Back to Case Studies
Apex Legal PartnersLegal & Professional Services

Building an Autonomous AI Intake System for Legal Services

Engineered an automated client intake & document synthesis agent using Claude, Zod, and Convex.

Tech Stack & Tools

TypeScriptNext.jsConvexClaude APIZodSentry

Project Impact & Results

  • 85% reduction in client intake processing time
  • Automated parsing of 1,200+ multi-page PDF documents monthly
  • Sub-second realtime dashboard updates for legal assistants

Deep Dive & Architecture

# Autonomous Legal Intake System ### Executive Summary Apex Legal Partners required a robust document synthesis pipeline to automate their inbound client consultation triage. ### Implementation Details We developed a serverless architecture using **Next.js**, **Convex**, and **Cloudflare Workers**. Inbound documents are uploaded directly to Convex storage, triggering background workers running Claude LLM function calling to extract structural data validated with **Zod schemas**. ### Key Outcomes The platform processed over 1,500 cases in its first 90 days with 99.4% accuracy.

Need a custom engineering solution?

Let's build your next system with TypeScript, Convex, and AI pipelines.

Get in Touch
Rocktown Labs — We Build the Future