Technology and data
Every system and every dataset, with what is known about it, how it was learned, when it was last seen, and who answers for it.
An inventory that knows what it knows
Name, vendor, product, version, environments, criticality, lifecycle, strategic classification, owner, entity, capability — and provenance on every row. Declared by a person, imported from a template, inferred from an assessment, or verified by someone who vouches for it. Datasets carry domain, sensitivity, expected refresh and schema version, and the relationships that make lineage traversable.
Connectors, not vendors
A connector contract, so an enterprise system — ERP, CRM, HR, database, warehouse, cloud platform, file transfer, messaging, iPaaS — is a module of one shape. At launch: a generic REST poller with your own field mapping, CSV upload for the parts of the estate the cloud cannot reach, and a telemetry API for systems that push. Pull, read-only, on a schedule you control; every read is read, refused or not read, and the third alarms.
Health you can argue with
Ten dimensions, each derived from evidence you can open, composed by a rule you can read. Your thresholds. A score that is one defined summary of the states and never their source. An object nobody has measured is unmeasured; it is never quietly healthy.