CMMS records
Work orders, failure codes, notifications, corrective hours, parts, interventions and maintenance history.
AI-powered Bad Actor Review
Your worst assets, ranked and explained.
AI finds the patterns. An engineer decides what they mean.
Sample output
Five assets. Eighty percent of the loss.
Every plant has a short list like this. The work is proving which five.
Cost of unreliability
Illustrative output
€882k concentrated in 10 assets
412 assets analysed · 12 months of work history · SAP PM export
Pareto
Where the money goes
Critical
High
Watch
Cumulative %
Five assets carry 81% of the loss. The rest is noise until those five are fixed.
Recurrence
The same failures, again
Loss split
By failure category
€882kTotal
Rotating / bearings 44%
Sealing 23%
Electrical 18%
Instrumentation 15%
81%
of the annual loss sits in 5 assets out of 412
4 of 6
events on the top actor share one failure mechanism
147 h
production downtime tied to the number one asset
R//
Figures are illustrative, not client data.
Engineer-validated read
4 assets ready for full RCA · 3 PM intervals to revisit · 2 duty corrections proposed.
Inputs
Move beyond a flat ranking.
A Pareto is the start. What sits behind it is the job.
Cost and downtime
Maintenance spend, lost production, unavailable hours, event frequency and agreed consequence measures.
Hierarchy and duty
Asset structure, operating service, redundancy, criticality, known modes, documents and engineer observations.
AI-connected evidence
Connect the ranking to the reasons behind it.
Each ranked asset linked to its events, losses, modes and causes.
AssetEventFailure modeDowntimeCostCauseAction
Rank#1 · Transfer pump P-204
Loss147 h downtime · 6 repeat events
Pattern4 events share bearing damage mechanism
ContextOff-design duty recurs during campaign change
PriorityDuty correction + lubrication control review
Workflow
From years of noisy history to a costed action portfolio.
AI does the sorting. Context decides what deserves action.
AI normalizes and ranks
Resolve asset identity, clean the measures and apply agreed ranking logic.
Cluster recurrence
Group repeat events, descriptions, symptoms and failure codes.
Add context
Relate top actors to duty, modes, history, documents and constraints.
Engineer review
Validate patterns, challenge false groupings and identify gaps.
Prioritize actions
Build a costed, owned plan aligned to value and feasibility.
Deliverables
A prioritized list with enough context to act.
Which assets matter, why they rank, what to do next.
- Ranked Bad Actor list with agreed loss measures
- Cost, downtime and recurrence Pareto views
- Repeated failure-mode and event clusters
- Context packs for priority assets
- Costed action portfolio with owners and timing
- Data-quality and uncertainty register
What remains connected
Today's ranking becomes tomorrow's reliability context.
Reviewed relationships and decisions stay in Factory Brain. Next time, you start from there.
- Normalized asset identity
- Loss and recurrence history
- Reviewed event clusters
- Known failure-mode patterns
- Prioritized actions and owners
- Assumptions and data gaps
Human validation
Rankings direct attention. Engineers decide what they mean.
AI surfaces patterns fast. Site context decides whether they are real.
R//Safe
Never trains a public model
Source-linked outputs
Engineer-validated
You keep the knowledge
Start with recurring loss
Find what is costing you most, in weeks, not months.
Bring your work history, your loss measures and your questions.
Start a Bad Actor Review