In 2016, Suman Katragadda—one of the founding members of the health care analytics practice at PwC US—came to Hyderabad to meet senior government officials about public health schemes. He was scheduled to return the same day, but a delay forced him to miss his flight. Instead, he decided to drive to his home town, four hours away by road.
“That night, however, I developed severe pain in my lower abdomen and back,” he recalls. “I was admitted to hospital and diagnosed with a kidney stone. They performed a blasting procedure and I was hospitalised for two days.”
After discharge, he began experiencing intense headaches and vomiting. “The pain worsened when I sat or walked but eased when I lay down,” he says. Reaching the doctor for a follow-up proved difficult. “The hospital kept saying he was in surgery, with patients or off duty,” says Katragadda. “It took nearly two days to speak to him.”
Suman Katragadda, founder, Heaps.ai
When he did, the doctor suggested it was likely a minor spinal fluid leak from the epidural and advised him to “drink plenty of water and lie down”.
Katragadda was stunned. “If it was that simple, why didn’t anyone tell me at discharge—or at least when I called the hospital? I could’ve avoided so much pain and anxiety,” he says.
“That experience,” he says, “exposed a critical gap in India’s health care system: the lack of continuity after discharge. If I faced this after a routine procedure, what must patients with serious conditions like heart failure or pneumonia go through? Their post-discharge journey must be even harder.”
Katragadda soon realised that post-discharge care gaps are not unique to India—they are a global issue. While consulting for a leading US medical institution, he reviewed two discharge summaries for men with congestive heart failure—one was 44, the other 65. “What struck me was that both summaries were identical and extremely basic,” he says. “They simply advised: if you have fever, chest pain, or shortness of breath, call 911 or return to the ER.”
Curious, he turned to the American College of Cardiology's guidelines and found that 26 potential symptoms can emerge after discharge. “My first question was: why weren’t all 26 mentioned?” he recalls. The answer? Patients wouldn’t remember them anyway. “But shouldn’t they at least be documented?” asks Katragadda.
He then learned that 50 per cent to 60 per cent of patients lose their discharge summaries before even reaching home, and most cannot recall specific instructions within a couple of days. This glaring gap in continuity of care became the seed for Katragadda’s idea to create Heaps.ai, a predictive model-based solution for medical care management.
The application analyses a user’s comorbidities, medications, living conditions and compliance history to forecast deviations in recovery. “We realised if we could predict when and how a patient might stray from the care path, we could intervene early,” says Katragadda. “That insight shaped how we designed the system.”
Initially focused on cardiology, the platform soon expanded. “Now, we cover 16 to 18 specialities and have care pathways for 36 chronic conditions,” he says.
Today, the platform does not just monitor vitals—it provides a comprehensive care management plan that includes scheduled doctor visits, lab tests, medication tracking, adherence nudges and targeted patient education. For instance, a patient with diabetes and high cholesterol receives an integrated roadmap, rather than siloed instructions from different specialists. The system synchronises appointments and required tests and begins nudging behavioural changes days before each medical touchpoint.
Heaps.ai does not charge patients. Its revenue comes from insurance companies, hospitals, governments and corporates. “For insurers, we help reduce readmission rates,” Katragadda explains. “For corporates, we manage employee health. And for governments, it is about population health management. Think about it: if a company pays Rs100 in premium and the insurer pays out Rs120 in claims, next year’s premium won’t stay at Rs100. It’ll go up. That’s where we step in to manage risk.”
Katragadda says that for governments, Heaps.ai manages populations covered under schemes like Ayushman Bharat and state employee health programmes. “The focus is on reducing chronic and repeat hospitalisation, shortening hospital stays and improving health outcomes,” he says.
The company credits its predictive model with multiple life-saving interventions—while also helping hospitals retain patients. Katragadda recalls one such instance involving a police constable who was admitted with chest pain and breathing difficulty. After treatment and a four-day hospital stay, he was discharged and appeared stable. But on the 21st day, the system flagged a high risk of recovery deviation.
A care coach called, guided by AI-generated prompts. “One of the first questions was: ‘Can you check if one leg looks more swollen than the other?’” says Katragadda. The patient initially said he felt fine but admitted to mild swelling on his left foot. At the nurse’s request, he sent a photo—revealing fluid retention. He was rushed to the hospital and underwent surgery the same day. “That intervention likely prevented a cardiac arrest—or a far more complicated emergency,” Katragadda notes. “And it ensured the hospital didn’t lose the patient to another facility.”
Katragadda, who holds a PhD in statistics, explains that traditional care coordination typically requires 10 to 12 staff for 1,000 patients. “With our system, one person can manage up to 2,500 patients,” he says. “It is a major leap in efficiency and scalability.”
Heaps.ai’s solution is trained on over 5 million data points from multiple countries and functions as an adaptive learning platform. It already serves clients in six countries.
Katragadda adds that a significant number of deaths and repeat hospitalisations occur because patients fail to return to the hospital despite being advised to. “That’s where AI can truly make a difference,” he says.