Security and deployment
Your settings files are sensitive. We built for that first.
Relay settings, facility ratings and topology are the kind of data your CIP program exists to protect. data phleet runs where your security team is comfortable, all the way down to fully on-premise and air-gapped.
01 / Deployment models
02 / Principles
Closed loop
Your data is processed inside the deployment you choose. It isn't sent to third-party AI services and isn't used to train shared models.
Deterministic results
Gaps and findings come from fixed rules, not model output, so they can be reproduced exactly, including years later during an audit.
Traceable
Every value keeps a pointer to its source document and page. Every result records the logic version, the reviewer and the time.
Least data
data phleet only asks for the documents each requirement needs, and de-identified inputs mean the cloud never has to know which plant a file belongs to.
03 / Questions
What security teams ask us
Where does our data go?
That depends on the deployment you choose. In the cloud model, data phleet runs in Microsoft Azure with TLS in transit and AES encryption at rest, and inputs can be de-identified before processing and re-associated with your own keys afterward. In the hybrid model, documents are read and processed inside your network and only results reach the cloud interface. On-premise, the entire platform, including the document-reading models, runs inside your firewall and nothing leaves your network.
Is it safe to use with CIP-sensitive information?
Protection settings, ratings, coordination data and topology can be BES Cyber System Information or otherwise sensitive. Putting them into a general-purpose AI tool raises CIP-011 information protection and CIP-013 supply chain questions about storage, retention, training use and access. data phleet is a closed system: your data stays inside the deployment model you pick, it's never used to train shared models, and every input and output is traceable.
Is our data used to train AI models?
No. Your documents and results are not used to train models shared with other customers.
Can it run air-gapped?
Yes. The on-premise deployment is self-contained and air-gap capable.
Bring your security team to the demo.
We'll walk through the deployment model, data flows and controls with them directly.