Your tools. Your science. Your data.
Customer control is a product requirement. These principles guide what we build and how we qualify it.
Keep scientific meaning intact.
Preserve identity, units, missing values, axes, coordinate claims and exact revisions. Make unknowns and conversion losses visible.
Make independence practical.
Our requirement is that customers can independently read and reuse their permitted data, including provenance and interpretations. Essential access and usable exit must not depend on renewing our application UI.
Respect the data authority.
Use supported OSDU APIs in OSDU mode. Check permissions at reads, execution and publication. Keep one authoritative backend for each asset.
Own the history, too.
Scientific lineage records how a result was made. Operational diagnostics and optional vendor analytics have different purposes and controls; they do not replace that history.
Make AI use explicit.
Our planned AI workflows use versioned datasets, recorded transformations and customer-approved execution. Possessing data does not automatically confer training rights. No cross-customer training without explicit agreement.
Ownership includes the science behind the data.
Explore how customer-owned derivation history, reproducible AI datasets and controlled diagnostics turn these principles into product requirements.