Deployment Evaluator
AI is not gap-neutral. Absent the right preconditions it concentrates advantage where capacity already exists — so the same tool can extend rural care or quietly hollow it out. Score a specific deployment across three forms of leverage and three preconditions, and read where the needle lands.
Everything sits at "partial" — the honest default. Score the cells below to move the needle.
Rows: what the AI does · Columns: what must be true to capture it
Can this site actually adopt it at all?
Governed, or just installed?
Who holds the risk when it fails?
Administrative automation with no accountability structure is the denials-engine pattern — the Lokken / Kisting-Leung fact set. A deployment scored this way doesn't just fail to close the gap; it actively widens it, because the contesting party is the one with the fewest alternatives.
The leverage isn't in the model — it's in the governance and infrastructure layer that decides who the model serves.