AI Learning Experiences

How LLMs work, what they can't do, and the mental models every lawyer needs.

An advanced first-principles course on what generative AI is actually doing — tokens, patterns, the ascent, representation space, the peak and descent, sampling, generation as a sequence of arcs. The follow-up to Foundations for readers who want to reason about AI capability and trajectory from first principles.

Vendor landscape, procurement, deployment patterns, and evaluation frameworks.

A craft course on directing AI for legal work. Build a portfolio of 25–40 production-grade prompts for your actual practice — drafting, research, review, citation, structured extraction.

Prompt patterns, tool use, and professional workflows — Claude for lawyers, from first task to personal playbook.

From solo productivity to orchestrated legal AI practice. Build 6–10 working Cowork-driven workflows — foundational document and matter work, scheduled intelligence, orchestrated lifecycles, war-room patterns — plus the operational playbook to defend before risk and compliance.

The vendor-neutral protocol for AI-assisted document review — collections, multi-pass review, calibration, privilege, and production.

How to know if your legal AI tool actually works. Build a complete evaluation harness — rubric, golden dataset, statistics, LLM-as-judge, adversarial set, drift monitoring, production evaluation, and a written report — for one tool you use or are considering. Buyer and maker tracks merge into a single discipline.

A guided drafting workshop for firm leaders, GCs, and risk officers. Leave with your firm's complete AI use policy drafted — preamble, definitions, approved tools, data classification, acceptable use, disclosure, supervision, training, incidents, exceptions, governance — plus a 90-day rollout plan.

The socio-economic impact of AI for legal practitioners — task-based framework, industry and macro views, global divergence, and the profession itself.

Foundations & first two apps — environment setup, codebase literacy, vibe coding, single-purpose tool plus a full-stack web app with auth and structured LLM analysis.

RAG, agents & capstone — research assistant over private documents, multi-agent workflows, deployed capstone with responsible-development analysis.

Foundations + Suites 1 & 2 — Research Monitor that scrapes/classifies/digests a legal information source on schedule, plus a Playwright test suite for your own legal AI app with rubric-based LLM testing, CI/CD, and a closed bug-fix loop.

Suites 3 & 4 — Docket Monitor against a real third-party legal system with robust failure handling, then a Filing Copilot or rationale memo. Capstone: a full enrichment pipeline with confidence scoring, source attribution, privacy controls, and a human review queue.

End-to-end litigation workflow mastery on LexGo AI — the practitioner certification track for AIDirect's Litigation Management System.