Stop paying for maintenance that doesn't prevent anything.
A critical, data-driven review of your entire PM program. We find the useless, the over-frequent, and the missing – with the savings quantified in hours and dollars, and the new plan ready to load in your CMMS.
Scoped per site · priced after a quick portfolio review on the first call
Industry studies consistently find that 30–50% of preventive tasks add no value. Yours is probably no exception – the difference is whether you know which ones.
01
Inherited frequencies.
"Every 6 months" because that's what the template said in 2009. Nobody has checked the intervals against actual failure data since.
02
PMs that address nothing.
Tasks that don't map to any real failure mode. They consume wrench time, spare parts and shutdown windows – and prevent exactly nothing.
03
Failures still get through.
Meanwhile the failure modes that actually hurt you have no task against them. You're over-maintaining half the fleet and under-maintaining the other half.
What You Get
Not a slide deck. Files ready to load in your CMMS.
PM kill-list, costed
Every task proposed for removal, with the hours and dollars it frees up.
Optimized frequencies
Intervals re-derived from your life data (Weibull), not from the template.
Coverage gaps closed
New tasks where failure modes currently slip through – RCM logic, prioritised.
Best-in-class benchmarks
Your PM frequencies and costs compared against industry references for the same asset classes.
CMMS-ready export files
Datasheets formatted for SAP PM, Maximo or your system – ready to push, no retyping.
Executive summary
One professional document with every improvement and the quantified gains – $ and hours.
How it works
From CMMS export to a leaner PM program.
STEP 01
You send one export
PM plan + 24–36 months of work-order history. Any CMMS, any format. NDA signed first.
STEP 02
Criticality screen
Assets ranked on a probability × consequence matrix, so the deep analysis goes where the money is.
STEP 03
AI-assisted analysis
We match every PM against your real failure history, then re-tune each frequency with Weibull, RCM logic and FMECA. Engineer-validated.
STEP 04
Quantified recommendations
Kill-list, new intervals, gap tasks – with CMMS-ready files and the executive summary.
Inside the analysis
Your failure history vs. your PM plan. Reconciled line by line.
The heart of the review is simple to state and hard to do by hand: every corrective work order in your history is coded to a failure mode, and every PM task is tested against those modes. What actually failed gets compared with what your strategy assumed would fail.
Input A · the strategy
PM plan & task list
Every task, its interval, trade, duration and asset – straight from SAP PM, Maximo or your CMMS.
Input B · the reality
Work-order & failure history
24–36 months of correctives, coded to failure modes – times-to-failure and suspensions included.
The task maps to no credible failure mode – or the failure keeps passing despite it. Candidate to remove or redesign.
Wrong interval
Life data shows over- or under-maintenance. The interval is re-derived from the Weibull fit, not the template.
Coverage gap
Repeated work orders against a failure mode that has no task. A new task is proposed – the right type, at the right interval.
Earns its keep
The task addresses a real failure mode at a supported interval. Kept – and now documented as to why.
The decision logic, applied to every task
Classic RCM questions, answered with your data instead of a workshop vote.
Does the task address a credible failure mode?
no → kill-list
Checked against your FMECA and your actual failure history.
Is the consequence worth preventing?
no → run-to-failure
Safety, environment, production loss – from the criticality screen. Run-to-failure is a strategy too, when it's a decision and not an accident.
Is the failure age-related or random?
random → condition-based
The Weibull shape parameter β tells us. When β ≈ 1, fixed-time replacement prevents nothing – the task switches to inspection or condition monitoring on the P-F interval.
Is the interval supported by the life data?
no → re-tuned
Too frequent: extended, hours given back. Too late: tightened before it costs you a failure.
Task earns its place – kept, at the right interval, for a documented reason.
Where the numbers come from
Your data first. References where it's thin – and we always tell you which is which.
01
Your CMMS history
Times-to-failure, suspensions, repair hours, parts consumed. The primary evidence for every interval decision.
02
OEM data & curves
Manufacturer recommendations as the starting point – then challenged against how the asset actually behaves in your service.
03
Industry failure-rate libraries
OREDA-type reliability data for the same equipment class, where your own sample is too small for a clean Weibull fit.
04
Best-in-class benchmarks
How top-quartile sites maintain the same asset classes – PM hours, task mix, condition-based share. Your program scored against it.
The AI does the matching at fleet scale. A senior reliability engineer reviews every kill, every interval change and every new task before it goes anywhere near your CMMS.
The Proof
The methods behind the numbers. On real data.
Every recommendation is backed by an established reliability method – Weibull life-data analysis for intervals, RCM logic for task value, root cause analysis when a failure keeps coming back. Here's what that looks like.
Weibull-V301-relief-valves.pdf
Method 1 · Weibull life-data analysis
From raw failure data to optimized PM interval.
Shape β
1.84
Scale η
3,840 cy
Savings
−38% PM
V-301 relief-valve fleet: the data showed the inherited interval was 38% too frequent. One asset class, −38% PM hours with no added risk – the same analysis runs across your whole fleet.
RCA-2026-031 · P-101A BFW pump
Method 2 · Root cause analysis
When a failure repeats, we go to root cause.
WHY 1
Pump tripped on high vibration (DE bearing > 12 mm/s)
WHY 2
Impeller imbalance from cavitation damage
WHY 3
NPSH-available dropped below NPSH-required
WHY 4
Deaerator level setpoint lowered 11 Mar
WHY 5 · ROOT
MOC review missed NPSH impact – procedural gap
Every "why" backed by evidence from SCADA, vibration FFT, operator logs and the OEM curve. 14 h of downtime, €38k – and a corrective action that stops the repeat, instead of another PM that wouldn't have.
Weibull, RCM logic, FMECA, RCA – the right method for each decision. The AI runs them at fleet scale; a senior reliability engineer validates every recommendation before it reaches your CMMS.
Not another AI pilot.You get a finished deliverable with quantified ROI. Most AI projects die in POC purgatory – this isn't one of them.
No IT project.A simple CMMS export is enough. No system access, no software to install, no integration.
NDA-first.Your data stays isolated, never trains public models, and is deleted after the engagement.
Remote-first. On-site when it helps.The review runs remotely – and we come to site for the criticality workshop if that's how your team works best.
FAQ
Questions you're probably asking.
How long does a full PM program review take?
Typically 6–8 weeks for a site-level program, depending on the number of assets and the state of the data. We confirm the timeline on the scoping call – and it's weeks, not the quarters a classic RCM project takes.
Will you touch our regulatory / compliance PMs?
Never without flagging them. Statutory and compliance-driven tasks are identified, marked separately, and kept out of the kill-list. You get the full picture, but regulatory tasks are only ever touched with your compliance team in the loop.
Our failure history is incomplete. Does the analysis still work?
Yes. Where your life data is thin, we lean on industry failure-rate libraries, OEM data and best-in-class benchmarks for the same asset class – and we tell you explicitly which recommendations rest on your data and which on references.
Is this just another AI proof of concept?
No. You receive CMMS-ready files and an executive summary with quantified gains – a finished deliverable, validated by a senior reliability engineer. The ROI is in the document itself, not in some future rollout.
Remote or on-site?
The analysis runs remotely from a CMMS export. The criticality review and the findings session work well as remote workshops – and if your team prefers a day on site for those, we do that too.
Built by a 20+ year reliability expert
Not theory. Practice.
Rob has run RCM, Weibull, RAM and LCC studies across LNG, power, mining and oil & gas – and killed more useless PMs than he can count. The AI runs his playbook at fleet scale; an engineer signs off on every change.
How much of your PM budget prevents nothing?
30-minute scoping call. Bring a PM plan export – we'll tell you where the likely waste is and what a review would return. No pitch. No obligation. If there's no clear ROI, we'll say so.
