Level 1 taught how AI systems work and fail on prints, routers, records and customer messages, and who owns each output. Level 2 puts that to work on the tasks that fill a plant's day. It is written for quality engineers and technicians, process and manufacturing engineers, supervisors who write work instructions, planners, buyers, estimators and inside sales, EHS coordinators and document control. Every module takes one plant task and builds a supervised, repeatable way to do it, starting from released documents at a stated revision and naming, by role, the person who approves the result.
Modules one to three build the instrument. Module one sets the operating model: triaging a task before any tool sees it, running an assisted task from the released documents to an approved record, reading a vendor accuracy claim as a buyer rather than a believer, and keeping the task AI use record while the work happens. Module two builds the layered prompt: a source block carrying the print, router, specification and customer requirement, a parameter block stating part, revision, material, units, tolerance basis and customer, constraints and required cross-checks, and a design that assumes text inside a supplier document may try to act as an instruction. Module three turns shared prompts into a plant library with owners, versions, tests before release, retest triggers such as a new edition, a customer requirement, a print revision or a model change, and deliberate retirement.
Modules four to seven are the core of controlled work. Module four builds the task source pack at a stated revision, screens it before upload, writes closed-file drafting instructions that forbid recall, and hunts invented values and smoothed-over gaps before the reviewer's read. Module five drafts work instructions from released engineering documents in plain language, with a translation a fluent person has checked, safety content checked against the OSHA standards themselves, and visual aids under approval and revision control. Module six drafts nonconformance narratives, containment messages, root-cause write-ups and corrective-action records from evidence, keeps describing what happened apart from deciding what happens to the parts, and covers when a problem leaves the quality system. Module seven covers what certificate packages, inspection and traceability records must show, using AI to list discrepancies but never to complete a record, audit preparation that generates no evidence, and electronic-record controls where a regulated product is involved.
Modules eight to ten face customers and suppliers. Module eight drafts status, lead-time and delay messages the file can support, removes unsupported promises, builds notification duties from the customer requirements actually held, and verifies any instruction that changes where money or parts go. Module nine screens an inbound RFQ package for export-controlled technical data, controlled unclassified information and customer secrets before any file moves, drafts the quote narrative from what remains, and records who screened what and on what basis. Module ten builds substantiation files for origin labels, environmental claims, testimonials and warranty wording, separates federal domestic-content clauses from advertising law, and applies the same test to the claims a vendor makes about its own AI.
Modules eleven to thirteen are the checks that protect the plant. Module eleven is the moment-of-use data decision for customer prints, process parameters, supplier files, controlled unclassified information, employee records and biometric or monitoring data, where notice and consent duties differ by state. Module twelve traces a standard or customer requirement to the body that issues it, confirms the edition in force, reads its scope for this part and this customer, and sorts law from contract, certification scheme and preference in a register the next person can reuse. Module thirteen reviews AI-drafted energy control, guarding, PPE, hazard communication, process safety and hazmat training content against the regulation itself, and keeps environmental reporting and waste decisions away from any estimate a model makes.
Module fourteen closes the level with what happens when something goes wrong: what the AI use record must answer and for whom, the first hour after an AI-related error reaches a part, a record or a customer, the reporting clocks for consumer product hazards and defence cyber incidents, state breach notice and who decides, and the assembly of the Level 2 portfolio.
Statements of authority are labelled with whom each binds and when it was checked; a voluntary consensus standard or certification scheme is never taught as law, and no state's rule is presented as a national one. The level ships with a printable workbook and nine templates, a final examination that draws forty scenario questions from a reviewed bank, and a capstone Assisted Task Portfolio built on fictional plants, parts, customers and suppliers.
Everything here is professional education. It is not legal, engineering, quality, safety, environmental or export-control advice, and it does not replace a released document, the employer's safety programme, a licensed professional engineer where a state requires one, the person your quality system names, an empowered customer representative, legal counsel, or the law that applies to your plant. Completing the level earns an independent educational certificate issued by AI Coalition Network with a public verification page. It is not an engineering licence, a quality-system registration, a safety qualification or a regulatory approval, it carries no professional development hours, and it qualifies no one as an auditor, a competent person or an authorised employee under any standard.