"Just write Not applicable."
A customer asks where you use AI. Someone senior says you don't. If a record you sent them says otherwise, every other answer on that questionnaire loses its value.
AI for Quality Teams · C14 · Self-paced
Your team already uses AI.Can you show how it's controlled?
Chat apps drafting NCR wording. A certificate reader on trial at receiving. An AI feature a software update switched on. Learn to find every use, set rules people will follow, approve tools on evidence and check a use before it touches a record.
Build the six records that let you answer "where do you use AI, and how is it controlled?" truthfully, and keep the answer true when a tool changes.
- About 4 to 5 hours planned
- Six modules + practical labs
- No IT background needed
- Fictional records only
- Human-reviewed case
Try the course first
Find the uses nobody told you about.
Open the real first lesson with Coach Q. Learn where AI hides in a quality system, how to ask so people actually answer, and why you sort uses by where the output lands.
Enter your name, work email and phone, then verify your email to open lesson 1. No payment card is required.
Every record in the course is fictional. You do not need to upload employer data to evaluate it.
Sound familiar?
AI arrived before anyone decided anything.
None of it is a scandal. It is just not controlled yet, and the records already show it. The risk is not that people use AI. It is that nobody can say where, on what terms, or who checked the result.
"I use my own account. Everyone does."
Company notes, part numbers and sometimes a drawing callout go into a personal chat account the company does not administer, cannot see and cannot get back.
"It came with the update."
A software vendor switches on a feature that suggests dispositions or calibration intervals. Nobody bought it, nobody decided on it, and people are already relying on it.
"Rafael read it before he filed it."
An AI-drafted NCR blames an operator the records clear and names a root cause nobody has found. Reading for sense is not checking each statement against the source.
See the method · fictional teaching example
95% right. Still not approved to work alone.
At fictional Torvane Precision, the receiving inspector wants a certificate reader to fill the receiving log. The heat number in that log is how every part traces back to its material. The criteria were written and signed before the test.
Illustrative practice case · not a measurement of any real product
40 certificates, 160 field checks.
Fields right · 152 of 16095%152 ÷ 160 × 100
Wrong and shown as if right · 3 of 1601.875%3 ÷ 160 × 100
The criteria for filling the log unattended were at least 99% and no silent errors. It fails both. Two of the three silent errors were heat numbers on certificates that list more than one heat, where only 12 of 16 checks were right.
Open the decision
Unattended entry: declined. 95% against 99%, and three silent errors against none.
As a draft the inspector checks field by field: approved with conditions. In the test the check caught all eight wrong values, and draft plus check averaged 4.5 minutes a certificate against 6.0 by hand. Test certificates were chosen in advance, so live time gets measured in the first month. Multi-heat certificates are keyed by hand, every entry carries an AI assistance line, ten are sampled a month, and any vendor model change means testing again.
A different decision can be sound if the records support it. The course assesses the reasoning, not how much AI you approve.
From mess to control
Six steps you can run in your own system.
Find → set the rules → approve tools → check before use → keep the records → review and change. Each step produces a record you can show, and each points to a process you already run instead of building a second system.
- 01 Find
- 02 Rule
- 03 Approve
- 04 Check
- 05 Record
- 06 Review
One line on the record.
"AI-generated" tells an auditor nothing. A traceable line names the tool and registered use, the sources each statement was checked against, who checked and when.
It is the smallest control in the course, and the one that makes every other record easy to follow.
The specimen is fictional. Try adapting it here; edits are not saved or sent.
What you bring back
Six records. One controlled system.
Build them in the labs on a fictional shop, then adapt the worksheets to your own system with information your company allows you to use.
AI use inventory
Before you write a single rule. Every use as a task, a tool and a person: the account type, what goes in under your classification rule, where the output lands and who checks it. Includes the features a vendor switched on and the personal accounts nobody mentioned.
One-page AI use procedure
Before you answer "what are the rules?". Scope, four named roles, a data rule that defers to your own classification, permitted, conditional and prohibited lists, and links to the purchasing, document control, competence and corrective action processes you already run.
AI tool register
Before anyone relies on a tool. The plan you hold, what its terms say, who read them and when, required settings, the data each tool may receive, the named uses it is approved for, and its status: approved, on hold, not approved or prohibited.
Pre-use check record
Before a use touches a record. The intended use, how bad a wrong output would be and how likely it is to be caught, a known-answer test with criteria written first, results with denominators, silent versus flagged errors, the decision, conditions and retest triggers.
Audit evidence pack and AI assistance line
When a customer or auditor asks. An index of which record answers which follow-up question, and one traceable line for every AI-assisted record: the tool and registered use, the sources checked, who checked and when.
Review and change record
When a tool changes. Corrections at the check and escapes found in sampling, each with its denominator, the change that triggered the review, what the figures can and cannot prove, the decision and owner, and the summary for management review.
Is this the right next course?
For the person who keeps the quality system.
A good fit if you…
- Own or maintain an ISO 9001, AS9100, ISO 13485 or similar quality system.
- Know people are using AI but can't yet say where or on what terms.
- Need to answer a customer, auditor or your own management truthfully.
- Want rules and records that fit the system you already run.
No IT or coding background is required. Supplied fictional records let you practise without uploading company data.
Bring specialists in when you need…
- Export-control, privacy or legal determinations about your data.
- Technical setup, integration or security review of a tool.
- EU AI Act obligations, if you place or use AI systems in the EU.
- ISO/IEC 42001 implementation or certification.
The course teaches you to mark these as unconfirmed and route them to the right owner. It doesn't make those calls for you.
Inside the course
The course, step by step.
About 4.5 hours, self-paced, in short sessions that fit your week.
Find every AI use
Find and classify every AI use that touches the quality system. 2 lessons.
Write the AI use procedure
Write a short AI use procedure that fits the existing quality system. 2 lessons.
Approve the tool, not the demo
Approve AI tools on evidence, for named uses, through purchasing control. 2 lessons.
Check a use before it goes live
Check an AI use before it goes live, with evidence in proportion to the risk. 2 lessons.
Keep records that answer an auditor's questions
Keep honest records that show how AI use is controlled. 2 lessons.
Review, change and retire
Keep control of AI uses as tools, people and tasks change. 2 lessons.
Inside the course
Six modules that build one control system.
Each module pairs a guided lesson with a hands-on lab, a work product and a coached reflection. Then you apply the method to an unseen case in a different industry.
01 / 06
Find every AI use
Your team says they don't use AI. Have you checked the software?
Look in three places: people, the software you already run, and purchases and accounts. Sort every use by where its output lands, and say plainly when the inventory is incomplete.
- Practise: Talk a worried engineer through what he has been doing, sort twelve real-world uses and write the interim note that keeps people telling you.
- Take forward: Build the AI use inventory.
02 / 06
Write the AI use procedure
A ban nobody follows is not control.
ISO 9001 does not mention AI, so decide how your existing requirements apply and write it down on one page. Three lists do most of the work.
- Practise: Check a policy a chat app wrote against what the company actually does, talk a general manager out of a ban, and write a prohibited list people will accept.
- Take forward: Build the procedure outline.
03 / 06
Approve the tool, not the demo
"Validated for ISO 9001" is a sales claim, not evidence.
Treat a tool whose output lands in your records like any provider that affects conformity. Read the terms for the plan you hold and approve named uses only.
- Practise: Sort brochure claims into contract or test, handle a pushy sales call and write the approval scope for a company assistant.
- Take forward: Build the AI tool register.
04 / 06
Check a use before it goes live
95% sounds good. Is it good enough for this?
Write the intended use precisely, match the evidence to the risk and test on answers you already know, with criteria signed before anyone sees a result.
- Practise: Work out a certificate reader test from its counts, separate silent from flagged errors, and decide two uses from one test.
- Take forward: Build the pre-use check record.
05 / 06
Keep records that answer an auditor's questions
The first question is easy. The next five are not.
Keep records that answer each follow-up, write an AI assistance line that can be traced, correct old records openly and never backfill.
- Practise: Run the sample before the auditor does, handle a gap in front of them and update a customer questionnaire truthfully.
- Take forward: Build the audit evidence pack index.
06 / 06
Review, change and retire
Approved in November. Different in February.
Watch corrections at the check and escapes past it, treat a vendor update as a change, retest before relying on a use again and retire uses cleanly.
- Practise: Pool monitoring figures correctly, deal with an update notice nobody routed and decide whether an approved use continues.
- Take forward: Build the review and change record; put it beside your first answer.
The learning experience
Make the calls on records before you make them at work.
01 · Practise
Decide, then see what happens.
Work one fictional machine shop's real mess: a survey, software settings, a vendor brochure, a test result, an audit sample and an update notice. Branching conversations with the engineer, the general manager, the salesperson and the auditor show the consequences of each answer. Seven short narrated videos with captions and transcripts and ten photographs support the lessons.
02 · Reflect
Connect each module to your own system.
A coached reflection closes every module. Six worksheets and six Course Hub document tools let you build your own inventory, procedure outline, tool register, pre-use check, evidence index and review record.
03 · Apply independently
Show the method works on a fresh case.
One assigned case from three fictional industries (medical device, food-contact packaging, aerospace distribution), six decisions and a separate 30-question exam. A person reviews your case.
Practice support with a clear boundary.
Coach Q is QMS Learning's AI quality and compliance colleague. In practice it points at the record or the gap, asks one question and gives hints that get more specific. It won't tell you a tool is safe, approved or validated. You check its suggestions and own what you use.
Coach Q's default allowance is 40 messages per rolling 24 hours. During the independent case and the exam, support is limited to navigation; content hints and Hub document drafting pause so the assessed judgment stays yours.
Illustrative practice exchange
"The vendor says it's 99.2% accurate and validated. Can I approve it?"
"Measured on whose certificates? What exactly would you approve it to do, and who checks the output before it lands in the log? What criteria would you write down before testing it on your own past receipts?"
AI suggestions can be wrong. Check sources, permissions and the limits of your authority before using an output.
Course Hub keeps your study plan, focus timer, progress, Documents and Apply at work tracker together. Record an authorized workplace application, the checks you made and the follow-up.
Plan about 4 to 5 hours, self-paced. The 260-minute planning estimate includes an 8-minute first attempt, six 27-minute modules, a 40-minute case, 20 minutes of decisions and a 30-minute exam. Allow extra time for practice and revision. This is a planning estimate, not a measured completion time.
Why pay for structured practice?
A policy template is easy to find. Control is harder.
Free templates and a chat app can give you a policy in a minute. The trouble is that a policy promising committees, certifications and validation you don't do becomes a list of nonconformities against your own document.
What the course is designed to add
- Decisions on supplied records, where the obvious answer is often wrong.
- Arithmetic with denominators: silent versus flagged errors, pooled versus averaged rates.
- Six linked records that fit the processes you already run.
- An unseen case reviewed by a person, and a separate exam.
What makes it worth considering
If you'll soon have to answer a customer or auditor about AI, decide on a tool someone wants, or explain why a vendor update matters, this gives you a place to make those calls and get them checked first.
The value is the practice and the assessment. It doesn't guarantee audit outcomes, customer acceptance or any saving.
If you want primary background first, NIST's AI Risk Management Framework is free. It's a general framework, not a quality system procedure, and it doesn't validate this course.
What informs the teaching
Built on requirements you already have.
The course doesn't invent new requirements. It shows how the ones in your quality system already apply when a process uses AI, and labels its own methods as methods. Every organization, person, tool and figure in it is fictional.
- ISO 9001: purchasing, documented information, competence, validation, audit, review and corrective action. The standard does not mention AI; how its requirements apply is your organization's decision. The course cites 2015 clause numbers; confirm numbering in your 2026 copy.
- FDA Quality Management System Regulation and ISO 13485:2016 clause 4.1.6 on validating software used in the quality system, used for the medical device case.
- FDA, Computer Software Assurance for Production and Quality System Software (final, September 2025): an example of risk-based, intended-use thinking.
- NIST AI Risk Management Framework: general background on governing and managing AI risk.
Sources inform the teaching. They don't endorse the course, and the course gives no legal, export-control or regulatory advice.
How you earn assessed completion
Earn the record by showing your reasoning.
- Complete the required labs and coached module reflections.
- Pass the human-reviewed independent case at 16/20 and decisions at 10/12, with no unresolved critical error.
- Pass a 30-question final exam at 24/30 (80%). You have three exam attempts, with a 24-hour wait after an unsuccessful attempt. Three versions support reassessment; you don't need to pass three exams.
Human review scores how you find and classify uses, set rules and approve tools, read a pre-use test, answer truthfully from the records and respond to a change. Approving less AI doesn't score higher; a decision supported by the records does.
Contact will@qmslearning.com about case review or revision. No fixed review turnaround is promised.
The verifiable certificate records assessed course completion. It isn't accreditation, ISO/IEC 42001 certification, approval of any tool or proof of competence in every workplace.
The itemized offer
Everything included, before the price.
Structured practice, six working records, coaching and independent assessment. No invented component values, former prices or savings.
Your complete course package
Six-module course and practical labs
IncludedGuided lessons, branching conversations, data decisions and coached reflections across six original modules. Seven synthetic-narrated videos with captions and transcripts, and ten photographs.
Six worksheets and six document tools
IncludedAI use inventory, procedure outline, tool register, pre-use check record, audit evidence pack index and review and change record. Adapt each before workplace use.
Coach Q for practice
IncludedCourse-aware practice support, with a default 40 messages per rolling 24 hours. Navigation only during the independent case and exam; Hub drafting pauses then.
Human-reviewed independent case
IncludedOne assigned case from three fictional industries, reviewed by a person against a five-part rubric. Six decisions are assessed separately.
Final exam and assessed completion
IncludedA 30-question exam with a 24/30 pass mark; three versions support three allowed attempts, with a 24-hour wait after an unsuccessful attempt.
Course Hub, study and application tools
IncludedStudy plan, focus timer, saved progress and the Apply at work tracker. Record real checks and follow-up on an authorized application.
Complete course · US dollars · per named learner
$129 one-time course price
Lifetime lesson access with no course subscription. Coach Q and AI tools follow the accepted usage and service terms; this isn't unlimited coaching or a promise of perpetual AI service. Applicable tax is shown at checkout.
People
Enroll for $1295 or more people get 10% off every seat. One-time purchase. 3-day refund under Academy policy.
Each seat belongs to one named learner. Choose 1 to 50 people. Five seats total $580.50 before applicable tax after the 10% seat discount.
Make an informed decision
Bring the questionnaire to your manager.
Before purchase
Try the first lesson.
Use the real guided lesson with Coach Q and decide whether the exercises fit your system. Start lesson 1 free ↑
For an individual Academy course purchase, request a refund within 3 days, subject to the Order. A refund ends access and invalidates a completion certificate. Read the full refund policy.
Lifetime lesson access, named seats and AI usage follow the published terms and the terms accepted at checkout. Coach Q has a default 40-message allowance per rolling 24 hours.
If your company pays
Use a concrete training request.
Explain the question you need to answer and the records you intend to build. Request a quote or PO arrangement. This opens your email application; nothing is sent by this page.
Straight answers
What to know before you commit.
Who is this for?
Quality managers, QMS coordinators, quality engineers and owners who keep an ISO 9001, AS9100, ISO 13485 or similar quality system and need to bring AI use under control.
Do I need an IT background?
No. The course is about decisions and records, not configuration. Account setup, integration and security review belong with your IT or tool owner, and the course shows you how to split that work.
Does ISO 9001 require an AI policy?
ISO 9001 doesn't mention AI. Its existing requirements, such as purchasing, documented information, competence and validation, still apply to processes that use AI. The course shows how to decide and record how they apply in your system.
Will this prepare me for ISO/IEC 42001?
No. ISO/IEC 42001 is a separate AI management system standard. This course works inside your existing quality system and makes no 42001 certification claim.
Does it cover the EU AI Act?
No. If you place or use AI systems in the EU, get qualified legal advice. The course doesn't give legal advice.
Do I need another AI subscription?
No separate AI subscription is needed for the included Coach Q and Course Hub tools. Coach Q has a default allowance of 40 messages per rolling 24 hours, and usage follows the accepted service terms.
Can I complete it without company data?
Yes. Every record in the course is fictional. At work, use only information your company authorizes, in tools it has approved.
Can Coach Q do the assessment for me?
No. During the independent case and exam, support is navigation only, and content hints and Hub drafting pause.
What do I need to pass?
Complete the required labs and coached reflections; score at least 16/20 on the human-reviewed case and 10/12 on decisions with no unresolved critical error; and pass a 30-question exam at 24/30 (80%).
What if I do not pass first time?
You have three exam attempts, with a 24-hour wait after an unsuccessful one. Case review feedback and any revision are handled through the reviewer; contact will@qmslearning.com with questions.
How long does it take?
Plan about 4 to 5 hours, self-paced: an 8-minute first attempt, six 27-minute modules, a 40-minute case, 20 minutes of decisions and a 30-minute exam. This is a planning estimate.
How long do I keep access?
A one-time purchase gives lifetime lesson access without a course subscription, under the accepted terms. Coach Q and AI tools stay subject to usage and service terms.
What is the refund policy?
For an individual Academy purchase, request a refund within 3 days; the Order controls if it says otherwise. Email will@qmslearning.com with "Refund Request" and your checkout email. A refund ends access and invalidates any certificate for the course.
Is the certificate an accredited qualification?
No. It records assessed course completion. It isn't accredited personnel certification, ISO/IEC 42001 certification or approval of any tool.
Can my company buy seats or use a PO?
Yes. Choose 1 to 50 named learners. Five or more seats get 10% off each seat; five seats total US$580.50 before tax. Contact will@qmslearning.com for a quote or PO.
Can I try a free lesson?
Yes. Enter your name, work email and phone, then verify your email to open lesson 1 with Coach Q. No payment card is needed.
Answer the AI question with evidence
Find it. Rule it. Check it. Keep it true.
Bring AI inside your quality system on purpose, with records that agree with each other.
- Find every use, including the ones a vendor switched on.
- Set rules on one page that fit the system you run.
- Approve tools for named uses on dated evidence.
- Test a use against criteria you wrote first, and retest when it changes.