The EAM platform with a
reliability engineer built in
IREAMS runs your maintenance — assets, work orders, PM, inventory, purchasing and cost, on desktop and offline mobile. Then it does what a CMMS cannot: computes your real reliability from your own history, and gives you an AI Reliability Specialist that cites every number it quotes.
Your CMMS records. Your reliability tool analyses.
Nothing closes the loop.
Almost every asset-intensive operation runs reliability engineering beside the maintenance system rather than inside it. That gap is where value quietly leaks away.
The analysis is always stale
Someone exports work-order history to a spreadsheet, cleans it by hand, and analyses a snapshot. By the time the study is finished, the plant has moved on — and the next study starts the cleaning again.
The failure data can't support the math
Weibull analysis needs clean failure histories at component level. If work orders close against vague locations with free-text failure notes, no amount of statistics will rescue the result. Garbage in, confident nonsense out.
Recommendations die as documents
An RCM study concludes; a PDF is filed. The PM schedule never changes because changing it means re-keying tasks into a different system that nobody owns. The binder becomes shelfware.
Nobody can prove it worked
Because the decision lived in one system and the outcome in another, there is no way to show that the revised interval actually reduced failures. Reliability stays a cost centre it cannot defend.
One loop, five steps
Each step feeds the next, and the last feeds the first. The longer IREAMS runs, the more it knows about your plant.
Bring your data in, in the right order
Start from a spreadsheet, a SAP export or another CMMS. The Migration Center walks the import in phases with dependency locks — the register loads before history, bills of material wait for assets and parts. A P&ID set can seed the register directly.
Run the day-to-day maintenance loop
Anyone raises a request. Planners turn it into a work order with a readiness score that shows what the plan still lacks. Technicians execute on mobile — offline if needed — while labour, parts and services settle to cost centres automatically. Autopilot merges overlapping PM visits into one shutdown window.
Measure reliability, honestly
Every completed corrective order becomes a coded failure event. From those, IREAMS computes MTBF, MTTR, availability, OEE and the rest of the SMRP vital few — each with its standard number and benchmark. Bad actors rank by real cost and downtime, not gut feel. This is where a CMMS stops and IREAMS begins.
Diagnose, model and decide
Pick a bad actor and the tools open pre-loaded with its history. Weibull fits life data with censoring handled correctly. Monte Carlo compares PM intervals against run-to-failure. RCA follows a physical → human → latent ladder with graded evidence. RCM writes the chosen task straight into the PM programme.
Let the Reliability Specialist keep watch
The Specialist reads the same data through governed tools. Each morning it briefs the team on bad actors, corrosion risk and backlog. On demand it investigates a failure, challenges an RCA, audits a PM programme or finds warranty money. It proposes; a person approves; the loop starts again with better data.
Good reliability math is a data-discipline problem first
This is the part most platforms skip. IREAMS structures every asset on the ISO 14224 hierarchy — Enterprise → Site → Unit → System → Equipment → Component — and work orders cannot be closed against a system. They must drill to the equipment or maintainable component that actually failed.
Authority dictionaries — work types, cost centres, failure modes — are centrally governed and lock on first transactional use, so history can never be contaminated retroactively. Technicians select standardized failure modes, causes and remedies before a critical work order can be completed.
That discipline feels like friction on day one. It is the reason that, a year later, your Weibull fits mean something and your Pareto is telling you the truth.
What the discipline buys you
- ✔️ Trustworthy MTBF — failure history recorded at the level that failed, not at the plant area.
- ✔️ Weibull fits worth acting on — coded failure modes give clean inter-arrival times and honest suspensions.
- ✔️ Cost that ties out — labour confirmations, goods issues and service receipts all settle to the asset and the cost centre.
- ✔️ Audit-ready by construction — append-only, tamper-evident logs of who changed what, when.
- ✔️ An AI that can cite — because the record is canonical, every number the Specialist quotes has a source.
One system, four working layers
Everything below runs on the same canonical record, with one permission model and one audit trail.
The EAM Backbone
Your operational single source of truth. Asset register on ISO 14224 taxonomy, requests and work orders with failure coding and TECO gating, preventive maintenance on time and meter triggers, scheduling with frozen zones, inventory with availability netting, purchasing, and full financial operations.
The Reliability Suite
Genuine engineering, not scoring. Censored Weibull life fits with confidence bounds, discrete-event Monte Carlo RAM, reliability block diagrams, Poisson spares optimisation, SAE JA1011 RCM and FMEA, evidence-graded RCA, Pareto defect elimination — and the PSC success layer.
Integrity & Compliance
Mechanical integrity as a closed four-step loop — assess, schedule, measure, evaluate. API 510/570/653 corrosion rates and remaining life from thickness readings, risk-based inspection screening, damage mechanisms, plus ISO 55001 audit workflows, LOTO permits and corrective-action tracking.
The Reliability Specialist
The AI layer that works across everything above. It never does the math itself — it drives the same deterministic engines your engineers use, cites every number back to your records, and drafts the work for a human to approve. An engineer's assistant with the whole plant record in its head.
Why this is not another CMMS
Maintenance software split into two camps: mobile-first CMMS tools that stop at “the work was done”, and enterprise suites that take a year to configure. IREAMS does both jobs.
| Capability | Mobile-first CMMS | Enterprise EAM suite | IREAMS |
|---|---|---|---|
| Field execution | Excellent — mobile, offline, simple. | Weak — desktop transactions; mobile is an add-on. | Mobile-first, offline-first, role-aware. Technicians land on My Work. |
| Reliability engineering | Not present. MTBF at best, computed loosely. | Add-on modules or consultants, configured over months. | Native. SMRP metrics, censored Weibull, Monte Carlo RAM, RCM, RCA, integrity — from live work history. |
| Failure data quality | Free text. | Catalogue profiles, rarely maintained. | ISO 14224 modes and causes, a review queue — and metrics that refuse to count uncoded failures. |
| AI | Chat assistant over work-order text. | Vendor copilot, roadmap-stage for most customers. | A Specialist that calls the product’s own engines, cites every figure and never writes without approval. |
| Cost & procurement | Parts list and a cost field. | Full — this is where the suites win. | Budget-checked POs, goods receipt, three-way match, settlement to cost centres — without the ERP project. |
| Time to value | Days. | Months to a year. | Days for register and work orders; reliability results as soon as history lands. |
It computes, it does not estimate
MTBF, MTTR, availability and OEE come from coded failure events using published SMRP definitions. No vendor-invented formulas.
Weibull that handles suspensions
Assets still running are treated as suspensions, not ignored — with confidence bounds shown. Most tools overstate failure risk by skipping this.
RCM writes the PM
The RCM decision becomes a task in the PM programme, not a spreadsheet nobody reopens. Run-to-failure is blocked where the consequence is safety.
An AI that cites its sources
The Specialist quotes figures from the same engines the screens use, links the record it read, and proposes changes for a person to approve.
Enterprise cost spine, without the ERP
Budget check on purchase orders, goods receipt, three-way match, settlement to cost centres — in a system technicians will actually use.
Migrate in phases, not in a project
SAP and generic templates, phased imports with dependency locks, and P&ID mining that derives the register from drawings.
Why the AI is only as good as
the system beneath it
Bolt a language model onto a messy CMMS and you get confident answers nobody can verify. The Reliability Specialist works because of what sits underneath it: a canonical record, coded failure history, and deterministic engines it can call but not overrule.
Ask it why an asset is failing and it queries the same Weibull engine your reliability engineer would open, reads the same ISO 14224 failure codes your technician selected, and cites the exact work orders behind its answer. It drafts the fix into the same work queue your planner already lives in.
That is what integration buys: an AI that can be checked. It advises and drafts; it never authorizes. No agent writes to your plant record on its own.
Agents orchestrate. Tools compute.
- ✔️ The model never does arithmetic. It selects the analysis and writes the explanation; the engine returns the number.
- ✔️ Every claim is cited to specific work orders, failure records or analyses in your own database.
- ✔️ Drafts, never writes. Proposals queue for human approval; autonomous action is not enabled.
- ✔️ Immutable audit of every agent run — who, what, when, and on what evidence.
- ✔️ Honest when data is thin. A capability appears only when the data behind it exists; you get an empty state before you get a guess.
Real screens, straight from the product
Captured from the live demo tenant — no mock-ups, no composites.
From preventing failure to
sustaining performance
An asset that runs is not the same as an asset that performs. IREAMS is the reference implementation of the PSC (Percentage of Success Centred) framework — the success-side mirror of classical RCM.
The Golden Spot
Each asset's optimal operating envelope, derived directly from the alarm bands your site already maintains. No new instrumentation, no new data entry — the envelope falls out of what you already record.
Success Rate & MTOP
MTOP is the success-side complement to MTBF: how long an asset sustains optimal performance. Success Rate measures availability against optimal operation rather than mere functioning. Target ≥ 90%; world class ≥ 95%.
Sub-optimal drift warning
Assets sliding out of their envelope are flagged with the limiting parameter named — long before anything trips an alarm. That window is where production loss is quietly accumulating.
SMEA — protect what works
Success Mode & Effects Analysis captures the conditions that sustain performance, ranked by SPN = Value × Sustainability × Monitorability. Where FMEA rewards what's easy to detect, SMEA rewards what can be actively sustained.
PSC is an additive layer. FMEA, RCA and classical RCM stay fully in place and remain authoritative for safety-critical cases.
Olorunfemi (2026), "A Success-Centric Evolution of Reliability-Centered Maintenance in Modern Asset Management," Science, Technology & Public Policy.
Aligned by construction, not by claim
The standards are not a compliance checkbox bolted on afterwards — they define the data model, the calculations and the workflows.
Data & asset management
ISO 14224 equipment taxonomy and failure data · ISO 55000:2024 asset management system · ISO 31000 risk-based prioritisation.
Condition & diagnostics
ISO 13374 six-layer processing spine · ISO 20816-3 vibration severity zones · ISO 17359 condition-monitoring workflow · ISA-18.2 alarm management.
Strategy & integrity
SAE JA1011 / JA1012 classical RCM · API 510 / 570 / 653 inspection intervals and remaining life · API 580/581-inspired RBI screening.
Governance & security
Row-level security with least-privilege role-based access · append-only, tamper-evident audit logs · a dedicated database per customer — your data lives in your own instance, not a shared tenant.
Start with a free reliability assessment
The fastest way to judge IREAMS is to point it at your own maintenance history. Send an export from whatever system you run today and get back what failure is actually costing you — bad actors in dollars, Weibull life fits with recommended PM intervals, PM waste, and recoverable warranty money.
- Any source: SAP PM, Maximo, MaintainX, Limble, eMaint, Fiix, UpKeep — or a spreadsheet.
- No commitment: the report is yours to keep either way.
- Findings in dollars: not a maturity score.