Tools organized by Data Sovereignty Zone
The Core Rule: Public tools get public data. Enterprise tools get internal data. Secure enclaves get PHI. Never cross the streams.
For publicly available information, literature synthesis, and creating your "Portable Context."
Google's source-grounded tool. I use this to "interrogate" PDFs and generate first-draft slides/visuals. Constraint: No PHI.
Consumer Voice AI. Great for dictating "Context Files" while walking. Note: Consumer cloud—no sensitive data.
Replaces Google for 90% of my queries. Provides citations for every claim.
For grant drafts, unpublished manuscripts, and sensitive logic. Data must not train public models.
Duke-licensed. Safe for internal memos and email triage. Copilot Notebooks provides source-grounded RAG (like NotebookLM)—won't pull from open web, only your uploaded documents. Word Mobile Dictation offers secure, structured dictation with custom instruction modes.
Duke's AI-powered grant proposal reviewer. Checks compliance, flags foreign components, validates citations, and provides "red team" feedback. All processing on Duke servers—never transmitted externally.
Run models entirely on your machine. Air-gapped capable. Your data never leaves your device.
The sandbox for building internal tools before they go to production.
For Protected Health Information (PHI). Secure enclaves only. No consumer AI tools.
Where I get my signal in a noisy world. These are public resources (Green Zone) for staying current without drinking from the firehose.
Ready to apply the framework?
← Back to The Horizon Mandate