Level 3 taught a practice group to run AI-assisted work as a system. Level 4 is the firm, and the change is not one of scale but of kind. A managing partner is accountable for output nobody senior has read, produced by people they do not supervise directly, using suppliers whose systems change without notice, under rules that differ by jurisdiction and are still being written. This level is about governing that honestly.

The fourteen modules move from accountability to evidence to defence. The first four establish who owes what: why a partnership diffuses responsibility by design and what it takes to place it somewhere specific, how to write a governance programme a partnership could actually adopt, how to keep a decision record that survives the departure of everyone who made the decisions, and how to build and maintain a map of which authorities bind the firm rather than trusting a survey written by somebody else.

Modules five to seven are the supplier relationship: assessing a legal AI supplier and telling an answer from an evasion, reading evidence claims and understanding why "validated" is usually a marketing word, and contracting for the things that actually go wrong, which are model change, sunset, exit, deletion and acquisition. Modules eight to ten cover the firm's exposure: placing AI inside the existing information security programme rather than beside it, defending processes against synthetic media and impersonation, and reconciling AI use with engagement terms, outside counsel guidelines and client consent.

Module eleven teaches a method for the questions with no clean answer, which at this level is most of the interesting ones. Modules twelve to fourteen close the programme: independent verification and audit readiness, leading adoption inside a professional culture that has good reasons to resist, and the implementation roadmap and partnership report that turns all of it into something a room of partners can vote on.

Every lesson follows the same format. It opens with why the topic matters now, poses a question worth sitting with, shows what goes wrong, gives the practical answer as frameworks, checklists, decision tools and worked judgement, and closes with the points to remember, a knowledge check and an exercise. Statements of authority are labelled: law or rule with its jurisdiction, professional guidance with its issuing body and date, best practice, emerging practice, or an AI Coalition Network recommendation. The course never presents one jurisdiction's requirement as universal and never presents a voluntary framework as a duty.

The course ships with a printable workbook, a governance programme template, a decision record, a jurisdictional obligations map, a supplier assessment, a contract clause checklist, a security control map, a scenario analysis worksheet, an audit readiness pack and a partnership report template. The final examination draws forty scenario questions at random from a reviewed bank. The capstone asks for two things: the governance programme for one organisation, and a written defence of one decision inside it where the authority is genuinely unsettled, marked on honesty as much as on conclusion.

Everything here is professional education. It is not legal advice, it does not replace the rules of professional conduct, court rules or other requirements of any jurisdiction, and learners must verify the laws, rules, court requirements and professional guidance that apply to their own practice. Completing the level earns an independent educational certificate issued by AI Coalition Network with a public verification page. It carries no professional education hours or approvals.