How AI is proving healthcare's commercial value

Apatient may receive early warning that he is at the risk of suffering a cardiovascular event after an AI model identifies cardiac patterns that would otherwise have gone undetected.

Another who has undergone a CT scan may never know that AI vetted the images before the radiologist signed the report.

In some hospitals, doctors are spending less time writing discharge summaries because Generative AI (GenAI) has taken over routine documentation.

These represent the first wave of AI applications moving from pilot stage to everyday clinical workflows across India’s healthcare system.

Healthcare has always been AI’s toughest laboratory. Unlike retail, finance or customer service, where new technologies can be rolled out quickly, medicine demands evidence, accountability and trust before innovation reaches the patient.

After all, a diagnosis or treatment recommendation is not just another algorithmic output; it affects human lives. Before AI can become part of clinical practice, it must first earn the confidence of doctors, hospitals, regulators and, ultimately, patients. The journey is well underway, say healthcare practitioners.

Sangita Reddy, Joint Managing Director of Apollo Hospitals Enterprise Ltd, says the early phase of AI in healthcare was about proving an algorithm works. The challenge now is to make it work safely, consistently and at scale in a live clinical environment. “It is not just about accuracy. The question is if it is clinically useful, trusted by care teams and capable of delivering sustained value,” she tells BT.

Apollo has already deployed AI in imaging, nursing workflows, discharge summaries, medical coding and remote patient monitoring. “Our investments have evolved accordingly. We now invest as much on electronic medical-record integration, secure infrastructure, governance and clinician adoption as we do on the algorithm itself," says Reddy.

Apollo has developed 20 certified algorithms with 12 in production and eight at different stages of integration and deployment.  

Beyond pilots

If there is one clear winner in healthcare’s AI journey so far, it is diagnosis. Whether it’s analysing scans, interpreting genomes or identifying cardiovascular risk, AI is proving most valuable in cases where data is abundant, workflows are structured and shortage of specialists is acute.

Policy think-tank NITI Aayog says AI has the potential to significantly accelerate India’s economic growth by contributing to a threefold increase in Gross Domestic Product (GDP) by 2035.

Kalyan Sivasailam, Co-founder and CEO of 5C Networks, a health-tech start-up that builds AI-powered diagnostic infrastructure, primarily for radiology, goes deeper into the topic.

Whether it is radiology, pathology, cardiology or neurology, these specialties generate vast volumes of structured data, making them ideal for AI-driven analysis, he says. More importantly, AI has moved beyond being a buzzword in diagnostics as it is increasingly becoming an integral part of everyday clinical workflows.

Will AI replace the medical technician? “The work is highly specialised, data-intensive and constrained by a shortage of experts. In such settings, AI is not replacing clinicians but augmenting them—helping specialists analyse information faster, prioritise cases and improve diagnostic efficiency. That is why diagnostics has emerged as one of the first areas where AI is seeing meaningful, large-scale adoption,” says Sivasailam.

Healthcare leaders caution against assuming AI is a solution to every problem. “There are certain sweet spots and low-hanging fruits which can be easily handled by AI,” says Vedam Ramprasad, CEO of medical testing lab MedGenome. “But it will not be able to impact every aspect of healthcare.”  

That distinction is increasingly shaping investment decisions. Rather than deploying AI across the board, hospitals and health-tech companies are prioritising areas where structured data, specialist shortages and measurable outcomes create a clear business case.

According to a KPMG report, due to the inherently human-centric nature of certain healthcare functions, AI’s impact has been more evolutionary than revolutionary.

In areas such as administrative functions, while AI has the potential to automate several back-office processes, adoption has been slower because of lower labour costs in India.

“For many hospitals, hiring additional staff is more economical than investing in AI-powered automation, particularly when the cost of deploying advanced AI models remains high. As a result, healthcare providers continue to rely on manpower to solve administrative challenges rather than redesign workflows around AI,” says Sivasailam.  

Measuring ROI

Does the use of AI pass the return on investment (ROI) test? Hospitals and health-tech companies are discovering that scaling AI requires years of investment in data, digital infrastructure, governance, cybersecurity, workflow integration and clinician adoption.

Above all, AI has to demonstrate measurable returns in day-to-day business. The conversation has shifted from whether AI works to where it delivers measurable clinical and business value.

“People think AI is cheap because tokens are available. Remove the tokens and it's going to be very expensive,” says Ramprasad.

This is particularly relevant in a price-sensitive healthcare market like India where providers must weigh AI investment against affordability. Even in radiology, one of healthcare’s most mature AI user segment, AI-assisted reporting can cost more than the conventional reporting by human radiologists, says Ramprasad.

Access, affordability

“AI is here to stay, and it will undoubtedly transform healthcare. But we need to be realistic. It won’t solve every problem,” says Ramprasad, adding thet there are areas where AI is creating tremendous value. The others will continue to require human expertise, he says. Cost remains a critical consideration.

Reddy says in healthcare, the return cannot be measured in terms of time or money. “We look at whether a technology enables earlier diagnosis, improves clinical outcomes, enhances patient safety, reduces unnecessary variation in care, and allows our clinical teams to spend more time with patients. Those are the returns that matter the most,” she says. “AI does not change the fundamentals by which healthcare should be judged. The primary measures remain outcomes, access and affordability,” she adds.

The earliest benefits are visible in faster documentation, improved scheduling and discharge planning, better use of hospital capacity, and less time spent on repetitive administrative work.

The larger opportunity lies in prevention and clinical outcomes. If AI helps identify risk earlier or enables intervention before a patient’s condition deteriorates, it may reduce complications and the cost of late-stage care.

An EY report says GenAI applications in health and life sciences could contribute an incremental $64 billion to India’s GDP by 2030, underscoring the far-reaching potential of GenAI in reshaping the industry.  

Responsible AI

While the discussion on AI implementation and ROI goes on, data privacy and governance are emerging as equally important challenges.

Healthcare leaders argue that AI can only scale if it is built on secure data-sharing frameworks, robust cybersecurity and clear regulatory standards. While India’s Digital Personal Data Protection (DPDP) Act is expected to provide the foundation, much of the healthcare ecosystem—particularly smaller providers—still lacks the infrastructure and processes needed for compliance. Strengthening governance, they say, is fundamental to building trust in AI.

Responsible AI also requires investment in data, electronic medical records, integration, cybersecurity, validation, trusted research environments, governance and upskilling. Its economic impact will ultimately be determined not by how many tools a hospital launches, but by how many become part of everyday care and deliver sustained value. This will only be tested over a period of time. Industry experts say investment alone will not determine AI adoption and success.

In an industry where trust is as important as technology, AI’s biggest test in healthcare is only just beginning.

 

@PalakAgarwal64