Pilot program
Ninety days. Named KPIs. No invented outcomes.
A VetBattleBuddy.ai pilot is a bounded evaluation of evidence readiness — not a production claims system and not a promise of a rating, a timeline, or a dollar savings on the public site.
Shape of ninety days
Scope. Operate. Measure. Decide.
Days 1–15
Scope and controls
Confirm the workflow, the human override path, the data boundary, and the measures that will count.
Days 16–45
Operate the bounded workflow
Run the educational readiness layer against the agreed population. Authorized humans remain in command.
Days 46–75
Measure
Capture first-pass yield, gaps named pre-submission, rework, traceability, and override. No public invention of results.
Days 76–90
Decide
Keep, expand, or stop. Publication happens only when the evaluation can be independently defended.
We publish the measurements we intend to prove. We do not invent the results.
Pilot results will be published when they are independently defensible. Until then, this is the contract.
| Measure | Use |
|---|---|
| Evidence First-Pass Yield™ | Primary |
| Missing evidence detected pre-submission | Primary |
| Evidence-related rework | Primary |
| Traceability of assisted outputs | Primary |
| Human override rate | Governance |
| Unsupported-output rate | Governance |
- Not a production claims-adjudication system
- Not a promise of a rating, award, or timeline
- Not a public dollar-savings claim
- Human override is mandatory
- Results stay unpublished until they are defensible
EFPY planning board
Illustrative target bands — not published company results
Replace these bands with pilot measurements when a bounded evaluation is complete. Values are planning placeholders so operators can see the board shape. They are not Evidence First-Pass Yield operating results.
Diagnosis documented
80%
Service-connection theory identified
70%
Medical evidence on hand
75%
Nexus / supporting statements
45%