AI Capability Sandbox
Synthetic rehearsal · not a real patient, not a real approval
← AI Fleet Control
K
AI Capability Sandbox
Rehearse an AMBER-tier worker against fabricated cases before completing its pre-deployment risk memo on the staging gate — so that memo is an informed decision, not a checkbox exercise. Nothing on this page is real, and nothing it produces can be sent, saved, or approved.
1. Choose a worker to preview
Only workers currently waiting on a staging-gate decision are listed here
Lab Summarizer turns a result set into a plain-language summary for the patient — reference ranges and trend, no interpretation.
2. Pick a synthetic case
Generated for this rehearsal only — not a real person, not drawn from any real patient's chart
3. What it would draft
A preview only — never sent, never saved, never logged anywhere a real draft would be
✦
Lab Summarizer
Synthetic Patient S-1
Select a case above to preview a draft.
This preview never touches a real patient record, is never sent to anyone, and leaves no trace once you close it — unlike a real draft from this worker, which is always logged whether approved or rejected.