Accelerating justice
On his first day inside a courtroom, Arghya Bhattacharya was struck not by the gravity of law but by the weight of paper.“There were towers and towers of paper everywhere. You could walk into record rooms and smell the rotten paper.
Outside, people were still using typewriters – not even keyboards.
Inside the courts, more work was happening on paper than on computers,” he said at a recent event hosted by the EkStep Foundation.In 2023, together with Utkarsh Saxena, who was a Supreme Court law clerk, Arghya founded Adalat AI to build a tech stack to reduce delays, streamline workflows, and improve access. A major Arghya Bhattacharya bottleneck in courts is the shortage of stenographers.
“Nearly every word needs to be written down… judges are often writing everything by hand, which slows proceedings and creates room for errors.” Adalat AI’s transcription engine understands jargon, Indian accents, and 12 languages “so nothing gets lost in transcription.
” The company has also created a court workflow system where voice AI lets judges navigate files and dictate orders. Paperless filing systems reduce reliance on typing, while WhatsApp-based chatbots help citizens track cases without navigating complex government websites. “Judges using the tech tell us they save significant time and can focus on core judicial work,” Arghya says. Utkarsh Saxena
Agentic tumour board
When his aunt was diagnosed with breast cancer, Ashish Makani witnessed firsthand the anxiety patients and caregivers experience while navigating complex medical terminology, fragmented records, and difficult treatment decisions.
The experience became a turning point, pushing him toward work at the intersection of technology, machine learning, and medicine.The engineering graduate and now consultant with Ashoka University has developed an AI-powered tumour board aimed at addressing one of oncology’s biggest structural challenges – limited access to multidisciplinary expertise. A tumour board is a group of cancer specialists – including medical, surgical, and radiation oncologists, radiologists, and pathologists – who meet to review complex cancer cases.
But bringing together such specialists is always difficult.
So, Ashish is building multiple specialised AI agents to simulate the collaborative reasoning of these experts. The system is designed to structure medical data, surface clinically relevant insights, and support more informed decision-making. It’s built using subspecialty guidelines such as those from the National Comprehensive Cancer Network.
The agents deliberate, challenge each other’s conclusions, and synthesise recommendations – mirroring the dynamics of a real clinical discussion.The platform is a work-in-progress, but has already drawn interest from academic and clinical stakeholders, including researchers from Harvard Medical School.
Helping ICU doctors, nurses
In govt hospitals, doctors and nurses in busy ICUs often struggle to balance patient care with growing documentation demands.
Srikanth Nadhamuni’s 10BedICU programme now uses AI across a range of processes to deal with such issues. The 10BedICU initiative, started in 2021, works to strengthen critical care capacity in public hospitals. To date, 216 10BedICUs have been established in 10 states. The state govt provides hospital space, medical staff, supplies, utilities, and signs an MoU to manage operating costs, wherein 10BedICU covers the capital expenditure by supplying ICU equipment, a software platform called CARE, teleICU technology, training, and community support.The latest AI tools built into the CARE platform perform a variety of tasks. They allow doctors and nurses to simply speak in their local language while AI automatically records and structures patient information into electronic medical records. They support nurses who may not always have immediate access to specialists. Nurses can ask questions in local languages and receive responses based on validated standard ICU protocols built into the system, reducing dependence on generic online information and improving confidence in clinical decisions. They also automatically generate discharge summaries using existing patient data, saving hours of manual effort and ensuring smoother continuity of care.
