Level 1 taught how AI systems work and fail with financial information, and which duties do not move when a tool arrives. Level 2 is where that becomes practice. It is written for staff and senior accountants, tax preparers and reviewers, auditors, bookkeepers, client accounting staff and CPAs in public practice and industry who draft, research and communicate under supervision. Every module takes one task that fills an accountant's day and builds a supervised, repeatable, defensible way to do it, with the reviewer's step shown explicitly.
The first three modules build the instrument. Module one sets the operating model: triaging a task on consequence, data class, verifiability and who signs before any tool is opened; a seven-stage workflow from source documents to a reviewed deliverable; reading vendor accuracy claims and benchmarks for who ran them and what they tested; and keeping the engagement AI use record while the work happens. Module two builds a layered accounting prompt with an engagement header, labelled authority and figure packets, an output contract with required tie-outs, and controls that limit the damage when client documents contain instruction-like text. Module three turns shared prompts into a firm library with owners, versions, release tests, retest triggers and deliberate retirement.
Modules four to six are production work. Module four drafts client letters, engagement letters and technical memos from a verified fact file in which every fact carries a status, then catches invented figures, upgraded representations and smoothed-over gaps before the reviewer's read. Module five reviews leases, loans, invoices and contracts against a checklist with quotes and locations, flags treatment questions without answering them, routes low-confidence and conflicting extractions to the page, and applies the amended PCAOB evidence requirements for electronic information. Module six covers what an AI-assisted workpaper must show, version discipline when drafts can be regenerated, the reviewer sign-off record, and retention and secure disposal.
Modules seven to nine are research and tax practice. Module seven follows a retrieval-grounded answer from question to citation and tests what a linked citation proves, claim by claim. Module eight is a six-step authority verification protocol: existence, pinpoint, scope, status, secondary summaries and the record. Module nine applies return position standards and preparer penalties, written advice under Circular 230, due diligence and competence as the IRS Office of Professional Responsibility's 2026 alert reads them, and reasonable reliance on staff, tools and client representations.
Modules ten to twelve turn to the client: status updates the file can support, removing unverified projections, when and how to tell a client that AI was used, and describing the firm's technology truthfully; intake, engagement scope, the consents that must come first and the independence screen for attest clients; and billing only time actually spent, crediting cost reductions, reviewing fixed fees and measuring savings from the firm's own records. Modules thirteen and fourteen close the loop: classifying engagement material, the moment-of-use decision for tax return information under the section 7216 regulations, disclosure outside the United States, third-party provider duties, and an engagement AI use record designed to scope an incident, with the first hour, a notification map and a practice review.
Statements of authority are labelled throughout as law or rule, professional standard, professional guidance, best practice, emerging practice, or an AI Coalition Network recommendation, with whom each binds. Licensing is by state, and the course never presents one state's rule, a professional body's code or a voluntary framework as binding on everyone. The level ships with a printable workbook and templates: a prompt pattern library, the authority verification protocol, a source document review checklist, an engagement AI use record, a client communication pack, a fees and AI disclosure worksheet, a data classification quick card and the capstone portfolio template. The final examination draws forty scenario questions at random from a reviewed bank, and the capstone is an Engagement Workflow Portfolio of three AI-assisted workflows run on fictional engagements.
Everything here is professional education. It is not tax, accounting, audit or legal advice, it does not replace the professional standards, IRS rules or state board requirements that govern accounting, tax and attest work, and learners must check the rules that apply to their own licence and practice. Completing the level earns an independent educational certificate issued by AI Coalition Network with a public verification page. It is not a CPA licence, carries no professional education hours, and does not satisfy any state board of accountancy requirement.