Aelix IQ · software-only predictive maintenance
Find the failure hiding in the
data you already collect.
Predictive maintenance that runs on the telemetry you already have. Live asset health scores, explainable degradation alerts, and the scheduled fix that follows. No sensors. No install. Live on your data in weeks.
Runs on your historian, PLC / SCADA, OSIsoft PI, Ignition or MQTT · Zero proprietary hardware
Agentic AI · The Maintenance Agent · Design-partner preview
Predict, then act. The agent does the legwork.
An alert fires. Before anyone touches it, the IQ agent pulls the asset's recent history, flags what changed, and finds similar past failures. Then it drafts a work order with the evidence and a recommended window, routed to your engineer for approval. Approve it, and the agent schedules the job and notifies the tech. The loop closes itself. You stay in command, and you can see exactly what the agent did and why.
Investigate.
Pulls the asset's recent history, the changes that matter, and similar past failures from the historian.
Draft the work order.
Attaches the evidence, recommends a window, and routes it to your engineer for review.
Wait for your yes.
Nothing touches the plant until a human approves. You decide what the agent is allowed to do.
Close the loop.
Once approved, the agent schedules the job, notifies the tech, and writes every step to the audit trail.
The guardrails
Explainable.
Every action shows the reasoning, the data it used, and the trade-offs.
Human in the loop.
Consequential actions require sign-off. By default.
Guardrailed.
Limits you set on what the agent may do on its own. Fails closed when uncertain.
Audited.
Every step, agent or human, written to the audit trail.
Status: Design-partner preview. In active validation. Not presented as fully autonomous or generally available.
The gap nobody else closes
Every sensor vendor says install our hardware. Every CMMS says manage the work after you already know. You sit in the middle: real telemetry, real downtime, no program.
Your historian already holds the warning signs. They just have not been turned into a heads-up. Aelix IQ owns that gap: point us at your last twelve months and we will show you a health score and one maintenance action you would have caught, before you spend a dollar on hardware.
What you get
No sensor tax.
Start predicting failures without a capital project. Aelix IQ ingests the telemetry you already collect and computes a live health score for every monitored asset. The sensor crowd charges $300 to $800 per asset for hardware. We run on data you already own.
Explainable alerts you own, and the fix attached.
On a plant floor, an alert your engineer cannot explain is an alert your engineer ignores. Your controls team authors degradation rules in plain language, so every alert arrives with the reason it fired. Then one click schedules the maintenance work order. Detect, explain, schedule, execute, in one loop.
Proof on your own data, in weeks.
See a real, quantified catch on your data before you commit budget. We ingest your last twelve months of asset history, backtest it, and surface at least one failure you would have caught. Weeks, not the multi-quarter, integrator-led project the enterprise suites require.
Demand forecasting, included.
A bundled machine-learning demand-forecasting module ships in the box, so one tool also helps you plan production against demand. No second vendor, no extra integration.
A glimpse of the work
Spindle vibration · Line 4
Fired because your rule vibration > 6.5 mm/s for 72h held, authored by M. Chen, Controls Engineer. Not a black box, logic your team wrote and can read.
Asset health surface · predictive maintenance
Why it lands
4 wks
Median time to first provable catch on partner data
$300-800
Per-asset sensor spend avoided
12 mo
Of historian data we backtest before you commit
Design-partner plant types: tier-2 auto stamping, food and beverage packaging, CNC job shops.
The question every engineer asks
"Is this just another AI black box?" No. We chose explainable, author-your-own rules over a black box, and we attach the scheduled fix the copilot crowd leaves off. Rules you trust, plus the fix, beats a confident probability that walks away.
Demand forecasting is ML · maintenance alerts are explainable by design
From data to catch
A short onboarding, a provable first catch, then the loop runs itself.
Connect
We ingest your historian and stand up your tenant. PLC / SCADA, OSIsoft PI, Ignition or MQTT. Nothing exotic, nothing installed on the line.
Catch
Your controls engineer co-authors degradation rules in plain language. We backtest on your history and surface the first failure you would have caught.
Close the loop
The operator floor view goes live, green, amber, red at a glance. The first real alert becomes a scheduled maintenance work order, in-tool.
Quantify
We pin a number to the avoided failure, downtime hours times cost, and hand you a story your team owns. Your rules, your data, your catch.
Design-partner cohort open
Point us at your last twelve months of asset data.
We will show you health scores and one maintenance action you would have caught. No sensors. No install. No bill until we have proven a real catch on your data.
Need this tailored to your environment?
Every Aelix product can be configured, extended, or built bespoke for your industry, data sources, and compliance constraints. Talk to our engineers about what would change.
Configurable workflows
Adapt rules, thresholds, and approval flows to match your operational policies.
Custom data integrations
Connect to your specific ERP, MES, SCADA, CRM, or proprietary systems.
Bespoke modules
Build product extensions tailored to your industry, region, or compliance needs.