The field manual for machine learning in manufacturing · updated September 1, 2026
ManufacturingML
Standards

Evidence rules we hold ourselves to

Reference material is only useful if you can tell where every number comes from. These are the rules.

Attribution

Every market figure is attributed inline to its source (e.g., IoT Analytics for market sizing and use-case shares). Figures we cannot attribute do not appear. Public company results (Renault, Georgia-Pacific) are cited as published.

Vendor-reported labeling

Results reported by vendors about their own deployments are always labeled vendor-reported and should be weighed as such. We do not convert vendor claims into unattributed facts.

Update cadence

Entries and the adoption report refresh on a quarterly cycle, aligned with the re-score cycle of our sister evaluation site. Material changes are dated.

Changelog
  • 2026-09-01: Initial publication: 15 entries, 2026 Adoption Report (10 insights), use-case library, AI & jobs analysis.

Ownership & independence disclosure

This site is educational reference, not a review site; vendors appear as labeled examples with vendor-reported figures, under the same attribution rules for everyone. Rubric-scored platform comparisons live on IndustrialProcessAI.com. No vendor pays for coverage on this site.

Reference
Twenty terms defined in plant terms, plus a model selection tree
Free, no login. One follow-up from TEEPTRAK, then nothing.
Get the glossary →
2026 State of ML in Manufacturing
Benchmark your AI maturity against peers: adoption data, ROI by use case, and where to start without a data-science team.
Get the Free 2026 Report →