Prashanthi Nuthi, VP at Enlace Health, helps organizations scale AI, technology and operations with strategic clarity and measurable impact.
Before I was diagnosed with cancer, I had already learned how overwhelming healthcare can become.
It started when something suspicious showed up on my mammogram. One test led to another: an ultrasound, a diagnostic mammogram, another diagnostic ultrasound, an MRI and eventually, a biopsy.
But there’s one moment I still remember clearly. I was sitting in the diagnostic mammography room when I was asked whether I wanted a 2D or 3D mammogram.
I didn’t know the difference. They explained that 3D provided more detailed imaging, but it could be more expensive and might not be covered by my insurance.
So, I asked what seemed like a reasonable question: “Does my insurance cover it?” They weren’t sure.
I was told I should have checked before coming in. If I couldn’t make a decision, I might have to reschedule. There I was, already worried because something abnormal had been found, trying to make a medical and financial decision I didn’t feel equipped to make.
I had taken time off from work. I didn’t want to delay the testing. But I also didn’t know how much it would cost me if I chose the 3D option. So, I chose the safest option (2D) I knew would be covered.
The tests continued. By the time I eventually had the biopsy, I was so overwhelmed that I blacked out afterward. And this was before I even knew I had cancer.
Years later, when I hear conversations about AI transforming healthcare, I often think about that room. Could all the intelligence we’re building have made that one moment easier?
A Choice Without Context Isn’t An Informed Choice
My experience isn’t simply about mammography or insurance. It’s more about fragmentation.
The clinical and benefits information existed, and so did the imaging options. But at the moment I needed to make a decision, those pieces didn’t come together.
Patients across the country are making decisions where cost and care collide. KFF reported that 36% of U.S. adults polled in May 2025 had skipped or postponed needed healthcare because of cost in the previous year. Even among people with insurance, 37% reported doing so.
For people dealing with cancer, there’s even a term for the burden created by healthcare costs: financial toxicity. The National Cancer Institute notes that people with cancer and cancer survivors are more likely to experience financial hardship than people without cancer.
The CARES Framework
This is where I believe AI has an opportunity to do something much more meaningful than automate another task. I think about that opportunity through a framework that I call CARES: connect, anticipate, respond, engage and strengthen.
C: Connect the data.
Healthcare has plenty of data. Patients shouldn’t have to connect the dots.
In my situation, the clinical recommendation and coverage information lived in separate worlds. What if the relevant clinical information, benefits and appropriate options could have been brought together before I was asked to make that decision?
Connected data doesn’t mean better dashboards. It should create better context for decisions.
A: Anticipate the need.
I was told I should have called my insurance company before arriving. But how was I supposed to know what imaging choice I would be asked to make before I got there?
If a procedure could create a coverage question, why not identify it before the patient is sitting in the examination room? AI may predict a clinical risk, but what if it can anticipate the next point where the patient’s journey might break down?
R: Respond with intelligence.
Knowing that a problem may occur isn’t enough. Someone or something has to act on it. That might mean that coverage is verified or the patient receives understandable information about the options beforehand.
A prediction sitting in a database doesn’t create impact. Intelligence becomes valuable when it changes what happens next.
E: Engage the human.
This is where I believe we need to be careful. AI isn’t for eliminating the human in the loop. If AI in healthcare is done well, it should help the right human become involved at the right time.
Technology can connect information and identify patterns, but there are moments when a patient needs someone who recognizes the fear behind the question. The goal should be to preserve compassion and remove friction.
S: Strengthen through learning.
Finally, we have to measure whether any of this truly made healthcare better. A good place to start is by asking the following:
• Did someone receive care sooner?
• Was a delay prevented?
• Was confusion reduced?
• Did the patient make a more informed decision?
• Did health outcomes, experience, equity or cost improve?
We also have to ask who the system missed—not only the people it failed to identify, but those it identified too late for the intervention to make a meaningful difference. Responsible AI is about continuously understanding its impact on real people.
Building A Healthcare System That CARES
Looking back at that day, from a patient’s lens, I wasn’t simply choosing between two imaging options. I was choosing my quality of life based on the limited context I had.
And that’s why, as healthcare leaders invest in AI, I believe we should ask not only “What can we automate?” but:
• Where are people getting lost?
• What information isn’t reaching them when they need it?
• Where can we recognize a need earlier?
• How can technology ensure the right people are brought in at the right time?
AI will undoubtedly make healthcare smarter, but smarter may not mean better. Perhaps one way to get there is to build an AI-enabled healthcare system that truly CARES.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?