Where the read takes place
Why access, licences and processing location are settled before the first analytical read.
Begin with the boundary.
Conversations about analytical intelligence often start with what a model can see. I believe the more useful starting point is what an institution is prepared to let it see.
Every source brings licence terms, access rules and, in some cases, restrictions on where it may be processed. Those conditions need to shape the design from the outset. They are part of the brief, alongside the analytical question the institution wants to answer.
Let the location follow the obligation.
Processing location may look like an infrastructure choice. For an institution, it also determines how the proposed deployment fits its governance requirements.
KADE’s approach is to define access, licences and processing locations for each engagement. My view is that those decisions should be explicit before the first analytical read: which sources are permitted, where processing may occur and how the resulting analysis will be reviewed.
Keep the read inside the workflow.
The architecture is designed to return the configured read into the institution’s existing workflows. Its systems of record remain authoritative, and the decision remains with the people responsible for it.
The contributing factors, configuration and version stay with each read. That gives reviewers a basis for examining how the analysis was produced, alongside where it took place.
For me, deployment starts by making the institution’s existing commitments explicit. The technology then has a clear boundary within which to work.
Explore deployment ↗An explainer based on KADE's published architecture and build-state information.