A team of students from Osgoode Hall Law School and the Schulich School of Business has won the OpenJustice Benchmarkathon, an international competition focused on advancing legal artificial intelligence (AI) held in May.
Hosted on the OpenJustice platform developed at Queen’s University’s Conflict Analytics Lab, the competition brought together teams from universities around the world to design and test AI systems for legal reasoning. The platform enables users to build workflows without coding, allowing legal professionals to translate analytical processes into structured AI outputs.
The Osgoode and Schulich team stood out for its structured approach to legal AI. Guided by Professor Martin Petrin, the students developed step-by-step reasoning frameworks that guide AI through legal decision-making. This method improves accuracy and reliability compared to using large, unstructured prompts.
The team includes Sneha Varadarajan, Blessing Uja, Tracey Gomes, Dongwoo (Tom) Lee, Caliyah Simon, Tirth Shah, Armaan Hayer, Scott Lu, Elvis Agbemelo, Arted Yzeiri, Syed Masood Raza, Alison McClean, Seif Abdelnaby and Abdulmawla Abou Merhi.
The 14 students combined legal training with hands-on experimentation using AI tools. Their work focused on translating legal reasoning into structured systems, highlighting the value of systematic approaches in legal practice.
Among the group were Osgoode–Schulich Hennick JD/MBA students Syed Raza, Caliyah Simon and Armaan Hayer, who described the competition as an opportunity to apply interdisciplinary skills.
“We built a deliberate, step-by-step process,” said Raza. “Instead of asking AI to produce a contract all at once, we guided it through a series of reasoning steps to generate structured outputs at each stage.”
Hayer compared the process to construction: “Each step builds toward a complete document.”The result was a more detailed and consistent output than typical AI-generated drafts, which can be incomplete. The team found that structured reasoning improved reliability by embedding legal analysis directly into the system. Collaboration was also key to the project’s success. “We had a very diverse group of students with different strengths,” said Hayer. “That helped us approach the problem from multiple angles. “Students also pointed to the Hennick JD/MBA program as a key factor. Training at Osgoode supported legal analysis and reasoning, while Schulich encouraged experimentation with emerging technologies. “The program pushes you to think across disciplines,” said Hayer. “You combine legal, business and technological perspectives.”
Simon emphasized accessibility in legal innovation. “You don’t need a background in coding to contribute,” she said. “What matters is understanding legal reasoning and how the law works in practice.”
Petrin said the project reflects broader changes in legal education and the profession:
“What stood out was their ability to translate legal reasoning into a structured process that AI can follow,” he said. “This reflects a shift in legal practice, where lawyers are increasingly designing how AI systems work.”
The team’s success points to a growing role for lawyers in shaping AI tools and workflows. It also highlights how legal education is evolving to prepare students to engage with emerging technologies.