaction: Claims, follow-up, patient feedback He suggested finding a segmentation that works for your individual practice and custom- izing it for your business, your practice, and your priorities. Speakers also discussed how they are cur- rently using AI in their practice. Dr. Rosenberg noted his practice's use of a generative LLM model to answer phone calls on the front end. There's failure built into anything that's new, he said, adding that his strong recommen- dation is to "fail quickly" and try again with something else if that's the direction you still want to go. If you fail quickly, you can take a step back and look at what doesn't work. "We were early to use generative LLM for phone calls," he said, noting that he's part of a private equity group. They needed to be able to scale, but deploying this initially had some issues. When they first started, one in four people calling were not getting where they needed to, but now the systems have become a lot more efficient, he said. Dr. Wortz suggested that if you're going to start using this technology, you should pick the thing that's the biggest pain point that AI can solve. "You're willing to invest more if it's a pain point," he said. Ms. Jacobs said her practice is using AI actively on a small scale and piloting in other areas. Noting the option to use AI for calls, she prefers to have a human answering calls, but she is interested in using AI/an answering service after hours or to help with things like scheduling an appointment or answering simple questions. The session went on to highlight some key vetting questions to use, including: • Does this reduce net complexity or add to it? ASCRS and ASOA Joint Session: Adoption, Integration, Implementation

• Cover