As India’s healthcare system grapples with a persistent physician scarcity, synthetic intelligence is more and more positioned as a software to ease the burden on overworked physicians.
Practically 70 per cent of healthcare professionals within the nation stated AI has elevated their capability to see extra sufferers, as per the Future Well being Index 2026 India report by Philips.
“Ours is a scientific AI platform targeted on a number of the particular realities of Indian healthcare, together with its illness burden, multilingual affected person interactions and use of native drug manufacturers.”
In an interplay with ETHealthworld, Rithika Reddy, Co-founder & CEO, and Pranav Reddy, Co-founder & CMO of ZenMD, discuss how the platform is constructed for India’s healthcare wants, the query of accountability, and what lies forward for AI in healthcare.
Q1. India faces a persistent scarcity of docs, forcing physicians to see dozens of sufferers in a single OPD session. Can AI meaningfully enhance scientific high quality right here, or is it merely serving to docs deal with an overburdened system?
Rithika Reddy: It may assist with each, however we ought to be cautious to not overstate what expertise can do.
AI can’t create extra hospital beds, prepare specialists, or resolve staffing shortages. These are a lot bigger infrastructure and coverage challenges.
The place it might make a sensible distinction is throughout the session itself. A physician seeing 70 or 80 sufferers in a single morning may additionally be reviewing outdated prescriptions, lab studies, handwritten notes, and drugs histories.
When the related historical past is organised and introduced clearly, and potential dosage points are flagged, the physician has fewer administrative distractions. The purpose is to assist docs use their restricted time and a spotlight extra successfully, reasonably than change them.
Q2. ZenMD helps features starting from drug dosing to documentation. Which use circumstances have demonstrated actual affect, and what proof suggests it improves decision-making reasonably than merely saving time?
Rithika Reddy: Proper now, the clearest and most rapid profit is time saved. Medical doctors generally use ZenMD to arrange OPD notes, dictate throughout consultations, translate regional medical paperwork, and test drug dosages.
The choice-support worth turns into extra seen when a affected person returns after a number of months. A busy OPD physician could have seen hundreds of sufferers since that particular person’s final go to. It’s troublesome to recollect each earlier prognosis, drugs, or lab outcome.
ZenMD helps deliver that historical past again into the session. It may join previous studies, medicines, and diagnoses so the doctor has a extra full image in entrance of them.
We’re cautious about making broad claims round scientific outcomes, as they’re influenced by many elements and require long-term examine.
Q3. Given India’s distinct illness burden, together with TB and antimicrobial resistance, and its multilingual affected person interactions, have you ever constructed methods for Indian scientific realities reasonably than counting on imported international datasets?
Rithika Reddy: Sure. That was one of many fundamental causes we constructed ZenMD particularly for India as a substitute of taking a world product and easily localizing the interface.
A mannequin educated primarily on Western information could acknowledge a medication equivalent to Tylenol instantly, however Indian docs and sufferers are sometimes referring to manufacturers equivalent to Dolo 650. The system wants to grasp each the native model and the underlying molecule.
Language is one other main issue. Many consultations transfer naturally between English and a regional language. Medical information could also be handwritten, lab codecs differ from one supplier to a different, and affected person histories are sometimes incomplete.
The system must account for the Indian scientific setting, the place circumstances like tuberculosis, dengue, and antimicrobial resistance are prevalent.
This fall. Hallucinations stay a significant concern with generative AI. If a software suggests an incorrect dose or overlooks a prognosis, the place ought to accountability lie, and what safeguards should exist?
Rithika Reddy: Healthcare AI ought to by no means encourage a health care provider to just accept an output with out reviewing it. The clinician stays accountable for the ultimate determination on the level of care, AI instruments ought to be a assist system.
Such methods additionally want a verification layer that checks outputs in opposition to established scientific parameters earlier than they seem within the product. For institutional use, generated references can embrace citations in order that docs can overview the supply behind the data.
Expertise safeguards are just one a part of the reply. Hospitals and clinics additionally want clear inside protocols for a way these methods ought to be used.
Q5. What’s the fundamental cause clinicians abandon healthcare AI instruments, and what should builders get proper to realize sustained adoption?
Pranav Reddy: The largest drawback is normally the workflow, not the expertise itself.
A physician who’s already stretched for time won’t hold utilizing a platform that requires a number of further logins, extra information entry, or a totally totally different method of working.
The worth must be apparent in a short time. Does it cut back documentation? Does it assist the physician discover data sooner? Does it permit them to complete their work earlier?
Healthcare merchandise are sometimes designed round how effectively they carry out in an illustration. They have to be designed round what an actual clinic seems like after ten or twelve hours of labor.
Q6. How vital is structured scientific information to the way forward for AI in India, and are initiatives equivalent to Ayushman Bharat Digital Mission (ABDM) constructing the precise basis?
Rithika Reddy: Structured information is prime. With out it, it turns into a lot tougher for methods to trade data, perceive affected person histories, or assist constant scientific workflows.
ABDM helps set up vital requirements for interoperability, however adoption is of course uneven. For instance, a big hospital in a significant metropolis and a small regional clinic could also be at very totally different phases of digitization.
So, platforms working on this setting need to work with the healthcare system because it exists as we speak. We have to course of handwritten, scanned, incomplete, and semi-structured data whereas additionally serving to clinicians create cleaner and extra standardized information going ahead.
Q7. Is there a threat that youthful docs could depend on AI for scientific reasoning, weakening unbiased diagnostic pondering?
Pranav Reddy: Sure, that threat exists, and it ought to be taken severely.
Nonetheless, college students and junior docs are already utilizing on-line search engines like google and yahoo and general-purpose AI instruments to reply scientific questions. So, the extra sensible query could also be how to verify they use these instruments responsibly.
A healthcare-specific platform ought to assist somebody test their reasoning, overview related data, and determine what they could have missed. It mustn’t turn out to be an alternative choice to understanding the basics.
Medical training can even have to adapt. Medical doctors ought to be taught not solely use AI, however query it, confirm it, and acknowledge when the output could also be incomplete or incorrect.
Q8. What measurable indicators over the subsequent 5 years ought to decide whether or not scientific AI has delivered actual worth?
Pranav Reddy: Time saved is a helpful place to begin, but it surely can’t be the one measure.
Hospitals ought to take a look at whether or not OPD groups can handle affected person move extra successfully, discharge summaries are accomplished sooner, coding and documentation turn out to be extra correct, and administrative strain on clinicians decreases.
When it comes to affected person expertise, some parameters to take a look at are reductions in prescription errors, sooner diagnoses and referrals, readability of directions, and fewer time spent ready or travelling unnecessarily.
Q9. Can expertise bridge healthcare gaps in tier-2 and tier-3 India, or will bodily infrastructure shortages stay the principle barrier?
Pranav Reddy: Bodily infrastructure will proceed to be an issue. Software program can’t herald a lacking specialist, create a hospital, or repair connections.
It may assist docs and well being employees use the sources they have already got extra successfully.
For instance, a basic practitioner in a tier-3 space may be capable of handle extra circumstances regionally if they’ve entry to good medical data and assist with selections.
This will reduce down on journey and prices for households. It isn’t an alternative choice to constructing infrastructure, however it might make healthcare simpler to succeed in whereas that infrastructure continues to be being constructed.
Q10. ZenMD is at present free to make use of. What’s your industrial technique and path to monetization?
Pranav Reddy: We’re maintaining our platform free at this stage is a deliberate determination. Our rapid precedence is to grasp how clinicians use the product of their day by day work and to make it genuinely helpful sufficient that they return to it frequently.
Our industrial mannequin is targeted totally on hospitals, clinic teams, and different institutional companions. These organizations could require system integrations, administrative controls, compliance options, personalized deployments, and enterprise-level assist.
On the identical time, we wish the core product to stay accessible to particular person practitioners. The broader purpose is to construct widespread scientific utility first and develop the enterprise enterprise across the further capabilities establishments want.
