Technically Legal – A Legal Technology and Innovation Podcast
Descrybe's Quest to Democratize Legal Research (Kara Peterson & Richard DiBona)
Why It Matters
Access to affordable, reliable legal research remains a major barrier for individuals and small businesses, perpetuating inequality in the justice system. Describe shows how AI can level the playing field, enabling non‑lawyers to navigate complex case law without prohibitive fees, and signals a broader shift toward AI‑driven, cost‑effective legal services.
Key Takeaways
- •Founders built AI-native legal research app without external funding.
- •ChatGPT API enabled summarizing 100+ million case law rows.
- •Two-person team leverages engineering and marketing skills synergistically.
- •Product offers free searchable case database and AI-powered answers.
- •Focus on solo practitioners reduces costs versus legacy research tools.
Pulse Analysis
The story behind Descrybe begins in the pandemic, when software engineer Richard DeBona faced a costly employment dispute and turned to Google Scholar for answers. Frustrated by dense opinions and prohibitive legal fees, he and his marketer wife Kara Peterson saw a broader access‑to‑justice problem. Leveraging the newly released ChatGPT API, they realized they could automate summarization of millions of case law entries, sparking the creation of an AI‑native platform designed to democratize legal research for non‑lawyers and small firms.
Technically, Descrybe’s advantage lies in its data‑first architecture. The duo cleaned and structured over 100 billion tokens, applying Retrieval‑Augmented Generation (RAG) techniques before the term was popular, to embed case fragments for fast, accurate vector searches. Their patented AI Citator evaluates citation authority, turning raw case lists into actionable insights. By offering a free searchable database alongside a chat interface that delivers concise legal answers, they undercut legacy providers that charge thousands of dollars per month, positioning the product as a cost‑effective, AI‑native alternative.
From a business perspective, Descrybe targets solo practitioners, small firms, and legal‑aid organizations—segments most burdened by expensive research tools. The two‑person, unfunded team capitalizes on complementary engineering and marketing expertise, avoiding traditional overhead while maintaining direct customer support. Their roadmap includes scaling features without massive hiring, leveraging AI to act as a virtual co‑counsel. As the legal tech market seeks affordable, high‑quality solutions, Descrybe’s focus on accessibility and price transparency offers a compelling value proposition for the next generation of legal service providers.
Episode Description
Kara Peterson and Richard DiBona, the husband-and-wife co-founding team behind Descrybe, discuss the legal research platform they built designed to "democratize access to the law." The discussion explores the unique dynamics of married cofounders and how they are leveraging Generative AI to disrupt a landscape long dominated by high-cost legacy providers.
Richard, a software engineer, and Kara, a marketing expert, share their journey from a personal legal issue to building a platform that processes over 100 billion tokens of legal data. They explain why they chose to build an AI-native system from the ground up rather than simply layering a "wrapper" over existing models—a decision that allows them to offer professional-grade tools at a fraction of the traditional cost.
In this episode, they discuss:
🔷 Why legacy legal research platforms have maintained such strong moats—and how AI changes that
🔷 The difference between "wrapper" AI tools and building a system from the ground up
🔷 How their structured, multi-step reasoning process improves accuracy and reduces hallucinations
🔷 What benchmarking against the Bar Exam reveals about legal AI performance
🔷 Where Descrybe is headed as they expand their toolkit
Comments
Want to join the conversation?
Loading comments...