By: KMK Rajendran, Business Head, India Operations at Forus Health
I have spent a good part of my career working around healthcare and technology. One thing I have come to realize is that the challenge doctor’s face is not always about having too much work. It is often about having too much information to process in too little time.
In India, much of the routine administrative work associated with a consultation like maintaining records, entering patient information, managing documentation and coordinating processes, is already supported by doctors’ assistants and other healthcare staff.
So, the real opportunity for AI lies elsewhere. It lies in helping doctors make sense of the information they already have.
Consider a patient who has been coming to a doctor for several visits. To understand whether the patient is improving, the doctor may need to refer to previous consultation notes, compare current findings with earlier ones, review multiple scans or images, look at test results and piece together how the condition has evolved over time. And none of this is optional. It is an essential part of good clinical decision-making.
But it takes time.
And time is one of the most constrained resources in healthcare. A doctor may have a fixed amount of time allocated to each patient. The more time spent reviewing records, comparing scans and connecting the dots, the less time remains for what patient’s value most, that is being heard and being able to understand what is happening to them.
This is where I see a real opportunity for AI.
AI can help bring together information from a patient’s previous and current visits, summarize relevant history, identify patterns and changes over time and help the doctor navigate multiple scans or data points more efficiently. The goal is not for AI to make the diagnosis instead of the doctor. The goal is to help the doctor arrive at an informed conclusion faster.
The additional time in hand allows doctor to listen more carefully to the patient’s symptoms. Enquire for additional information that might otherwise be missed, explain the diagnosis in simpler terms, discuss treatment options and next steps. Last and most important of all it gives the patient an opportunity to ask questions and leave the consultation with greater clarity and confidence.
When we talk about AI, the conversation often turns to one question: Will AI replace doctors? I don’t think that is the right question.
A more useful question is: Can AI help doctors spend less time processing information and more time with the patient?
For me, that is where the real promise of AI in healthcare lies. It is not about replacing clinical judgement. It is about augmenting it.
It is not about taking the doctor out of the decision-making process. It is about giving the doctor better access to the information needed to make that decision. And ultimately, it is about giving something back that healthcare often struggles to provide: time.
But there is one important caveat. AI in Healthcare cannot compromise on safety.
There is a fundamental difference between using AI to optimize a routine business process and using it in a clinical environment. If an AI system makes an error in a business workflow, it may result in inconvenience, lost time or financial impact. In healthcare, an error or wrong analysis can have consequences for a person’s health and life.
The future of AI in healthcare cannot simply be about making systems faster. It has to be about making them reliable, explainable, clinically validated and designed to work alongside healthcare professionals.
So, I don’t believe we should blindly accept an AI output just because it is produced quickly or looks intelligent. AI can summarize information, identify patterns and suggest possibilities. It can help a doctor navigate information and conclude a patient visit more effectively and faster. But the doctor must still be able to question it. Clinical judgment cannot be handed over to a machine or computer program. The doctor should remain responsible for the final decision and there must always be a clear way to check on the suggestions/analysis shared by AI.
There is another point that I think is often missed.
We should not introduce AI simply because it is available, trendy or the next big thing. If a doctor has to learn a complicated new system, enter information in multiple places and then spend more time checking what the AI has done, we have not really saved time. We have only moved the work around. The technology has to be designed to navigate around a doctor’s clinical workflow. It should be simple and should reduce steps rather than add them. And most importantly, the doctor should understand what the system is doing and where its limitations are.
In my view, good healthcare technology should become almost invisible in the workflow. It should help the doctor without constantly demanding attention. I also believe that the success of AI in healthcare will depend on trust. Doctors need to trust that the information being presented to them is accurate. They need to know when the system is uncertain. They need to be able to question an analysis or suggestion shared and, when necessary, ignore it. AI is only as good as the data it receives and studies. A doctor should be able to judge the output shared based on his true world experience and studies.
Patients also need to feel that technology is being used to improve their care and not to create more distance between them and their doctor. Because ultimately, healthcare is not only about data.
A doctor may notice something because of the way a patient speaks. A patient may share something important because they feel that the doctor is actually listening.
AI can process information incredibly fast. But it cannot replace the human connection between a doctor and a patient.
AI does not have to replace the doctor to transform healthcare. Perhaps its biggest contribution will be much simpler – giving doctors back some of the time they have lost to everything around medicine, so they can spend more of it on medicine and more importantly, on their patients.