Connect intake to the case
Inquiries from websites and Meta ads enter a tracked intake workflow. When someone becomes a client, their intake history carries into the case.
LegalCRM
From first inquiry to case closing.
I built LegalCRM’s Personal Injury edition to bring intake, case management, and daily practice operations into one system. My work spans product design, full-stack development, workflow automation, and the custom AI assistant at its center.
The idea is to carry context forward: from the first inquiry to the case, its documents, and the decisions that close it. Each step should build on the work already done.
Personal Injury edition.
How I built it.
Engineer the assistant for its users
I built the AI harness—the layer that connects the assistant to case information, files, memory, and tools that perform tasks. It follows each user’s role and permissions, asks for missing information, plans multi-step work, and requests approval where needed. It checks results before reporting completion.
Support the full practice
The system supports records requests, document drafting, communications, tasks, and discovery through settlement and closing review. The assistant helps carry out the work; attorneys retain legal judgment and final approval.
Tech stack
- Next.js / React / TypeScript
- The application interface and workflow logic.
- PostgreSQL / Drizzle
- Case records, relationships, and workflow state.
- Custom AI harness
- Connects the assistant to case information, tools, memory, and approval steps.
- Document engine / Integrations
- Drafting, review, lead capture, and connected communications.
