How to start
Start with a trial.
Move to commercial.
A structured trial on the SCADA history you already hold — three to ten facilities, three to six months, graded at every gate. You finish with events from your own history, their likely causes and a business case costed against your KPIs. No new field hardware. No waiting for data to accumulate.
- History
- 12–24
- months of SCADA
- Scope
- 3–10
- facilities
- Trial
- 3–6
- months
- Your time
- 30
- min every 2 weeks
- P0 · Data
- P1 · See
- P2 · Use
- P3 · Trust
- P4 · Scale
EVT-0142 · from your history
Compressor blowdown
Station 07 · Unit 2 · 02:14 · 37 min
- Volume
- 41 MCF
- 95% · 27–66
- Detection
- Strong
- 3 sensors agree
- Likely cause
- ESD · high disch. temp
- Good
Your review
The gates
It earns trust in phases.
Five gates, five questions. Each one is graded on a number and on a capability your team gained, against thresholds we agree together at kickoff.
- Phase 0 Month 1
Do we have what we need?
Data staged, tags mapped, validation basis agreed — before results exist is the right time to agree what they'll be judged against.
data · validation basis
- Phase 1 Month 2
Can we see it?
Learns your normal, then pulls first fingerprints from your own history. Your team opens AroIQ and reviews an event that was not in the records, with the SCADA evidence behind it.
first events · evidence
- Phase 2 Months 3–4
Can we use it?
Our analysts label the patterns AroIQ finds. Your team reviews a handful — thirty minutes every two weeks — and each answer carries to the whole pattern. Then the real test: using AroIQ alone, you decide whether an event needs a crew.
usable in daily ops
- Phase 3 Month 5
Can we trust it?
AroIQ is scored blind on events your team held back. Every number traces through its calculation, ready for an auditor.
blind test
- Phase 4 Month 6
Can we scale it?
Events costed against your KPIs — the scaling business case. The commercial question opens at the Phase 2 review, not at the end.
business case · decision
Shown for a six-month trial · shorter trials run the same gates on a tighter clock
What a trial needs from you
Less than you'd think.
The trial runs on data you already hold. Five asks, and only one of them recurs.
Not needed to start
- New sensors or field hardware
- A tag list or tag templates
- A live historian connection
- Clean data
- 1
12–24 months of historical SCADA.
Send everything — every tag, at the finest interval you have, not rolled up. Historian exports in any format; a large CSV is fine.
- 2
A P&ID, PFD or general arrangement per site — helpful, not a blocker.
AroIQ can rebuild a station's map from how its sensors behave together. Drawings get us to equipment-level answers sooner.
- 3
Maintenance records and known events.
your validation basisWork orders, known blowdowns and ESDs, past survey findings, spills, O₂ and H₂S excursions — even if the timestamps are rough.
- 4
Thirty minutes every two weeks.
one answer, every matching eventOur analysts do the labelling. Your engineers review a handful of events — confirm, override or flag — and each answer carries to every event in the same pattern.
- 5
Named technical and operations contacts.
An operations or reliability lead, an emissions lead, a facility engineer, and someone from IT for the security session.
What you get
A trial report. Then the product.
The trial closes with a report you can take to a budget meeting. Commercial picks up where it leaves off: the same events, continuously, across every site you bring.
At close · trial report
Phase 4- Results by event type, with volume ranges and ranked causes — including what didn't work
- Which events showed precursor conditions, and which didn't
- The business case from trial data, costed against your KPIs, with the method shown
- Data-gap recommendations: which tag would tighten which result
- A scaling roadmap: which sites next, and what each one needs
Commercial · AroIQ dashboard
See every event. Act on the few that matter.
Triage and the Investment Plan for the people who decide. Events Investigation and the RCA Workbench for your engineers. The calculation audit for your auditors.
- Triage
- Patterns
- Sites
- Investment Plan
- Events Investigation
- RCA Workbench
Commercial · model-as-a-service
Data into your own stack.
Raw outputs — events, volumes, causes, confidence — delivered where your teams already work. Not an either/or: run one or both.
- Snowflake
- Ops centre
- Work orders
FAQ
Questions we get on every first call.
Short answers here. The long ones are better on a call with the people who built it.
Ask us yoursDo we have to start with a trial?
No. The trial is there so you can judge AroIQ on your own data before you commit. If you already have the business case, start commercial: we stage your history the same way, and the same gates run as your onboarding — under a commercial agreement, across the sites you choose.
Do we need to install anything?
No. That's most of the point. We ingest and stage a historical export first, so you see what AroIQ finds on events you already know about — before anything touches a live feed.
How does AroIQ fit with OGI surveys and sensors?
AroIQ needs no new field hardware. It reads the operating record you already collect — continuously, at every SCADA site — and gives each event its equipment, likely cause, volume and cost. Its limit: a small component leak that changes no pressure, flow or temperature leaves nothing to read. A site visit or an OGI camera can confirm a specific event while it's under way or when it recurs, and your survey obligations stay with the approved methods.
Is our SCADA good enough?
Almost certainly. Drift, gaps and inconsistent tag names are the normal condition of SCADA data, and the cleaning step is built for them: self-tested gap fills, drift correction, telemetry codes stripped. A trial on your own export is the quickest way to know.
Our SCADA is sparse — does that matter?
Sparse doesn't mean missing. Most SCADA writes on change, so a quiet tag usually means nothing moved — and the model reads it that way. Every result carries a confidence that states what the data supports, and where one more tag would tighten it, we'll tell you which.
Does this create new reporting obligations?
AroIQ is an internal operational analysis, not a regulatory measurement method. It supports your compliance program with per-event evidence. Whether a finding triggers a duty to repair, record or report depends on your regulator's rules, so we walk through it with your regulatory team early in the trial.
Has this been independently tested?
Yes. METEC at Colorado State University independently evaluated our SCADA-based methodology (January 2026) and found operating states identified from SCADA alone were statistically distinct and correlated with measured emissions. Phase 3 of the trial then scores AroIQ blind on events from your own history.
Is this a language model?
No. Detection and causation models are built in-house by our machine-learning data scientists and trained on sensor data. Language models only draft the plain-English cause notes — and people review them.
Who on our side uses it?
Operations, reliability and emissions — the same events, different views. Operations sees the upset and the production it cost; reliability sees the equipment and the recurring pattern; emissions sees the volume, the range and the audit trail.
Every trial starts with an export
See your own SCADA
find its first event.
Tell us which facilities and what you're working on. We'll bring the right people — and the security session comes first.