Tau
AI for manufacturing · 11 lessons

AI in manufacturing, from shift notes to inspection

Use AI to write shift handovers and work instructions, find what causes breakdowns and plan smarter maintenance and quality checks. For plant, quality and maintenance teams and supervisors.

Who it’s for: Plant, quality, maintenance and process engineers; production supervisors.

On a factory floor, small problems add up fast. A vague shift handover, a missing work instruction or a repeat breakdown can cost hours. This course shows plant, quality and maintenance teams where AI genuinely helps, and where it doesn't.

You'll start with the paperwork. You'll turn messy shift notes into a clean handover with open issues and owners. You'll capture an expert's know-how as a step-by-step work instruction with safety callouts, then check it on the line. You'll use AI to clean up downtime codes and comments. Then you'll build a Pareto chart, which ranks the causes from biggest to smallest, and list root causes to test. You'll also write 8D and CAPA problem-solving reports faster, with every claim tied to evidence.

Later lessons look at bigger projects. You'll choose the right first machine for predictive maintenance, which uses sensor data to warn you before a failure. You'll build the defect library an AI inspection camera learns from, and prove it matches your best inspectors before it goes live. You'll also cover machine safety, the security of the systems that run your equipment and the protection of trade secrets.

By the end you’ll be able to…

  • Turn shift logs and downtime data into a ranked root-cause list
  • Draft and validate work instructions and 8D reports with AI
  • Scope a predictive maintenance or visual inspection pilot with KPIs
  • Prove an AI inspector against expert judgement before go-live

Short on time? Start with the essentials

These 4 lessons cover the most useful ideas in this course. Each one takes about a minute.

Essential 1

Where AI actually helps on the factory floor

Distinguish where generative AI, machine learning and computer vision fit across maintenance, quality, planning and documentation.

Essential 2

Turn messy shift notes into a clean handover

Use a chat assistant to summarise shift logs into open issues, actions and owners without inventing detail.

Essential 3

Find your top downtime causes from raw logs

Use AI data analysis to clean downtime codes and free-text comments, build a Pareto and propose root-cause hypotheses to test.

Essential 4

Write faster 8D and CAPA reports without losing rigour

Use AI to structure an 8D or CAPA, prompt 5-Whys and fishbone thinking, and keep every claim traceable to evidence.

Every lesson in this course

This course has 11 video lessons of about a minute each, from the basics to expert habits. Each one ends with a question, three takeaways and its sources.

Foundations 2 lessons

  1. Where AI actually helps on the factory floorEssential

    Distinguish where generative AI, machine learning and computer vision fit across maintenance, quality, planning and documentation.

    “Every alarm, scrap tag and downtime code in your plant is data AI can learn from.”

  2. Turn messy shift notes into a clean handoverEssential

    Use a chat assistant to summarise shift logs into open issues, actions and owners without inventing detail.

    “Three pages of scribbled shift notes become a six-line handover in 30 seconds.”

Building skills 3 lessons

  1. Capture expert know-how as work instructions

    Convert an SOP or a recorded expert walkthrough into a step-by-step work instruction with safety callouts, then validate it on the line.

    “Your best setter retires next year. Where does everything they know go?”

  2. Find your top downtime causes from raw logsEssential

    Use AI data analysis to clean downtime codes and free-text comments, build a Pareto and propose root-cause hypotheses to test.

    “'Other' is your biggest downtime code. AI can read the comments and tell you why.”

  3. Write faster 8D and CAPA reports without losing rigourEssential

    Use AI to structure an 8D or CAPA, prompt 5-Whys and fishbone thinking, and keep every claim traceable to evidence.

    “AI can draft your 8D in minutes. It can't go to the gemba and see the fault.”

Advanced 5 lessons

  1. Pick the right first asset for predictive maintenance

    Choose one asset for a condition-monitoring pilot by its cost of failure, how much warning its failure modes give and the sensor data you already have.

    “Run-to-failure versus weeks of warning: what changes for your most critical pump?”

  2. Build the defect library an AI inspector learns from

    Define defect classes, limit samples and image conditions so an AI visual inspection system learns what your customers actually reject.

    “An AI inspector only knows the defects you show it. How many photos is enough?”

  3. Keep AI safe: machine safety, OT security and IP

    Identify when AI touching machines triggers safety assessment (e.g. the EU Machinery Regulation from January 2027), OT cybersecurity controls (e.g. IEC 62443) and trade-secret protection, and check what applies at your site.

    “From 20 Jan 2027, EU machinery law covers safety parts that learn and change.”

  4. Update a PFMEA after a process change

    Use AI to propose new failure modes, effects and causes for a changed process step, then have the cross-functional team rate and own each row.

    “New fixture on Op 30. Did anyone update the PFMEA? AI drafts the rows in minutes.”

  5. Roll out AI with operators, not to them

    Run a two-week operator trial of an AI alert with a one-tap "useful or false alarm" response, and change the alert based on what operators report.

    “If operators don't trust the alert, the AI saves nothing.”

Expert 1 lessons

  1. Prove an AI inspector matches your best inspectors

    Run an attribute agreement study on a blind sample of good and defective parts to measure an AI inspector's escape and false-reject rates against expert judgement before go-live.

    “Your AI inspector is "99% accurate". How many bad parts did it pass?”

Apps we recommend

These are the tools this course uses, and what each one is best for. We have no partnership with any of them.

Augury

Best for: Predictive maintenance on rotating equipment

Wireless vibration and temperature sensors plus AI diagnostics flag faults on pumps, motors and fans before failure.

Paid

LandingLens

Best for: AI visual inspection without a data science team

Quality engineers label defect images, train vision models and deploy them to inspection stations with little code.

Freemium

Tulip

Best for: No-code frontline apps for operators

Builds digital work instructions, quality checks and line dashboards connected to machines, with built-in AI features.

Free trial
Hands-on project

Cut one downtime or scrap loss on your line

Take 4–8 weeks of downtime or scrap data from one line. Using approved tools and your site's data rules, clean the codes, build a Pareto, confirm the top root cause on the line and write an A3. Then rank three AI use cases for your plant by value, data readiness and safety or OT risk.

What you’ll make: A one-page A3 (Pareto, confirmed root cause, countermeasures) plus a ranked list of three AI use cases for the plant.

Check your own work

  • Every Pareto bar reconciles to the raw export, with the recoding rules written down
  • The root cause was confirmed on the line or with data, not taken from the AI alone
  • Each countermeasure has an owner, a date and a metric that shows whether it worked
  • Each AI use case states expected value, data source and a machine-safety and OT-security check
  • No customer drawings, recipes or supplier prices went into an unapproved tool

Try a ready-made prompt

Copy a better prompt for a real task, and see why it works better than the version most people type.

Further reading and viewing

We’ve chosen these articles, research papers, videos and guides from other trusted sources.

Guide · NIST Manufacturing Extension Partnership · 25 min

Artificial Intelligence: Key Considerations and Effective Implementation Strategies

Plain-language Industrial AI guide for small and mid-sized manufacturers, built around their real questions.

Article · World Economic Forum · 15 min

Global Lighthouse Network

Case studies of factories that scaled AI and digital tools, with the operational results they measured.

Article · IBM Think · 10 min

How can AI be used in predictive maintenance?

Clear explainer of sensor data, failure prediction and how predictive maintenance differs from scheduled PM.

Guide · NIST · 40 min

AI Risk Management Framework

The reference framework for governing AI risk, useful when writing your plant's AI policy and pilot criteria.

Questions about this course

How is AI used in manufacturing?

There are three main kinds. Generative AI, like ChatGPT, helps with writing such as handovers, work instructions and reports. Machine learning finds patterns in sensor and production data, which is the basis of predictive maintenance. Computer vision checks parts on camera for defects. The course explains where each one fits, and why a chat assistant isn't the right tool for every problem.

What is AI predictive maintenance?

It uses sensor data, such as vibration or temperature, to spot signs that a machine is heading for a failure. That gives you time to plan a repair instead of reacting to a breakdown. Tools such as Augury do this on rotating equipment. The course shows you how to choose a first machine by its cost of failure, how much warning its faults give and what data you already have.

What safety rules apply when AI touches machines?

When AI affects how machines run, safety and security rules can apply. In the EU, the Machinery Regulation can require a safety assessment. IEC 62443 is a widely used standard for securing the systems that run industrial equipment. The course uses these as examples. Check what applies at your site, keep process data and trade secrets inside approved tools, and validate any AI-written instruction on the line.

How long are Tau's manufacturing lessons?

Each lesson takes about a minute. You watch a short video, answer one question, then get three takeaways and the sources. The course has 11 lessons, and you can start with the essentials if you're short on time. Your first 5 lessons in Tau are free, and Tau Pro gives you every course and lesson.

Get early access to this course

Tau opens soon. Join the waitlist and we’ll email you as soon as you can start this course.