Three hundred years. That’s how long it would take to resolve the backlog of 50 million pending cases in India’s legal system. Meanwhile, people languish in overcrowded cells waiting for simple bail applications to be heard. One prisoner spent 12 years in jail waiting for his trial to begin. It’s not an isolated instance: An alarming 70 to 80% of India’s prison population comprises these “undertrial prisoners”—people who have been arrested but not yet convicted of a crime.
“There is profound human suffering at the back of these numbers,” says Utkarsh Saxena MPA/ID 2020, noting that families slip into extreme poverty just to pay for a decade’s worth of travel and legal fees to attend endless, postponed court hearings. “For countless Indians, justice is excruciatingly slow and effectively denied.” In 2023, he co-founded Adalat AI to tackle the crisis.
Adalat AI builds tailored AI models specifically trained in the complexities of Indian courtrooms. With real-time transcriptions, document digitization, and other AI-fueled innovations, Adalat has sped up justice for thousands of people since its launch. The organization is now expanding across the Global South, where many nations experience similar bottlenecks.
“So much of the conversation in the AI world is about chips and frontier models, but there isn’t nearly enough discussion about public institutions, state capacity, and how technology can work in low-resource, low-tech settings."
The venture’s success has drawn significant recognition, including a top prize in Harvard University’s President’s Innovation Challenge. Adding to that momentum, premier startup accelerator Y Combinator invited Adalat AI to join the fall 2026 cohort at its startup school. This will provide the venture with intensive mentorship, access to a global network, engineering guidance, and support for security and scaling. Says Utkarsh, “It is very rare for nonprofits to be admitted to the Y Combinator program, and Adalat AI is one of the few Indian-founded nonprofits ever chosen.”
Utkarsh’s path to founding Adalat AI began with firsthand frustration. After receiving his LLM from Harvard Law School in 2014, he returned to India as an idealistic young trial lawyer, having just finished a clerkship on the Indian Supreme Court. The reality of the situation, though, set in when he began practicing in district courtrooms. “There was paper everywhere,” he says. “Clerks and judges were writing everything by hand.” In one of his first weeks in court, a clerk needed to retrieve a file. “He got a ladder and dragged it across the shelves trying to find my file, and it took two hours to get the right document out.”
Over his years of trial practice, Utkarsh identified many problems, including an acute shortage of court reporters and stenographers. He decided to step away from court to work in government policy, including consulting for the chief economic advisor in India’s Ministry of Finance. To gain expertise to address the issues he’d observed in India’s courts, he then enrolled in Harvard Kennedy School’s MPA/ID program.
“At HKS, I learned to apply quantitative and economic methods to understand how systems work, and I realized that court systems often missed that kind of scrutiny,” he says. “I was also learning about remarkable nonprofits working directly with public schools to improve education and public hospitals to strengthen health care.” He was especially inspired by classmate Abdulhamid Haidar MPA/ID 2021, whose nonprofit, Darsel, uses a personalized, AI-powered chatbot to educate K-12 students in low-resource areas in the Global South. “I kept wondering, where were the nonprofits working directly with public courts to improve justice delivery?” says Utkarsh.
In the summer of 2023, as generative AI took off, Utkarsh realized that the technology had advanced to the point where it could have a big impact on the problem of extreme legal delays—and Adalat AI was born. He describes the solution as an “end-to-end justice stack”—unified layers of software that handle court cases from start to finish, rather than solving one task at a time. From digitizing paper filings, automating case analysis, autogenerating hearing summaries, and providing real-time speech-to-text transcription, Adalat AI tackles the biggest pain points in the judicial system.
He and his co-founder Arghya Bhattacharya, Adalat AI’s chief technology officer, built a speech platform to solve a linguistic hurdle. Indian courtrooms are multilingual, where proceedings mix English with regional dialects. Standard, Western-trained AI models routinely crash when encountering these shifts. To solve this, Adalat AI built proprietary models fine-tuned across six regional languages, affecting more than 6,000 courtrooms in 11 Indian states—in other words, 25% of India’s judiciary.
The impact has been immediate. Adalat AI’s flagship real-time transcription tool cut case resolution times by 30% to 50% across 4,000 courtrooms in nine Indian states. In historic firsts for judicial technology, the High Court of Kerala mandated the tool be used for witness depositions and the Sikkim High Court required it for electronic filings. Next month, Adalat AI will be compulsory for courtroom transcription and case-flow management in the Andhra Pradesh High Court.
Utkarsh’s innovative solution caught the eye of TIME, which last month named him to its fourth annual list of the 100 most influential people in AI. “It was a pleasant surprise,” says Utkarsh of the honor. “So much of the conversation in the AI world is about chips and frontier models, but there isn’t nearly enough discussion about public institutions, state capacity, and how technology can work in low-resource, low-tech settings. Those questions were a big part of my education in the MPA/ID program, so I’m glad that this recognition is bringing more attention and discussion to them.”
Utkarsh, who is currently completing his PhD in Public Policy at the University of Oxford, has a deeply personal stake in the legal system. In 2022, he and his partner, Ananya Kotia, petitioned the Indian Supreme Court to legalize same-sex marriage—and as a lawyer, Utkarsh presented the oral arguments to the court. While the effort ultimately failed, Utkarsh says the experience was a reminder of the importance of systems and institutions in determining people’s rights and freedoms.
As Adalat AI continues to scale its work, Utkarsh remains focused on the people most affected. He says, “We think of ourselves as the plumbers of the court system, unclogging the pipes of judiciaries across the Global South from paper-based and manual processes. Judges and court staff are the frontline workers of justice, often working in some of the hardest conditions in the world. Our job is to give them back time to do the thing that actually matters: Deliver timely justice to some of the poorest and most vulnerable people in the world.”
