The Clinician's Perspective: On bedsides, boardrooms and buzzwords: What AI can—and can’t—fix in healthcare

This article is part of our series, The Clinician’s Perspective, where we explore the intersection of AI and healthcare through the eyes of our team – former clinicians who understand the realities of patient care firsthand.

‍

I’ve worked and continue to work with several hospitals and care systems, and if there’s one thing that unites them all, it’s how chaotic things can get—each in their own unique way. People often talk about volumes or pressure as the biggest challenges - according to my job application, I apparently thrive in those conditions. I’ve struggled (and will likely continue to struggle) at the bedside and boardroom on all matters clin ops and AI. But its ok to struggle, because that’s how we learn and try to do better…

At the bedside

• Outpatients: missing or incomplete records: There’s nothing quite like the jolt you get walking into a clinic only to realize the patient records for your session aren’t there. You do a quick scan of the room, hoping they’ll magically appear. Sure, electronic health records (EHRs) have helped. But let’s be real—it sometimes feels like we’ve just replaced paper chaos with digital chaos. Those 37 clicks to confirm a patient, the endless PDFs uploaded into the system (ECGs, EEGs —culprits, I know too well). We high-five each other for going paperless, only to discover we’ve just built a new maze.

• Emergency departments: sticky notes and stress - “busy” doesn’t begin to describe it. You’re already on edge from the sheer volume of patients when you realize there aren’t enough computers to go around. So you revert to “post-it but don’t lose it”—scribbled notes stuck to every surface, handed off to colleagues to later enter into the system. It’s not inefficiency by choice; it’s self-preservation.

• Operating rooms - the awkward waiting game: have their own brand of frustration. The hardest part isn’t the complexity of the cases—it’s the silence when a patient or their family asks, “When will surgery start?” and you don’t have a clear answer. Who's going to be brave enough to call out the surgeon or anesthesiologist for showing up late? That silence sticks. It’s not about data or dashboards—it’s about coordination, timing, and human communication that too often gets lost in the shuffle.

• Inpatient units - communication, not capacity: Everyone talks about staff shortages. But what stresses me out isn’t the number of people—it’s how we communicate with the ones we have. Orders scrawled in block letters, tasks flagged “STAT” (for something that’s obviously routine), yet somehow things still lag. It’s not because people don’t care—it’s because the process doesn’t flow.

In the boardroom

Here’s the irony. We're trained to communicate—hundreds of hours spent learning how to talk with patients, families, and each other. Yet when we step into boardrooms or pitch digital initiatives, that skill somehow evaporates. Suddenly, we’re stumbling through PowerPoints, trying to justify “why AI,” “why this vendor,” “why now.” The language becomes corporate and stilted. We lose the same clarity and empathy that make us good clinicians.

So, we call in consultants to help. They polish the business case, build the slide deck, and coach the delivery. But deep down, we know: the most convincing arguments are those rooted in our lived clinical reality, not just the numbers. Almost every product demo I've done has ended with a: “AI-tool sounds great, but what's the added-value?”  That’s not cynicism. That’s curiosity. It’s a fair question: show me the connection between the data and the day-to-day reality.

On social media platforms

Can we please chill with the AI buzzwords and ground ourselves in the reality of our healthcare systems?

Start with human discipline, not AI wizardry:  AI can absolutely help. From documentation and clinical decision support to interpreting results, it can make our lives easier and safer. But that’s the catch—it helps. It supports. It doesn’t and cannot fix broken systems. If we’ve got sloppy workflows, incomplete data, or inconsistent follow-through, no algorithm can make that magically better. AI thrives on structure—it’s only as good as the processes we build and the discipline we maintain. In other words: do the work first. Then let AI amplify it.

Every day, hospitals collect masses of data without even noticing: timestamps on patient encounters, billing information, staff scheduling, bed occupancy rates, even how long it takes to transfer a patient from ward to imaging. It’s not glamorous data—it’s the kind that lives in the corners of Excel sheets and EHR backends. But it tells the true story of how care actually runs.

Hospitals have been on this for a while now - “Performance management units,” “business intelligence teams,” “health informatics divisions”—they all sound different, but they’re solving the same puzzle: how to stitch together the clinical and the operational to get a full picture of performance. Across the GCC, we’re seeing some great examples of this principle in action. In Saudi Arabia, the Ministry of Health’s “Seha Virtual Hospital” didn’t start by deploying AI—it started by fixing patient flow across 130 hospitals. Once that foundation was solid, AI-powered triage and imaging analysis tools actually worked, because they were feeding off clean, consistent data. In the UAE, Cleveland Clinic Abu Dhabi and SEHA have been piloting predictive analytics for readmission risk. But what’s interesting is that the teams spent months cleaning data and aligning workflows before the algorithms were even switched on. The result? Clinicians trust the output because they trust the process behind it. In Oman, for instance, data-driven workforce planning helped one regional hospital cut overtime costs without cutting quality of care. In Qatar, Hamad Medical Corporation’s analytics teams use machine learning to predict staffing bottlenecks days in advance. These are the quiet, unglamorous steps that make the real difference—the stuff that rarely makes headlines but changes outcomes.

The bottom line

AI is our sidekick—the one that amplifies our strength once we’ve done the groundwork. If our patient flows are smooth, our documentation consistent, and our communication lines open, AI can be transformational. But if not, it’ll just show us our chaos in higher resolution (which could in fact be que buzzword transformational). The real work is still ours: fixing the basics, owning our processes, and communicating—clearly, consistently, and with purpose—whether at the bedside or in the boardroom. That’s when AI stops being a buzzword and starts being a bridge between care and clarity.

About Dr Sara Abdel-Karim

Sara is Tandem Health's Director for the Middle East. She is a physician, public health specialist and biochemist. For the last 14 years, Sara has worked with hospitals and ministries across the GCC in the capacity of a clinician, services commissioner and management consultant.