HYDERABAD: Indian hospitals are grappling with rising patient volumes, mounting documentation demands and an influx of fragmented AI tools that promise greater efficiency but often struggle to deliver measurable value. With this in mind, Baby Memorial Hospital Group (BMH) signed an MoU with UK-based DeepInfinityAI in Hyderabad recently signalling a shift towards a more integrated and accountable approach to AI adoption.
The partnership aims to bring multiple applications under one governed framework, with a focus on measurable outcomes and a clear principle: technology can assist, but the doctor always has the final say. In an interview with CE, Dr Anup Warrier, group chief of medical services, BMH, discusses how an integrated approach to artificial intelligence can reduce clinical burnout, streamline hospital workflows and strengthen patient care, while ensuring that doctors remain firmly in control of clinical decisions.
How can an integrated AI model support hospitals better than fragmented tools?
Electronic Medical Records (EMRs) have largely remained transactional. AI can build on these systems and assist both clinical and non-clinical staff. Having a single capable partner covering areas such as documentation, radiology, laboratory reporting, discharge summaries and back-office operations makes integration and ongoing governance considerably easier. Lifecycle management is equally important, as AI systems continuously learn and evolve.
How can AI help doctors with everyday clinical tasks such as preparing notes or handling laboratory reports?
Many hospital tasks are repetitive. Patient information travels through various stages, from registration to discharge, yet clinicians often have to enter the same details repeatedly. AI can make this information available wherever it is needed, eliminating the need for manual re-entry. In the case of laboratory reports, AI can generate real-time alerts for abnormal results and suggest appropriate next steps. This can reduce unnecessary escalations while providing junior clinicians with structured guidance.
Can AI help doctors make quicker decisions, particularly in critical care settings such as the ICU? Who retains the final say?
In clinical workflows, large language models often surface information but stop short of translating it into action. That gap is left to the human clinician, who remains the safety net. In the ICU, pattern recognition and predictive analytics can flag potential patient deterioration hours in advance. Image-based AI tools are already assisting in radiology. However, these technologies remain decision-support tools. Fully autonomous clinical decision-making is neither currently available nor desirable. The doctor remains the final decision-maker.
How do you ensure patient data remains safe and that technology does not overstep its role?
Rigorous safety and security evaluations of technology providers are essential. Hospitals should have an AI governance committee to monitor vendors and solutions throughout their lifecycle. Patients must also be assured that no clinical action is taken solely on the basis of a machine’s recommendation.
How far are Indian hospitals from making AI a regular and trusted part of patient care?
Three major challenges remain: fragmentation across vendors, inadequate lifecycle monitoring to identify performance drift, and a lack of rigorous value measurement. Accuracy alone does not necessarily translate into value. The government can accelerate this process by strengthening regulations, funding local studies to measure the value of AI and creating incubation environments where technology companies can work alongside hospitals to understand real-world clinical workflows.
What should the approach to AI adoption look like going forward?
AI should be implemented only in areas where it demonstrably delivers value to both patients and hospitals, and that value must be measured continuously. Clinician burnout is a major concern, with more than a third of it linked to documentation burdens. By simplifying workflows through carefully selected AI solutions, we can free up clinicians to spend more time on direct patient care. Technology must support clinical teams, not replace clinical judgement.