AI is everywhere right now. And the real challenge right now isn’t access, but knowing what to trust.

Every vendor has a solution. Every platform claims to be “intelligent.” Every conversation seems to come back to what AI is going to do next.

But when you step into your practice, the question becomes much simpler:

What actually works for me?

In this episode of Power Hour, host Eugene Shatsman sits down with Dr. Abdullah “Abed” Sarhan, CEO and Co-Founder of RetinaLogik, who has been working in deep learning and AI long before it became widely used. 

And what makes this conversation different is that it’s not theoretical, but practical.

They break down what AI is actually capable of in eye care practices today, where it can genuinely improve workflow and patient experience, and just as importantly, where ODs should be far more skeptical.

What AI Actually Is (And Why That Matters in Practice)

One of the first things Abed clarifies is something most people misunderstand: AI isn’t one thing.

It’s a broad field, and within it, deep learning represents a more advanced layer, where models analyze massive amounts of data across thousands of variables to make decisions.

When someone says “AI,” it can mean anything from a simple rule-based system to a highly complex model trained on large datasets. And in eye care, that difference shows up in how tools are built, how they perform, and how much you can actually trust them.

In fact, Abed points out that even today, many tools lack proper validation and structured evaluation, which is one of the reasons adoption still feels inconsistent across practices.

Where AI Can Actually Improve the Practice

Despite the noise, there are areas where AI is already making a meaningful impact, and interestingly, they’re not always where people expect.

Abed highlights a few areas that are practical right now:

• Voice agents to handle calls, scheduling, and basic patient inquiries
• Pre-testing and exam workflows to reduce friction and staff dependency
• Patient education tools that help explain conditions and treatment paths more clearly
• Data integration to bring fragmented systems into a more unified view

These are not futuristic ideas. This is real-world technology that ODs have already implemented in their practices.

For example, a significant percentage of patient calls still go unanswered, and staff are often stretched thin managing repetitive tasks. AI, when implemented correctly, can offload that pressure and improve both efficiency and patient experience.

But there’s an important nuance:

The goal isn’t to replace people. It’s to free them up to focus on higher-value interactions, especially the doctor-patient relationship and patient experience. 

What’s Being Overhyped (And Where to Be Careful)

While some areas are completely operational, others are being oversold.

Abed is very direct about this:

Diagnostics, particularly “black box” AI that claims to detect conditions like glaucoma or diabetic retinopathy without a clear explanation, should be approached with caution.

The issue isn’t that AI can’t assist.

It’s that many tools:

• Don’t clearly explain how decisions are made
• Use misleading metrics that don’t reflect real clinical performance
• Haven’t been validated in real-world conditions

And that’s where many practices get stuck.

Not because AI doesn’t work, but because it’s hard to tell which tools actually do.

How to Evaluate AI Tools (What Doctors Should Be Asking)

One of the most practical parts of the episode is how Abed breaks down evaluation.

When selecting AI tools in your practice, you need to ask the right questions.

For diagnostics:

• How is the decision being made?
• Can the model explain its reasoning?
• What data was it trained on?

For operational tools:

• Does it integrate with your existing systems?
• Has it been tested in real clinical environments?
• Can you trial it and see how it performs in your workflow?

And more broadly:

• Is the company transparent about limitations?
• Does the team have the technical depth to keep improving the product?

One of the biggest risks a practice can take is choosing the wrong AI tool that won’t evolve with your practice. 

AI in eye care is no longer a future conversation. It’s a present one.

The real opportunity is understanding what actually works, what doesn’t, and how to make informed decisions in a space that’s evolving quickly.

Eyecare BOSS Live Summit

If you’ve been to conferences before, you’ve probably left with many great ideas and intentions, and then when you get back to your practice, nothing actually changes.

Eyecare BOSS Live is designed specifically with that in mind so you can not only know what to do to reach your goals, but also have the tools, strategies, and plans to implement lasting changes once you’re back. to change that.

This is a 2 and a half-day, invitation-only event (September 16–18 in Cleveland) designed specifically for practice owners who want execution, not just inspiration.

Inside the room:

• 200 growth-focused operators
• Peer-level masterminds
• Real conversations around revenue, staffing, leadership, AI, and specialty growth

And everything is built around one outcome:

You leave with a 90-day plan you can actually implement.

If you want to be considered, click the link below:

• AI Is Not One Thing | Understand the difference between basic machine learning and advanced deep learning, and why that distinction matters when evaluating tools in your clinic.
• Focus on Practical Use Cases First | Learn where AI is already delivering value today; including voice agents, workflow automation, and patient education, without overcomplicating implementation.
• Be Cautious with Black Box Diagnostics | Discover why diagnostic tools without clear explanation or validation should be approached carefully, despite strong marketing claims.
• Ask Better Questions Before Investing | Get a clear framework for evaluating AI tools, from understanding data sources to testing real-world performance in your clinic.
• The Future Is Digitized, Not Replaced | See how AI will improve efficiency and workflows while keeping the doctor-patient relationship at the center of care.

Connect with Dr. Abdullah “Abed” Sarhan

Dr. Abdullah “Abed” Sarhan is the CEO and Co-Founder of RetinaLogik, with a PhD in deep learning and over a decade of experience in artificial intelligence before its recent mainstream adoption. His work focuses on building and evaluating advanced AI models in healthcare, with a particular emphasis on transparency, validation, and real-world clinical application. He brings a technical and practical perspective to how AI should be developed, assessed, and integrated into modern eye care practices.

Dr. Sarhan has more than a decade of experience in data science and software. He has contributed to various peer-reviewed journals and conferences in high-ranking venues. Additionally, Dr. Sarhan delivered numerous talks, workshops, and courses locally and internationally related to data science and healthcare.

• LinkedIn: https://www.linkedin.com/in/abdullah-sarhan-phd-99bb1157/ 
• Website: https://www.retinalogik.com/ 

Connect with Eugene Shatsman

Eugene Shatsman is Host of Power Hour, Managing Partner of National Strategic Group, and a TEDx Speaker. A growth strategist and consultant to hundreds of optometric practices nationwide, Eugene specializes in scaling independent eye care businesses through structured strategy, marketing science, and operational clarity.

• LinkedIn: https://www.linkedin.com/in/eugeneshatsman/ 
• Website: eugeneshatsman.com