The week of Thanksgiving, my dad drove my mom to the emergency room for stomach pain so severe she could not stand up straight. Three days of tests. Specialists rotating in and out. No clear answer. Then, almost miraculously, she started feeling better on her own. The doctors sent her home on Thanksgiving Day with some pain management and a collective shrug. We sat around the table that night genuinely grateful, thinking the worst was behind us.

The next morning, my mom could not finish a sentence without gasping for air. We called 911. A different hospital this time. They intubated her within minutes of arrival. Over the next three days, a battery of tests finally revealed the answer: A diabetes medication her endocrinologist had prescribed more than a year earlier had caused diabetic ketoacidosis and acute pancreatitis. Once the medication was stopped, she recovered completely.

Eight doctors. 120 years of cumulative medical experience. Six days of fear that I would not wish on anyone. I want to tell you what happened next, because it changed how I think about medicine, about technology, and about what is actually coming for all of us.

The artificial intelligence diagnostic experiment

Weeks after my mom came home, I could not let it go. Not out of anger. More out of obsession. I kept thinking, was there a faster path to the answer? So I decided to find out. I compiled everything I knew about my mother’s health. Her age, gender, where she grew up, her diet, every diagnosis she had ever received, every medication she was taking, and her symptoms going back to that first ER visit. I fed all of it into an artificial intelligence system.

Then I waited three minutes. The output identified her diabetes medication as the likely cause of her symptoms. It flagged diabetic ketoacidosis and acute pancreatitis. It listed the exact tests that would confirm the diagnosis. The same tests that eventually led the doctors to the right answer six days later. The confidence level was 84 percent. I sat there staring at my screen for a long time.

Understanding the nuance of artificial intelligence

I want to be careful here because the nuance matters. The AI did not have access to her labs, her imaging, or her chart. It worked entirely from the structured information I provided. It can be wrong. An 84 percent confidence rating means a 16 percent chance of being off entirely. And nothing about a language model replaces the physical exam, the clinical intuition that comes from years of training, or the relationship between a doctor and the patient sitting across from them.

But I keep coming back to the same uncomfortable question: What if the physician on call that first night had a tool that could cross-reference her complete medication history with her presenting symptoms and flag this interaction as an 84 percent probability? Does she spend six days on a ventilator? Does my family spend six days not sleeping? I do not know the answer. But I think about it a lot.

The historical pattern of technological disruption

I have spent 22 years in digital marketing watching technology rewrite industries from the inside out. I watched search engines change how patients find doctors. I watched social media change how practices build trust. I watched AI reshape advertising so completely that the agencies who refused to adapt are mostly gone now. Every single time, the people closest to the work said the same thing: This does not apply to us. Our field is too nuanced. Too human. Too complex for a machine to understand. And every single time, they were wrong.

I am not saying medicine will follow the same path. The stakes are incomparably higher. A bad ad wastes money. A missed diagnosis costs a life. But the underlying dynamic, a profession built on deep expertise now being asked to integrate a technology that processes data faster than any human ever could, is the same dynamic I have watched play out over and over again. The physicians who thrive in the next decade will not be the ones who resist AI. They will be the ones who figure out how to use it without losing what makes them irreplaceable.

Applying artificial intelligence to personal health

After my mom recovered, I kept going down the rabbit hole. I used AI to build a nutrition plan based on my specific eating habits, my schedule, my health history, and my goals. I used it to design a workout routine I would actually stick to. I fed it my numbers and it gave me back a version of healthy living that was built for me specifically, not for some average person on a chart.

In just over two months, I lost 25 pounds, my hemoglobin A1C improved, and my blood pressure came down. The plan was not magic. It was just relentlessly personalized in a way that no generic advice ever manages to be. Same principle. Comprehensive data in, specific and actionable output out. Applied to wellness instead of diagnosis.

An inflection point for modern medicine

We are sitting at a genuine inflection point in medicine, and nobody is quite sure how to talk about it honestly. AI will not replace physicians. I believe that. The empathy, the judgment, the presence of another human being who has dedicated their life to understanding suffering and healing it, that cannot be replicated by a model trained on text. But AI will absolutely change what physicians can do, how fast they can do it, and what it means to miss something that the data was trying to tell you all along.

My mom was lucky. She had a family that kept pushing, a second hospital that moved fast, and doctors who eventually found the answer. A lot of patients do not have all three of those things working in their favor. The question is not whether AI will enter medicine. It already has. The question is whether you will be the one holding the tool, or the one who finds out too late that you should have been.

I watched my mother almost die from something a machine identified in three minutes. I chose to rise and greet it. I think you should too.

Uday Rajaram is a medical marketer.