Prove one state change
through Runtime revalidation and evidence.
The first PoC does not replace C2, Digital Twin or the AI model. It connects one approved T0 state, changes it to T1, measures Runtime revalidation, and returns a review state plus verifiable evidence to the existing human review layer.
CREAIM attaches as an independent Runtime Validation & Decision Assurance layer.
Start with one representative mission.
A narrow PoC gives both teams a measurable baseline. Current/simulated state and predicted future information may be supplied by the partner; CREAIM revalidates only the agreed information boundary and returns evidence for review.
Route / asset / constraint review
Surveillance / response-option review
Threat / resource-allocation review
Multi-domain state review as follow-on
Detailed integration material is not published on the public site.
Detailed schema mapping, interface contracts, test vectors, security profiles, implementation limits, conformance details, target measurements and delivery artifacts are released only after approval under the Partner/NDA process. Core implementation, scoring, reduction logic, solver strategy and MetaCore remain CREAIM Internal.
Agree the pass criteria before the test.
Proposed gates, not achieved results: (1) approved synthetic/de-identified input and baseline, (2) T0→T1 change trace, (3) per-run Run ID and input/result evidence integrity, (4) fail-closed HOLD on missing or inconsistent evidence, (5) partner-agreed latency and repeatability thresholds, (6) expert review and human authority. Target hardware, real data and live integrations require separate approval and measurement.
Prove Digital Twin state change under controlled conditions
T0 → T1 RUNTIME → VERIFIED REPORT. Current-Run SHA-256 and a reproducible replay receipt are proposed acceptance criteria, not claims of live partner delivery, independent Core execution or military field certification.
Proceed from interface agreement to measurement.
Do not begin with all military data.
Initial work can use de-identified, synthetic or specifically authorized test data. The exact partner schema is agreed privately during the PoC design review.
Judge the PoC with measured evidence.
Input change → Runtime result
Measure only the affected Runtime scope
Detect critical constraint violations
Compare with military expert judgment
Replay result for identical approved input
Verify result and evidence integrity
Observe controlled behavior during network loss
Confirm final authority remains human