Rich Berman granolacowboy
I build and deploy applied-AI systems for regulated organizations, grounded in more than ten years of legal-operations work. My focus is the integration, evaluation, and adoption work that moves a promising model into a production workflow.
See the workWriting
-
What regulated-industry buyers actually need before they'll adopt AI
Why trust artifacts, data boundaries, human approval gates, and production evidence matter more than demos.
-
Anatomy of a legal intake automation
A concrete teardown of a legal intake workflow, its deterministic tools, human gates, and evaluation strategy.
-
Deploying AI in a change-resistant vertical: field notes from a decade in law firms
What a decade inside law firms taught me about discovery, integration, evaluation, and production adoption.
Featured project
intake-triage-mcp
An open-source MCP server for legal intake triage. Deterministic conflict checks, matter validation, and follow-up templating, with no LLM calls inside the server. I built it to show what a secure, auditable intake pipeline looks like in practice.