Build an AI-assisted legal workflow you can explain, test and defend. Start with the client’s problem, compare what the models produce, and take responsibility for the advice.
Harvey, Legora, CoCounsel, Copilot — every one of them promises to transform legal work. Some of it is real. Much of it does not survive contact with a regulated practice, a privileged document, or a partner's standard of care. The cost of guessing wrong is measured in wasted licences, exposed clients and eroded trust.
This masterclass gives legal leaders an evidence-based way to decide: what these systems actually do, where they fail, how to govern them, and how to tell a genuine capability from a demo. No vendor script. No breathless futurism. The same “does it deliver?” discipline applied to the tools reshaping the profession.
In teaching for Erasmus School of Law master’s students, JJ used a financing case to put AI-assisted legal work under pressure. Groups developed advice, compared findings from different models and vendors, and presented their reasoning for challenge.
The assignment paired a term sheet with a client file. Each student worked with a model; another group member checked the work using a different model. Teams kept a prompt log, investigated unsupported claims and omissions, and explained what a lawyer would need to verify before signing the advice.
The lesson is practical: decide what the model needs to see, test what it says against the documents, and know what remains your judgment.
General Counsel and heads of legal, law-firm partners and senior associates, and legal operations leaders. Adapted to the room, from master’s students learning the method to experienced practitioners deciding how their teams should use AI.
A focused, practitioner-register session — in-house for a firm or legal department, or for a professional body. Length and depth tailored to the room, from a half-day intensive to a full programme.
Tell me who is in the room and what you want them to walk out able to do. I will come back with a proposed shape and dates.