Wirewalk Instituteof AI
mfg-235 Intermediate Available

Building AI into Quality Management: SPC, CAPA, ISO 9001 and IATF 16949

For US quality engineers, customer and warranty quality engineers and QMS owners at manufacturers: you leave able to build, test and govern AI for specific quality jobs while a named person stays the decider of record.

Level
Intermediate
Study time
16 hours
Lessons
18 in 6 modules
Exam
20 questions

About this course

Kittredge Precision Components, a fictional composite supplier with plants in Indiana and Ohio, has just been put on Controlled Shipping Level 1 by its largest customer for cracked seat-heater controller housings. The CEO has promised an "AI-enabled quality program": exit controlled shipping by December 31, cut the plant's ppm with that customer from 38 to under 15 by June, and take $1.5 million out of the cost of poor quality, all before an IATF 16949 recertification audit in March. The escape itself has already put the ppm promise out of reach as stated, and the evidence will support less than a third of the savings; the course is about finding that out early, proving what is real and saying so. The course follows the quality engineers, the warranty analyst, the MSA engineer, the QMS administrator and the internal auditor through that program, one business problem per module.

Each module compares the approaches practitioners actually choose between in October 2026: classical run rules against Hotelling T-squared and EWMA on a replay; Microsoft 365 Copilot over an 8D library against an enterprise project workspace and the eQMS vendor's built-in summarizer; an exact-match script against an LLM reviewer for PFMEA linkage; keyword rules, an LLM and an embedding classifier for warranty narratives; vendor validation reports against an attribute agreement study on the plant's own parts, with a human review band for the borderline cracks the station misses. Each module shows a realistic failure (limits recomputed around a drift, a root cause copied from a similar 8D, a post-inspection field leaking into a model, a burning-smell narrative routed on model confidence, an AI summary with the wrong quantity in an audited record) and how it was caught. Standards are taught by edition and in the course's own words: ISO 9001:2026 was published in September 2026 while IATF 16949:2016, built on ISO 9001:2015, remains the automotive scheme until its planned Revision 2. The US reporting clocks that warranty AI must never delay are quoted from the regulation text (49 CFR 573.6 and 573.3; 15 U.S.C. 2064(b) and 16 CFR 1115.14). Vendor products are named to show the categories in use, never as endorsements, and vendor outcome figures are not repeated. Pharma GMP CAPA is covered in mfg-170 and mfg-270, PPAP depth in mfg-220 and building vision inspection in mfg-215. Every module ends in a Decision Lab, a branching scenario with business metrics, a cliffhanger and several endings with an expert debrief, and the course closes with a strategic capstone simulation and an AI-in-QMS procedure you score against a rubric as practice. Educational content, not legal advice.

What you will be able to do

  • Evaluate an SPC monitoring layer on a replay, comparing run rules with a multivariate or drift detector on detection delay and false alarms, routing each signal to the reaction plan without moving control limits, and auditing an automatically chosen capability distribution
  • Build and evaluate a retrieval-grounded 8D draft in which every number traces to a record, D4 explains both occurrence and escape and how each was verified, and a similar past 8D is not copied
  • Analyze PFMEA, control-plan and work-instruction linkage with a script and an LLM, reject unsupported rating changes, and decide whether a predictive-quality model may enter the control plan
  • Evaluate a warranty-narrative classifier against expert codes and place a deterministic safety bypass ahead of any model so that product-safety decisions and US reporting clocks never wait on a model
  • Analyze requirements documents with AI by diffing customer-specific requirements revisions against a gold list and checking audit findings for edition-correct ISO 9001 and IATF 16949 citations
  • Validate vendor AI for a stated intended use with thresholds fixed before testing, decide go or no-go with re-test triggers, and qualify an AI vision station on plant parts, including borderline ones, before it is credited as a detection control
  • Evaluate a quality AI business case and a management-review narrative by separating booked cost-of-poor-quality reductions from avoided temporary costs and freed hours, net of run costs, and by reading monthly and rolling ppm correctly
  • Create an AI-in-QMS procedure and a release decision for one quality AI use, and test the procedure against scenarios from the course

Syllabus

6 modules · 18 lessons · about 16 hours

Module 1: Two Hundred Fourteen Charts: An SPC Layer That Routes Every Signal

A process quality engineer replays six months of a machining cell to choose a detector, keeps control limits a controlled decision, routes signals to the reaction plan and audits an automatically chosen distribution.
3 lessons
  1. Replaying Cell 7: Run Rules, Hotelling T-Squared and EWMA Compared on Detection Delay and False Alarms
  2. Drift, Tool Wear and Limits That Do Not Move: Routing Each Signal to the Reaction Plan
  3. The Ppk That Looked Too Good: Auditing an Automatically Chosen Distribution on a Two-Cavity Flatness Study
  • Decision Lab · 12 minTwo Hundred Fourteen ChartsCut a plant's SPC alarm load without hiding the next oversize bore: choose whether to move limits, which detector to trust, and how a signal reaches a person.

Module 2: The Cracked Housing: An 8D Grounded in Evidence, Not in Last Year's 8D

A customer quality engineer under controlled shipping builds the complaint pack, chooses the assistant setup the data allows, drafts D2 to D5 with retrieval and scores the draft before the customer sees it.
3 lessons
  1. Where the Evidence Lives: Building the Complaint Pack and Choosing Copilot, a Project Workspace or the eQMS Assistant
  2. Drafting D2 to D5 with Retrieval: Is/Is-Not, Occurrence and Escape, and the Past 8D That Must Not Be Copied
  3. Scoring the Draft Before Arden Sees It: Ten Checks for Invented Numbers, Copied Causes and Read-Across
  • Decision Lab · 13 minControlled ShippingDraft, check and submit an 8D under controlled shipping, choosing which tool may see the evidence, whether to trust a precedent, and how far the fix must reach.

Module 3: Living PFMEAs and Predictions That Want to Be Controls

An APQP engineer measures linkage checkers on seeded gaps, keeps ratings with the team, and decides how a predictive-quality model may be used.
3 lessons
  1. Finding Linkage Gaps Between the PFMEA, Control Plan and Work Instruction: Script, LLM or Both
  2. When the Assistant Suggests a Lower Occurrence Rating: Keeping S, O, D and Action Priority with the Team
  3. A Predictive-Quality Model for Short Shots: Label Leakage, an Inspection-Budget Threshold and the Control-Plan Wording
  • Decision Lab · 12 minThe Model That Wanted to Release PartsDecide what a predictive-quality model for short shots may do: which fields it sees, whether it releases parts or flags them, what happens to the PFMEA, and how it survives a mold change.

Module 4: Field Signals, Reporting Clocks and Customer Requirements

A warranty analyst codes field narratives behind a safety bypass and finds an emerging lot cluster, and a customer quality engineer diffs a revised supplier quality manual.
3 lessons
  1. Coding Four Hundred Warranty Narratives: Keyword Rules, an LLM and an Embedding Classifier Compared on Kappa
  2. The Safety Bypass: Fire, Injury and Loss-of-Control Narratives Reach a Person Before Any Model Decides
  3. Revision G of Arden's Supplier Quality Manual: Extracting and Diffing Requirements Without Missing the Ones in Tables
  • Decision Lab · 14 minWeek SixSpeed up warranty coding without letting a model stand between a burning-smell narrative and the people who decide, then escalate an emerging lot cluster.

Module 5: Proving Vendor AI Works: The eQMS Summarizer and the Vision Station

A QMS administrator tests a default-on eQMS summarizer for a stated use and reaches a no-go, and an MSA engineer decides what an AI vision station, with a human review band for borderline cracks, can be credited for.
3 lessons
  1. Release 26.3 Turned On an AI Summarizer: Intended Use, a Fixed Test Set and Thresholds Set in Advance
  2. Go, No-Go and the Next Vendor Release: Scoring Summaries and Writing the Re-Test Trigger
  3. Can the Vision Station Replace the Visual Check? Attribute Agreement, Borderline Cracks and Re-Qualification Triggers
  • Decision Lab · 12 minThe Vision Station DemoDecide what an AI vision station can be credited for on a special characteristic: whose evidence counts, what to do about borderline cracks it accepts, and what happens when the vendor updates the model overnight.

Module 6: Audit, Management Review and the AI-in-QMS Procedure

An internal auditor checks AI-drafted findings for the right edition, a quality manager ties an AI narrative to the numbers and the arithmetic of a ppm promise, and the learner writes and tests the procedure that governs every use in the course.
3 lessons
  1. Planning the Lima Process Audit with AI Without Mixing ISO 9001:2015, ISO 9001:2026 and IATF 16949:2016
  2. The Management-Review Narrative That Contradicted Its Own Table: Checking AI Against Cost of Poor Quality
  3. Capstone: The AI-in-QMS Procedure and the Release Decision for One Quality Use
  • Decision Lab · 13 minFive Days Before the AuditorIn the week before recertification, decide how to issue AI-drafted audit findings, what management review will say about cost and ppm, and how to answer an auditor who finds AI text in a CAPA record.
  • Capstone simulation · 35 minClearline: From Controlled Shipping to RecertificationRun a supplier's AI-enabled quality program from controlled shipping to recertification, with three promises already made, a fixed budget, a CEO who wants good news and a customer reading every number.

Exam and certificate

The exam unlocks once you have completed every lesson and played each Decision Lab to an ending. It draws 20 questions from a larger bank, spread across the modules, with the answer options shuffled for each attempt. You have 40 minutes and need 80% to pass. After a failed attempt you can try again after 24 hours, up to 3 attempts in total. Feedback is given per module, never as the answer key.

Passing earns a certificate of completion and assessment from the Wirewalk Institute of AI with a unique ID that anyone can check at institute.wirewalk.com/verify. It is not an accredited qualification or a vendor certification. How certification works.

The Wirewalk Institute of AI is independent. This course teaches publicly documented products and is not affiliated with or endorsed by Advantive, Minitab, Siemens, Octave, MasterControl, Veeva, Microsoft, Anthropic, or OpenAI. Product and company names are trademarks of their respective owners.

More in AI in Manufacturing and Industrial Operations

All 16 courses in this track
mfg-001 Free intro Available

AI in Manufacturing and Industrial Operations: The One-Hour Briefing

A free hour built around real plant jobs: a maintenance planner turning overnight vibration alerts into work orders, a technician diagnosing a drive fault with a manual-citing copilot without touching lockout/tagout, a quality engineer setting the threshold on an AI terminal-lock vision check, an OT engineer deciding read-only versus write-back for a vendor's cloud link, a machine-safety engineer classifying ML person detection, and a pharma QA specialist drafting a deviation summary under human review.

1 h · 8 lessons Free intro
mfg-110 Basic Available

AI for Manufacturing and Plant Professionals

Practice with a fictional Ohio components plant using complete inline records. Check AI summaries, quality-review metrics, maintenance evidence, schedule feasibility and controlled-document claims before recommending a bounded next step.

8 h · 18 lessons See price at checkout
mfg-210 Intermediate Available

Building and Deploying AI in Industrial Operations

Evaluate plant AI through a cumulative synthetic US food-manufacturing case: data lineage, alert-to-work-order evidence, causal soft sensors, controlled-document RAG, control-model tests and read-only gateways. Assemble a reviewable decision dossier; no production authority or safety qualification is granted.

22 h · 23 lessons See price at checkout