When I co-founded Caredose in 2017, the problem we set out to solve was deceptively simple: get chronic patients in India to take their medication on time. What we discovered over the next four years, managing over one million medicine doses across India, partnering with institutes such as Apollo Pharmacy, Max Hospitals, WHO, USAID, and the Gates Foundation, was that medication adherence is not a simple problem at all. It is arguably the most complex, deeply human, and structurally ignored crisis in healthcare in India and beyond.
Nearly a decade later, I watch the same problem from a different chair, as an institutional investor at Cedars-Sinai Technology Ventures, where I have led over $20 million in health tech investments and evaluated more than 500 companies. And I am now convinced of something I only suspected back then: technology will transform healthcare. And, AI has taken that potential for transformation to another level altogether – it will fundamentally change drug discovery, diagnostics, robotic surgery, among other aspects. And, a similar, significant impact will take place in the quiet, unglamorous, urgent problem of whether a patient actually takes their pill.
The Scale of the Problem India Is Ignoring
The numbers are staggering. The WHO estimates that approximately 50% of patients with chronic conditions globally do not take their medications as prescribed. In India, the picture is worse. Studies show that approximately 60% of Indian patients do not adhere to anti-diabetic medications, and non-adherence to anti-hypertensive medication can be as high as 80%, depending on the region and care setting.
Consider what that means against India’s chronic disease burden. India had approximately 74.2 million diagnosed diabetics in 2021, a number projected to reach 124.9 million by 2045. Chronic diseases already account for 53% of all deaths and 44% of disability-adjusted life years lost in India, with cardiovascular disease, respiratory conditions, and diabetes killing around 4 million people annually. A significant proportion of those deaths are not caused by a lack of access to medicine. They are caused by medicine that was prescribed, dispensed, and never taken.
This is not a provider or patient failure. It is a system design failure. And it is exactly the kind of large-scale, behaviour-driven, data-rich problem that AI is structurally built to address.
Why AI Fits This Problem Better Than Any Previous Technology
The IoT era, which Caredose operated in, tackled adherence through smart packaging, connected pill dispensers, and physical reminders. It worked, but it had a ceiling. Hardware costs, distribution friction, and patient behaviour variability limited how far it could scale.
AI changes the equation fundamentally. The global medication adherence market is currently valued at $5.56 billion in 2026, projected to reach $10.63 billion by 2031, growing at a CAGR of 13.86%, with Asia-Pacific forecast as the fastest-growing region at 15.02% CAGR. That growth is being driven almost entirely by intelligent, software-first tools, conversational AI that sends personalised reminders in a patient’s language and dialect, predictive analytics that identify who is likely to drop off before they do, and ML models that are now reaching ~90% accuracy in forecasting daily adherence patterns, enabling far earlier clinical intervention.
For India specifically, this shift is transformative. Smartphone penetration exceeds 700 million users. WhatsApp is already embedded in how patients communicate with pharmacies and doctors. The infrastructure for AI-powered adherence nudges, conversational, low-data, vernacular, already exists in the hands of the patients who need it the most.
Where Indian Startups Should Focus
Three areas stand out for founders building in this space.
First, personalised behavioural intervention at scale. AI can analyse dosing history, patient demographics, and dropout patterns to generate hyper-personalised adherence protocols, something no human pharmacist can replicate across a patient base of millions.
Second, the pharmacy channel. India has 850,000+ pharmacies, the most frequent touchpoint in a patient’s healthcare journey. AI tools that integrate with pharmacy dispensing workflows, predict refill drop-off, and trigger outreach at the right moment are largely unbuilt.
Third, disease-specific adherence for TB and diabetes. India carries the world’s highest TB burden and is on track to become the global diabetes capital by 2045. Both conditions have well-documented, catastrophic consequences of non-adherence. Both have documented, replicable AI intervention points. The opportunity is not theoretical; it is urgent.
Research shows that every percentage-point improvement in medication adherence saves healthcare systems between $2 billion and $7 billion annually. For India, even a 10% improvement in adherence among its diabetic population alone would represent an enormous reduction in hospitalisation, complications, and preventable deaths.
We spent years at Caredose proving that adherence can be moved, from under 50% to above 80% for our provider partners, using technology and behaviour design together. AI makes that outcome possible at a scale we could never reach with hardware alone.
The question for India’s health tech ecosystem is not whether AI will transform medication adherence. It is whether Indian founders will build the solutions before the market is captured by global platforms that do not understand the pharmacy corner in Chennai, the TB patient in Patna, or the diabetic grandmother in Jaipur who gets her prescription refilled once a month, if at all.
This is not a niche problem. It is one of India’s largest, solvable healthcare opportunities. And it is waiting.
Kinshuk Kocher is Director of Investment Operations & Special Projects at Cedars-Sinai Technology Ventures, Los Angeles. He co-founded Caredose, a MedTech startup that created the simplest way of managing regular, chronic medication. He holds an MBA from the University of Oxford.