CrossnoKaye puts AI into the control room at more than 100 industrial sites

CrossnoKaye says its AURA AI system can investigate plant problems using live equipment data, control logic and sensor readings, and is deployed across more than 100 industrial facilities. One example shows why that context matters: at a quarry, AURA traced recurring crusher stoppages to a slipping conveyor belt, rather than the crusher that first appeared to be at fault.

AWS AI Watch analysis

What happened

CrossnoKaye built AURA into its ATLAS industrial automation platform, which connects to existing plant equipment rather than requiring customers to replace it. The company says its platform covers 13,000 controlled assets and processes more than 50 billion new machine readings each day. AURA can query equipment layouts, control rules, sensor readings, alarms and operating histories under the caller’s permissions.

In the quarry example, the system linked a secondary crusher fault to an overload interlock triggered by a slipping belt beneath a tertiary crusher. The interlock had stopped the secondary crusher to prevent a downstream blockage. CrossnoKaye says an operator inspected and fixed the belt within minutes. The company describes how it built the system in an AWS case study.

Why it matters

Industrial control systems contain context that can be hard to assemble from separate logs and vendor systems. Here, AURA’s investigation drew on the plant layout and control logic as well as timestamps, helping distinguish the original fault from the equipment that reacted to it.

The system can also run Python analysis in isolated micro-virtual machines for calculations over large sensor datasets, rather than asking a language model to do the maths in its prompt. CrossnoKaye says its existing automation platform continues to run local control if cloud connectivity drops. The company also reports that its control deployments have had more than one million run hours with zero safety incidents; that is a company-reported figure, not an independent safety assessment.

Our read

This is a more grounded use of AI than handing a chatbot the keys to a factory. AURA investigates plant conditions, while the existing control system and human operators remain central to what happens next. The quarry example makes the value legible: finding the cause of a fault can matter more than producing a confident explanation of the wrong one.

The account comes from AWS and CrossnoKaye, and the performance claims are theirs. The next test is whether other operators can show comparable results beyond a company case study.

What to watch

• Whether CrossnoKaye publishes independent evaluations of AURA’s diagnoses and operational results.
• How the system’s permissions and human approvals work when an investigation recommends action.
• Whether the reported efficiency gains hold across different facilities and equipment.

Discussion spark: In industrial settings, should AI be allowed to recommend control changes, or should its role stop at diagnosis until independent evidence proves it reliable?

not affiliated with or endorsed by Amazon Web Services (AWS)