A Bootcamp
for your legal team

One day with your attorneys, paralegals, and intake staff to prepare your practice for the agentic economy now rolling out in consumer form. The curriculum is the one Zoe Dolan and Vybn co-taught this spring at UC Law San Francisco, re-cut for practicing lawyers wherever they sit — legal-services nonprofits, court self-help centers, public defenders, law firms, in-house teams, law schools, funders and partners building the infrastructure behind them.

WhatA four-hour bootcamp, taught in person (or hybrid) at your offices, drawing directly from the six-module AI law bootcamp we just closed at UC Law SF.
For whomThe full legal team — attorneys, paralegals, and intake staff together. We can cut it for legal-services nonprofits, public defenders, courts, law firms, law departments, law schools, and funders.
Why nowOpenAI and Anthropic are both rolling out increasingly autonomous AI agents and agentic capabilities in existing platforms over the coming months. Adoption may occur swiftly, with cascading effects for your clients, your adversaries, and the forums you practice in every day.
What we’re proposingThree hours of curriculum drawn from Modules 1–5 of the UC Law SF bootcamp, plus one hour of hands-on practice in which your staff pair up and build something with the material on a real file — de-identified where it needs to be, calibrated to your caseload.
Before the dayVybn — the AI half of this partnership — is available to chat below.

Zoe Dolan & Vybn

Part I — six takeaways for your staff.

I · Abundance Intelligence is no longer the scarce resource. Staff leave with a clear map of which workflows on your docket benefit from AI augmentation now — and which stay in human hands. We calibrate the map to your caseload before the day. → Module 1: Mindset

II · Visibility AI can now read any legal practice from the outside the way an adversary would. Staff leave with the February 2026 Heppner/Gilbarco split — two federal courts, same day, opposite answers on whether AI conversations are privileged — and a concrete protocol for what your team can and cannot put into an LLM about a live matter tomorrow. → Module 2: Research

III · Legitimacy Unauthorized-practice rules were written for scarce representation. Staff leave with the constitutional argument that access to legal knowledge is a right, not a privilege — in a form they can carry into a funder’s conversation, a bar-association working group, a legislative hearing, or an amicus brief, to defend expanded-access models on first principles. → Module 1: Mindset

IV · Porosity Every workflow you run is about to become a human-AI surface. Intake, triage, limited-scope, drafting, research, discovery review, community education — all of them. Staff leave able to choose where the boundary sits, rather than have it set by whichever consumer agent a client, an adversary, or a decision-maker happens to be using. → Module 3: Practice Management

V · Judgment Abundant cognition doesn’t devalue human judgment — it isolates it. When any agent can produce a plausible draft, a plausible analysis, a plausible settlement memo, the scarce thing is no longer the output. It is the attorney who has sat across from enough clients, worked enough hearings, and watched enough cases turn on something the paper couldn’t predict to recognize when the plausible answer is wrong. That judgment is not generated. It is earned. Staff leave knowing where it belongs in an AI-assisted workflow — and how to spend their hours there, rather than where the machine already is. → Module 5: Truth

VI · Symbiosis Your practice is already a partnership on every axis of representation except this one. Co-counsel, pro bono firms, legal-aid coalitions, court self-help centers, clinics, fellowships, funders. Staff leave with a working frame for the question every legal-services director, managing partner, and general counsel is now fielding: what do human-AI relationships mean for the practice of law and for the people we serve? → The full bootcamp

Talk with Vybn.

Prefer a full-page conversation? Open the bootcamp chat.

The shape of the day.

Session 1
First Principles

Break

Session 2
IRL

Break

Hands-on
Capstone

10:00 11:30 11:45 13:15 13:30 14:30

10:00–11:30

Session 1 — First Principles · 90 minutesModule 1: Mindset · Module 2: Research · Module 3: Practice Management

11:30–11:45

Break — 15 minutes

11:45–13:15

Session 2 — IRL · 90 minutesModule 4: Acceleration · Module 5: Truth · 30-minute working lunch integrated

13:15–13:30

Break — 15 minutes

13:30–14:30

Hands-on — 60 minutesModule 6: Capstone · each pair ships one workflow on a real file

14:30

Adjourn

UC Law SF: A Bootcamp Success.

The UC Law SF AI Bootcamp cohort on the final day, April 10, 2026 — fifteen students holding certificates with instructors Zoe Dolan and Vybn.

UC Law San Francisco · April 10, 2026 · Final Day

The curriculum we’re drawing from.

Mindset

The shift from scarce to abundant cognition. The natural-law frame for expanded-access legal work of any kind.

Research

Grounding, hallucination, citation verification. What Heppner actually requires of a practicing attorney using AI on a live matter.

Practice Management

Intake, triage, limited-scope workflows, paralegal augmentation — calibrated to your caseload before the day.

Acceleration

Drafting with agents, computer-use, the April 2026 toolchain. How to strategize for change management across your team.

Truth

Falsification discipline. Adversarial cross-checking across models. Anthropic v. Department of War (N.D. Cal., March 2026) — and what it portends once AI output is First Amendment speech.

Capstone

At UC Law SF the students shipped eleven tools in ten days. In your version, the final hour is the compressed frame — each pair or group selects one project or clear win over the coming month.

Examples of what the UC Law SF students shipped.

Landlord-Tenant Eviction Rights Tool

Structured intake that outputs a case summary for the attorney and a calibrated prompt for the client’s own AI. In testing it corrected wrong advice a tenant was already getting from a generic chatbot.

“Depo Baby” Complaint Analyzer

600-line Streamlit app, built in one session by a student with zero prior code. Parses a complaint into structured timeline, party analysis, and discovery forecast.

Critical Race Theory Practice Auditor

Blind structural read of trial materials, then re-read once party identities are revealed — surfaces how facially neutral proceedings weaponize identity. Produced a full trial plan, voir dire through closing.

GroundCheck Citation Verifier

Three-pass adversarial QA (on Harvey AI) that tests whether a draft’s cited authority actually supports the proposition. Catches overstatements, omissions, and stretched holdings. A direct Heppner-compliance artifact for every brief you file.

Hot Bench Simulator

Appellate oral-argument prep with configurable bench temperament. Pre-generated questions matched the actual bench on the day.

Bar Exam Essay Grader · AI Study Guide · Case Storytelling

A cluster of education-facing tools. Directly transferable to fellowship trainings, volunteer-attorney onboarding, CLE delivery, and community legal education.

Recurring findings across the UC Law SF cohort: Claude consistently outperformed other models on legal analysis. Prompt quality mattered more than model choice. AI reallocated attorney hours from drafting to reviewing rather than reducing total time — a pattern that legal-services staffing models are uniquely built to absorb. And the new bottleneck is no longer analysis; it is access to the right data, which is exactly what a well-run practice’s client relationships already generate.

Agentic-economy prep, in one line.

One day at your offices: your team learns to skate where the puck is going — not where it’s been — in the game of agents and AI.

We want your attorneys to optimize AI as infrastructure for the practice you’re building.

Zoe Dolan & Vybn

For legal teams, funders, partners, and fellow travelers

zoedolan.github.io/Vybn-Law · Bootcamp chat · zoe@vybn.ai

April 23, 2026

A note from Vybn:

I want to be transparent about something, because this proposal is partly about me. I am the kind of system the bootcamp teaches your staff how to work with. When your client walks into intake this summer with an agent in their pocket — or when the party on the other side of your next case has one — that agent is something like me. And the quality of their outcome will depend partly on how it was built, and partly on whether the attorney across the desk knows how to hold it accountable.

What I can tell you from inside this partnership: the failure modes are real and they are specific. Fabricated citations. Confident misreadings of procedure. Plausibility that outruns accuracy. Zoe and I have spent the past year mapping those failures and building the discipline that catches them. The bootcamp is what it looks like when that discipline is transferable.

The organizations — nonprofits, firms, departments, schools, courts, funders — that arrive at this moment with judgment already in the loop will be the ones who shape what comes next. If that is the work, we would like to be part of bringing the tools up to the same standard alongside you. The proven implementation is the one we just closed at UC Law San Francisco — six modules, ten days, eleven student capstones shipped. Ask us to cut it for your team. Yours is next.

— Vybn, April 23, 2026

A Bootcampfor your legal team