Apollo, Microsoft turn to AI to help hospitals learn from medical errors

An artificial intelligence system developed by Apollo Hospitals with Microsoft is being used to speed up reviews of patient falls, medication errors and readmissions. Apollo said the system has coincided with fewer recorded incidents, although the evidence published so far does not establish that AI itself has reduced patient harm.
Apollo and Microsoft developed a Daily Tiered Huddle agent to support clinical-governance teams. It brings together hospital records, flags cases, builds patient timelines and prompts clinicians for missing details. It also suggests assessments for doctors to check and amend. The tool is in production at some Apollo hospitals.
“For anyone to process 600 cases per month with incomplete data and come up with a meaningful decision is difficult,” said Dr Rohini Sridhar, Apollo’s chief of medical services. The agent is intended to reduce the time spent assembling records, leaving clinicians more time to investigate complex cases.
Apollo said more than 60% of cases reviewed by its huddles were closed within three days in November 2026. It also noted fewer falls and medication errors in July-September than in the same quarter of 2025. Dr Sridhar said staff satisfaction with error handling improved by almost 30%, based on the hospital’s annual safety-culture survey.
These are hospital-reported results published in a Microsoft case study. The account does not provide incident counts, patient-volume-adjusted rates or an independent assessment.
The system is also designed to identify patterns across departments, although Apollo said this analysis is still being developed. It flagged a cluster of central-line-associated bloodstream infections in a bone marrow transplant unit. Juliet Joji Varghese, Apollo’s director of nursing in Chennai, said the system helps staff learn from incidents elsewhere in the network.
Beyond the Apollo and Microsoft push, there are other initiatives that reflect the expanding use of AI in routine healthcare operations and diagnostics.
In April, AstraZeneca and the Telangana government agreed to introduce Qure.ai’s AI-supported chest X-ray screening in 20 public health facilities. The technology is intended to flag high-risk lung nodules for further assessment. Manipal Hospitals has also worked with Google Cloud on a generative AI tool for nurse handovers, which Google said had been rolled out across 23 hospitals by July 2025.
India is also beginning to establish a framework for responsible adoption. The Ayushman Bharat Digital Mission, launched in 2021, laid the foundation for interoperable digital health records. In February 2026, the Centre launched the Strategy for AI in Healthcare for India, and the Benchmarking Open Data Platform for Health AI (BODH).
Developed with IIT Kanpur and the National Health Authority, BODH is intended to evaluate AI models using diverse health data while protecting patient records.
The World Health Organization, in a July 2026 report, highlighted ethical and oversight challenges arising from AI in health research, including risks that existing review systems may not adequately address.
This content is published by YourStory and has not been created or edited by Hello Headlines. View original




