ArborAI could be the partner boutique law practices didn’t know they needed
The system, created by UT law student Patrick Ragozzine is a research platform for boutique law practices in East Tennessee to help with case and trial preperation.
Patrick Ragozzine was sitting in a legal research class at the University of Tennessee when he realized the artificial intelligence tools his future profession leans on were not very good. Some fabricated cases that did not exist. Others missed the ones that mattered. Most cost a fortune. So Ragozzine, a rising third-year law student with no background in technology, decided to build something better himself.
The result is Arbor AI, a research tool aimed squarely at the solo practitioners and small firms that make up much of Tennessee’s legal landscape. This summer, Ragozzine is getting a beta version into the hands of a small group of personal injury attorneys in East Tennessee, betting that the lawyers priced out of expensive legal software are the ones who need it most.
“I was shocked to learn how expensive these systems were and how mediocre they are,” he said of the tools he encountered in class. The high-end platforms could run a single attorney more than $1,000 a month, he said, and still return unreliable results. “A small firm can’t afford that.”
From lemonade stands to law school
Ragozzine grew up in North Georgia and wanted to run a business for as long as he could remember, from childhood lemonade stands to knocking on doors to cut neighbors’ grass. He was convinced he would launch a company straight out of high school. He tried, and it did not work.
“It was a lot harder than I thought it would be in my head,” he said.
So he gave school another shot, starting at a two-year program before deciding on law school and earning a spot at UT. He is the first in his family to attend college, and law school was not something he had ever pictured for himself. What drew him was the idea that a law degree could be a lever for helping people.
Teaching himself the technology
Building Arbor AI meant learning an entirely new field. Leaning on the resources around him, Ragozzine taught himself the fundamentals of large language models (LLMs), neural networks and how AI systems actually work. From there he reverse-engineered the harder problem: how to turn raw judicial opinions into material a model could learn from and use to produce structured legal reasoning.
The lesson that stuck was a familiar one in machine learning. Quality in, quality out. When he compared his early results against general-purpose tools like ChatGPT and Gemini, he said his own system returned more thorough answers.
UT’s entrepreneurship community helped him turn the idea into a venture. He sought feedback from his professors, including guidance on the ethical standards for using AI in legal research, then pitched in the Graves Business Plan Competition at the university’s Anderson Center for Entrepreneurship and Innovation. He went on to take first place at the Boyd Venture Challenge. Connections at 121 Tech Hub put him in a room with other founders.
“They have been guiding me through the startup process, and I’ve learned a lot just by being around those folks,” he said.
What Arbor AI does
The workflow is built for a lawyer’s caseload. An attorney uploads intake notes from a client meeting, case facts, police reports or other evidence. The system extracts and organizes the facts, pulls relevant case law, flags potential risks or gaps in the record, and keeps track of documents the attorney still needs to obtain.
Ragozzine has deliberately kept the scope narrow, starting with one jurisdiction and one practice area. “The AI reasoning is dramatically improved when it focuses on one specific jurisdiction and one specific practice area,” he said. “So my jurisdiction is Tennessee and my focus area is personal injury.” Proving the tool with a small group of boutique and solo firms, he said, lets him refine it before opening it more widely.
The point, he said, is access. The best legal representation too often goes to those who can afford it, and he sees tools like Arbor AI as a way for smaller firms to offer strong, affordable advocacy. The goal, in his words, is “to improve efficiency for solo and small firms so that they can take more clients, they can lower their fees, and they can ultimately perform better advocacy.”
“We hope to empower these small firms to go toe-to-toe with the big firms,” he said.
Ragozzine still has the bar exam ahead of him, roughly a year out, and the grind that comes with a final year of law school. But building a company has already changed his plans. He started out wanting to practice business law. Now he is eyeing a future in patent and intellectual property work, the kind of law that lives closest to the thing he built.
Like what you've read?
Forward to a friend!