Legal Tech Doesn’t Need More “Lamborghinis” ft. Kara Peterson and Richard DiBona

Everyone claims AI will revolutionize legal, but Kara Peterson and Richard DiBona want to fix the systemic bottlenecks that have broken the system for decades.

In this episode of Between the Briefs by Steno, Adrian Cea and Joe Stephens sit down with Kara and Richard, cofounders of Descrybe, an AI company focused on making the law easier to find, understand, and verify. Their story began during the pandemic, when a personal legal issue pushed them into the frustrating reality of trying to understand and access the law as nonlawyers. That experience became the starting point for a platform designed to make legal information more accessible, accurate and affordable.

What You’ll Learn:

Why the justice gap is fundamentally an operational supply-and-demand mismatch.
How inaccurate AI outputs could trigger harsh judicial regulations.
Why Descrybe moved from D2C pro se tools to empowering legal aid providers.
How issue-level citation analysis outperforms traditional "good law/bad law" tools.
Why high-quality legal tech does not require enterprise pricing tags.

This episode is a grounded look at what responsible legal AI can actually do: not replace lawyers, not magically fix the system overnight, but give under-resourced legal professionals better tools to serve more people.

Highlights:

00:00 Introduction

01:09 Kara and Richard’s Unusual Founder Story

02:21 Why Access to Justice Matters in Legal AI

02:51 AI, Legal Demand, and the Justice Gap

04:29 How AI Could Make the Gap Smaller or Bigger

05:10 Richard’s Software Engineering Background

06:58 Building From Harvard’s Case Law Access Project

07:33 Why Access to Justice Is Like a Public Health Issue

08:41 The Hallucination Problem in Legal AI

09:46 Why Descrybe Builds for Attorneys and Legal Aid Providers

13:02 Why Legal Research Is More Complex Than It Looks

13:27 The Moment They Learned About Shepardizing

14:36 Why Legal Librarians Matter So Much

16:29 Issue-Level Citation Analysis

19:03 The Four Levels of AI Accuracy in Legal Research

21:40 Why Foundation Models Will Not Eat Everything

23:47 How Descrybe’s Anthropic Collaboration Happened

25:36 What the Claude for Legal Launch Changed

27:49 Why Courts Need Better Hallucination-Checking Tools

29:31 How Descrybe’s Brief Checker Works

31:23 Scaling After the Anthropic Collaboration

33:22 Why Fair Pricing Matters in Legal Tech

36:10 Advice for AI Founders

39:03 Why the Legal AI Ecosystem Matters

40:03 Why Kara Does Not Believe in Declaring War on Incumbents

42:19 Kara and Richard’s Hot Takes on Legal AI

45:49 Final Thoughts

Quotes:

“We can’t “human” our way out of this problem. It’s just too labor intensive, there’s too much need, the cost structures don’t work. People can’t afford lawyers.”
“AI could either make the justice gap smaller or bigger.”
“We realized the better we could serve people serving the people who need to be served is where we could actually have the most impact.”
“Legal tech is really weird. You don’t have to trade really good tech for fair pricing.”
“Not everyone needs these gigantic end-to-end solutions.”