Defense + Government · Fed Supernova 2026
Decision advantage is changing course together.
Grids turns changing facts, dependencies, constraints, and policy into one inspectable model for coordinated, accountable decisions. This complete web brief develops the model, cases, evidence, and evaluation path.
01 · The coordination gap
The facts can be current. The decision can still be stale.
Operational plans, logistics assessments, policy, and authority often move on different clocks. People reconcile them in exports, briefings, and private spreadsheets after a consequential fact has already changed.
Revision snapshot
Operations
Plan Echo · as of 06:12
Logistics
5 of 6 · as of 06:18
Command
Green · as of 06:12
Documented institutional pattern
GAO reported that planners gathered operational and logistics information separately, then combined it in documents and slides to judge whether a course of action was feasible. A later experiment integrated two systems so planners could see logistics problems and adjust the course in real time.
Source: GAO-25-106454, Defense Command and Control
The central question is: what authoritative model sits behind the common operating picture and keeps the decision current as its source facts change?
02 · The category
A picture shows the state. A model shows what the state means.
| Common operating picture | Common operating model |
|---|---|
| Answers: What is happening? | Answers: Can the plan work? |
| Current status and overlays | Facts and dependencies |
| Alerts and summarized indicators | Constraints, uncertainty, and policy |
| Useful views for different roles | Bounded alternatives with inspectable reasons |
| A picture does not encode the decision logic. | Attributable human decision evidence |
Grids is proposed as executable decision infrastructure: authored mission or policy logic, evaluated by a common runtime, available through the interfaces each role needs.
The interface can change without rebuilding the meaning of the decision: one model state, visible logic, and multiple role-specific views.
03 · How it works
From one changed fact to an accountable decision.
- 01Configured factA source revision enters with identity and provenance.
- 02Affected logicOnly dependent calculations need to update.
- 03Bounded choicesAuthored alternatives are evaluated inside explicit constraints.
- 04Reason visibleEvidence, assumptions, status, and model revision remain inspectable.
- 05Human decisionA person authorizes; selected views and integrations receive the revision.
Grids
Model, recalculate, evaluate, and explain.
Integration
Source authority, identity, and delivery receipts.
People + organization
Policy, role authority, and operational action.
A model makes its inputs, dependencies, and consequences inspectable so people can challenge the evidence and understand the decision before acting.
The Grids ecosystem
Continue into the working system.
Run Grids in the cloud, browse signed packages in the public namespace, or read the developer documentation.
Case 01 · 72-second film
Every system works. The system does not.
Documented pattern
GAO described operations and logistics being gathered separately, then reconciled in documents and slides to assess course-of-action feasibility.
The film
One logistics fact changes. Three views remain valid for different revisions. A configured binding updates one decision model; a human records the external decision.
Measure
Time to revised feasibility, manual reconciliation steps, conflicting revisions, synchronized assessments, and rejected options with an inspectable reason.
The film makes the configured branches, evidence revision, human decision, and downstream state visible as one measurable workflow.
Case 02 · 42-second film
Inventory is not availability.
A part has value only when identity, location, condition, demand, route, and release evidence make it usable in time.
Documented pattern
GAO documented backlogs, duplicate requisitions, inadequate visibility and prioritization, and systems that were not fully interoperable. Later integration gave thousands of users near-real-time shipment and stock visibility.
Grids proposition
Author identity, condition, freshness, priority, and provenance. Match demand to eligible stock. Test a bounded allocation and route. Present the evidence and margin for human authorization.
Grids keeps source data, route constraints, solver logic, and command authority explicit throughout the decision.
Sources: GAO-04-305R, GAO-05-775, and GAO-12-138 · Watch the film and read its text version
Case 03 · 48-second medical-planning film
The medical plan must move before the next patient does.
This medical-planning scenario compares alternatives across changing demand, capacity, transport, and time while qualified personnel define the clinical semantics, review the evidence, and authorize action.
Planning state
Changed inputs
Surge +6 · window closed · capacity 8 → 3
Capacity view
North 5 · Central 3 · South 5 · Reserve 3
Evaluated result
Alt A: capacity exceeded · Alt B: window + capacity failed · Alt C: feasible within bounds
Research basis
A 2025 peer-reviewed proof of concept reported limits in static casualty estimates and a need for predictive evacuation and medical-resupply planning.
Decision evidence
The model compares authored alternatives and exposes the supporting evidence and uncertainty for review.
Sources: PubMed 40523074, Automated Battlefield Trauma System and Joint Trauma System journal watch
04 · Trust by design
Move faster without hiding the authority boundary.
AI with human authority
One explicit model and one governed path: AI authors, a compiler validates, a runtime executes, and a human authorizes.
Coalition interoperability
Local models publish explicitly configured, typed, revisioned outputs. Signed packages and governed interfaces can coordinate logic without automatically publishing every local detail.
Public priorities reinforce the need for designed interoperability: Open DAGIR emphasizes government-owned interoperable repositories and a multivendor ecosystem; GAO's AI framework centers governance, data, performance, and monitoring; NATO's FMN and CWIX demonstrate why interoperability must be designed before a mission.
Deployment teams define and verify releasability, privacy enforcement, interface accreditation, and approved cross-domain transfer for each operating context.
05 · Evaluation
Proof starts with one bounded workflow.
Bring one changing fact, one hard constraint, and one accountable decision. Establish the manual baseline first, then compare the modeled workflow.
Good first evaluations
Course-of-action feasibility; asset visibility; maintenance readiness; medical surge.
Measure
Manual baseline time; changed input to revised conclusion; stale or contradictory states detected; options with explicit status; explanation completeness; human decision evidence and downstream acknowledgment.
Supported architecture
Incremental graph; bounded solvers; predicates and WHY; selected exchange; local and server modes.
Prove in context
Domain-model validity; source integration; latency; usability; downstream behavior; mission impact.
Operational readiness
ATO and accreditation; cross-domain behavior; DDIL synchronization; field integration; operator acceptance.
Evidence standard: Measure the baseline and the modeled workflow to establish performance and mission impact.
Start with one decision
Bring the workflow that cannot wait for the next briefing.
A first conversation can map the facts, dependencies, constraints, authority boundary, and measurable baseline, then define what a common operating model would need to prove.
Primary sources
- GAO-25-106454 · Defense Command and Control
- GAO-04-305R · Logistics During Operation Iraqi Freedom
- GAO-05-775 · Defense Logistics
- GAO-12-138 · Afghanistan Supply and Distribution
- PubMed 40523074 · Automated Battlefield Trauma System
- Open DAGIR · Government-owned interoperable repositories
- DoD AI Rapid Capabilities Cell · Priority use cases
- GAO-21-519SP · AI Accountability Framework
- NATO · Federated Mission Networking
- NATO · CWIX 2026