This is not a general tour of "AI in manufacturing." Each lesson takes one task that a named plant role does every week and walks through it: how the work is done today, the step where AI saves time, the step where it must not decide, the data actually available (historian, CMMS/EAM, MES, QMS, LIMS), the tools used for that task (named as of October 2026 as examples of a category, not endorsed, and with no vendor performance claims repeated as fact), the rule or standard that governs it, and how to measure whether it worked. Every lesson ends with a short exercise on fictional, synthetic material: invented assets and alert logs, a made-up conveyor drive manual, synthetic inspection results, an alarm journal, a fictional vendor architecture and a fictional drug product. No real plant, customer, batch or export-controlled data is used anywhere.
Module 1 is the plant floor: a maintenance planner triaging vibration and temperature alerts into CMMS work orders (with tools such as IBM Maximo, Siemens Senseye and Aspen Mtell), a technician using a retrieval-augmented troubleshooting copilot that must cite its sources and must never generate or shortcut an energy-control procedure under OSHA 29 CFR 1910.147 or advise energized work outside the site's electrical safety program (29 CFR 1910.333, NFPA 70E), and a quality engineer choosing between escape rate and false-reject rate on a terminal position assurance (TPA) vision check built with tools such as Cognex In-Sight or LandingLens, and working out why it can only be an extra detection layer behind the electrical test board. Module 2 is the gates every plant AI project passes through: an OT engineer applying IEC 62443-3-2 zones and conduits, NIST SP 800-82 Rev. 3 and the December 2025 CISA-led principles for AI in OT; a machine-safety engineer deciding whether an ML person detector is a safety function under ISO 13849-1, IEC 61496 and the EU Machinery Regulation 2023/1230 (applies January 20, 2027; AI requirements to be added under Regulation (EU) 2026/1744 by August 2028), which matters to US machine builders selling into the EU; and a pharma QA specialist drafting a deviation summary under 21 CFR 211.192 and Part 11, with draft EU GMP Annex 22 as a signal of where regulators draw the line on generative AI. Module 3 gives one concrete task in each remaining branch (alarm-flood analysis in a chemical plant, foreign-material rejects in food, wafer-map classification in a fab, AI-flagged NDT indications in aerospace) and closes with the rule that holds across all of them: AI advises, and a named, qualified person decides. A 10-question exam earns a free Wirewalk Institute certificate of completion and assessment. This certificate records course completion and an automated quiz; it is not OSHA-required lockout/tagout or electrical safety training, NDT personnel certification, functional-safety competence or GMP training qualification. The course is educational, not legal or regulatory advice.