FactoryCortex: root-cause analysis that keeps up with the line
FactoryCortex watches every process signal, tells you what moved and in what order, ranks the likely causes against the evidence, and closes the loop only when the fix is verified.
Co-founded byVerto SolutionsandInfinity Technologies
- Spot → story → causes → fix
- The whole RCA loop on one screen
- 8 read-only evidence tools
- Historian, alarms, CMMS, SOPs, SPC and more
- 4 colours, 1 meaning
- Needs action · needs a look · normal · settling
Try it
Walk a fault from first signal to verified fix.
Three production lines, two of them in trouble. Pick one and follow it through FactoryCortex: what happened, the likely cause, the work order, and the chart that proves the fix held.
- 1Spot it
- 2See what happened
- 3Rank the causes
- 4Act
- 5Confirm the fix
Line 01 · Parquet assembly is flagged. Open it to see what happened.
Workspace › Production intelligence
All your lines
Good evening, Alex
2 of 3 lines need you. The rest are said out loud below, so “no news” never hides “no data”.
Needs attention
2Past its limit · Line 01 · Parquet assembly
Layer offset is past its limit.
Process moved · Utility 02 · Boiler & utilities
Valve tracking error is heading for its limit.
Running normally
- Cell 03 · Precision machining — all 6 signals normal.
How it works
A dashboard says 2.4 mm. FactoryCortex says whether that matters.
The loop, in order: control charts notice the process moving, the story explains it, the investigation ranks what might be causing it, and work orders get it done — then the charts say whether it held.
- 01
Spot it early
Every signal gets its own centre line, control limits and sigma zones. Western Electric and Nelson rules catch a shift, run or trend long before a single reading looks wrong — and a projected-crossing rule warns while product is still in spec, when the fix is cheapest.
- 02
See what happened
Instead of a wall of statistics, each problem leads with its story: what it is doing now, which signal moved first and which followed, what else happened at that time — maintenance notes, alarms, changeovers, failed quality checks.
- 03
Rank the causes
A deterministic engine scores the line's known failure modes against the evidence, shows what supports and what contradicts each one, and keeps an explicit reserve for causes nobody has modelled.
- 04
Act, then verify
Steps come verbatim from the matching SOP. A person decides to raise the work, the technician reports what was found, someone accepts it — and the same chart shows whether the fix held.
Built to be trusted
Fast is only useful if it’s right.
The most expensive investigation is the one that spends two shifts diagnosing a process that never moved, because the instrument watching it did. FactoryCortex is designed around the ways RCA usually goes wrong.
Who it’s for
Plant directors. One verdict per line, meaning the same thing on every screen.
Line managers. What needs you, most pressing first, each with the one button that deals with it.
Maintenance. Work that arrives with the evidence, the SOP step and a clear owner.
Quality engineers. Control limits you can set, capability that’s honest, and improvement projects with real gates.
Measurement first
Before a trend is treated as the process, the instrument is checked — gaps, stale feeds, resolution, redundant sensors that disagree. If it can’t be believed, no rules run on it.
Correlation is never cause
Every association is labelled as one. “Moved 22 seconds earlier, r = −0.78” is a hypothesis to test, not a confirmed root cause.
AI never touches the numbers
An optional AI analyst can word the explanation over the same evidence. Scores, candidates and cited readings come from the engine; anything else is discarded.
Every decision auditable
Tool inputs and outputs, evidence IDs, the observation window, the scoring formula and the SOP version are kept — exportable with a SHA-256 digest.
FAQ
About FactoryCortex
What is FactoryCortex?
FactoryCortex is a root-cause analysis (RCA) console for manufacturing and process lines, co-founded by Verto Solutions (Toronto) and Infinity Technologies. It watches production signals with statistical control charts, explains what moved and in what order, ranks the known failure modes against the evidence, and turns the fix into tracked work that is verified on the same chart.
Who is building FactoryCortex?
FactoryCortex is co-founded by Verto Solutions, a Toronto technology company that builds custom software, automation and data systems, and Infinity Technologies, which designs and builds enterprise AI agents and AI-powered products.
Is FactoryCortex available?
FactoryCortex is in development and has not been released yet. Manufacturers interested in an early pilot can contact Verto Solutions at info@vertosolutions.ca.
How is FactoryCortex different from a dashboard?
A dashboard shows that a value is 2.4 mm. FactoryCortex says whether that is unusual for this specific line, which signal moved first, which known failure fits the sequence, which signals argue against the alternatives, and what the matching SOP says to do — so a team can act in minutes instead of spending a shift reconstructing events.
Does FactoryCortex use AI?
The ranking of causes is done by a deterministic engine with a visible scoring formula and an explicit reserve for unknown causes. An optional AI analyst can write the explanation over the same evidence, but it never changes the numbers, and every record it cites is validated before anything is saved.
What data does FactoryCortex work from?
Process historians and sensor trends, alarms, controller logs, quality samples, maintenance notes (CMMS) and SOPs. Before a trend is trusted, FactoryCortex checks the measurement itself — gaps, stale feeds, resolution, and disagreement between redundant instruments.
Early access
Run a line that should explain itself?
We’re talking to manufacturers about early pilots. Tell us about your line and where investigations stall today.