The top of the pathway is not about using AI better. It is about teaching other small business owners to use it well, in a chamber workshop, a franchise meeting, a college evening class or a coaching session. The owners in that room have no compliance function behind them, and the instructor may be the only person who ever gives them the caveat. An instructor who teaches a rule wrongly makes the error once for every owner who heard it.
The first three modules set the foundations. Module one describes who is in the room from what the surveys establish rather than from assertion, designs for one action an owner takes on Monday instead of for coverage, and explains why every teaching recommendation in this course is practice, never a borrowed research finding. Module two designs a session backwards from the behaviour, such as checking a quote against the rate sheet, with objectives an observer could verify, a timetable that survives a late start, and a checking habit that holds when the same person does well with AI on one task and badly on the next. Module three keeps material current when rules are narrowed, delayed, vacated, set aside, withdrawn or re-enacted: four corrections every instructor must be able to explain, a currency log naming each claim, its source, its last check and its checker, and a written correction to the people who heard the old version.
Modules four to seven are delivery. Module four facilitates a room of mixed trades, where a landscaper, a dental practice manager and an online seller need different answers: surfacing the condition that decides it, three traps that catch a whole room, naming the licensing board and stopping, and the owner who argues, the one who is lost and the one who already bought. Module five scripts four live failures on fictional data, an invented rule, a stale price, an arithmetic slip and instruction-like text in a forwarded enquiry, and debriefs for the checking habit rather than distrust. Module six builds a fictional business the whole room can work on, keeps real customer, employee and payment data out of the room, and runs the lab when learners use different tools or none. Module seven teaches verification without fearmongering: what the official record shows when a learner asks for the scary example, how the FBI describes generative AI in impersonation and payment fraud without invented numbers, and why no reported case is not proof of safety.
Modules eight to ten are the content hardest to teach accurately. Module eight teaches the size gate: federal employment thresholds as a ladder that starts at one employee, privacy tests built on revenue and volume, partial exceptions and scaled duties that are not exemptions, and a biometric statute with no gate at all. Module nine teaches what differs by state, city and trade: three state privacy tests built differently, recording consent and monitoring notice, two states' different approaches to AI, and the official licence directories. Module ten teaches the evidence honestly: two credible surveys that report different adoption figures because they asked different questions, what a Census figure can carry, and labelling every productivity study by what it measured before quoting it.
Modules eleven to thirteen cover assessment, support and access. Module eleven writes scenario items whose wrong answers are real mistakes, tasks that reward asking about size, state, trade and data first, rubrics scored the same across cohorts, and fair, recorded retake and integrity decisions. Module twelve diagnoses why a learner is stuck, supports learners between sessions at low cost, and records resistance that is sound judgement. Module thirteen covers accessible handouts, captions and plain language, private accommodations in a room of eight, exactly whom accessibility duties bind, and why an instructor never tells a room that a tool has made anything compliant.
Module fourteen sets the limits of the role: saying "I do not know" and then finding out, declining a particular business's legal, tax, employment or licensing question with a referral that names the right holder, disclosing vendor ties and teaching rather than selling, and the claims never made about a course. It prepares the Teach-Back, the capstone: one teaching segment for a fictional audience on fictional data, checked claim by claim against its sources, delivered and reflected on from evidence of what learners did. The level ships with a printable workbook and fourteen templates, and the examination draws forty scenario questions from a reviewed bank.
Everything here is professional education. It is not legal, tax, accounting, employment, privacy or licensing advice, and it does not replace an attorney, a tax professional, a licensing board or the law of the state where an owner does business. Completing the level earns an independent educational certificate issued by AI Coalition Network with a public verification page. It is not a teaching credential or a training-provider registration, it carries no professional education hours, and a session taught by a holder earns none, satisfies no licensing requirement and gives no one authority to teach a licensed trade.