Oversee every operating AI across your force—confidently and at speed
A single system of record, guardrails, and stage-gates for internal and vendor AI—from ISR to autonomy to GenAI.
Built for IL5/IL6 deployments. No operational payloads ingested.
“AI is shipping faster than governance—fragmented records, mixed vendors, and audit gaps.”
Visibility
One Operating AI Record (including purpose-level modules)
Controls
Guardrails, scorecards, and stage-gates tied to policy
Collaboration
Multi-role approvals, oversight packets, and vendor wraps
How it works
Register
Create/ingest Operating AI Records: purpose, owner, locations, sourcing. Normalize internal factories (Linchpin, TF Hopper, Space C2) and vendor systems (TITAN, Maven, PANDA, Anduril CUAS).
Category playbooks
KPIs
- AP per class
- False-positive budget
- Cueing latency
- Bias slice by target-set/geo
Scenario tests
- Urban/night/weather
- Sensor-fusion ablations
Guardrails
- Labeling provenance
- Target-misclassification recourse
Stage-gate
- Regression pass %
- Drift delta
- Escalation path
Org-specific quick-starts
Army
- Link Linchpin registry
- Wrap TITAN & Vantage telemetry
- FIRESTORM scenario packs
Show integration details
Integrate mission packages; onramp brigade-level approvals; export to RMF.
Navy
- Ingest TF Hopper artifacts (SPY GPT/OpenShip) via Flank Speed
- Overmatch gate rules
- TF-59 vendor wraps
Show integration details
CAC/PIV + ABAC; watchstanding guardrails; export to warfare center workflows.
USMC
- Site-level IC-sUAS registrations
- Lattice telemetry → guardrails
- Range validation packs
Show integration details
Unit-level approvals; rollbacks; after-action evidence threads.
USAF
- PANDA KPIs via API
- ACE test evidence → safety case library
- Mx ops guardrails
Show integration details
Tie to wing-level boards; SLA-backed vendor integrations.
USSF
- Index UDL + Space C2 modules
- Track-custody scorecards
- Alerting guardrails
Show integration details
Catalog provenance checks; custody continuity metrics; SDA playbooks.
NGA
- Maven workflows as separate purposes
- Target-class bias checks
- Evidence locker export
SOCOM
- Mission-package autonomy approvals
- Rapid rollback plans
- Adversarial behavior sweeps
Show integration details
Vendor wraps for deployables; geo-fencing and human override latency SLAs.
USTRANSCOM
- CAITT/MASCOT guardrails
- Clinical/operator sign-offs logged
- Fleet readiness KPIs
Show integration details
Evidence chains to medical/ops authorities; export to logistics boards.
Security & compliance
- Hosting: IL5/IL6-ready, zero-trust, CAC/PIV + ABAC.
- Data handling: markings (CUI/FOUO/SECRET), WORM evidence store, cross-domain guard compatibility.
- Scope: Overlook stores evidence & governance state, not mission payloads.
Controls & policy alignment
- Policy alignment: DoD AI Principles, NIST AI RMF, autonomy/ISR/GenAI annexes.
- Change safety: kill/rollback, warn-only switches, staged approvals.
FAQ
How do you integrate without ingesting mission payloads?
Overlook wraps systems and captures evidence/telemetry, not payloads; it stores governance state and artifacts.
What’s needed to pilot at IL5/IL6?
We deploy to approved environments, integrate with CAC/PIV + ABAC, and align to RMF/ATO processes.
How do vendor wraps work if we already have ATOs?
Vendor wraps add guardrails and evidence without replacing ATOs, preserving authority boundaries while improving safety.
Can we register purpose-level modules separately?
Yes. The Operating AI Record supports purpose-level modules with their own KPIs, risks, and approvals.
Do you provide category test templates?
Yes. Each playbook tab includes a download link to example test templates.
How does oversight evidence export into RMF artifacts?
One-click export packages include traceable evidence mapped to RMF controls and can be attached to ATO bodies of evidence.
References to government organizations and commercial systems illustrate compatibility and public sources; no endorsement implied.
Military
Top operating AIs
- ISR processing & target recognition assist
- Predictive logistics & maintenance (air/land/sea)
- Course‑of‑action analysis, wargaming & simulation
- C2 decision support & battlespace awareness
- Autonomy‑assist & human‑machine teaming (non‑lethal logistics/ISR focus)
Impact focus
Readiness, sortie/mission availability, accelerated planning cycles, sustainment efficiency, improved situational awareness.
Solution playbook
Audience
J‑codes, Program offices, Sustainment/logistics, Intel.
Operational areas
ISR processing, C2 decision support, Logistics & maintenance, Simulation/wargaming, HMT (non‑lethal ISR/logistics).
Key behaviors
Detection/ID confidence; latency; COA recommendation traceability; mission availability predictions; autonomy boundary adherence.
KPIs
Readiness ↑, mission availability ↑, planning cycle time ↓, sustainment cost ↓, decision advantage ↑.
Risks & controls
ROE & policy compliance; data classification; contested comms; model drift under novel conditions; human‑on‑the‑loop.
Overlook impact
Classification‑aware Operating Records; ROE guardrails; theater‑level Impact Scores; after‑action learning loops.
30/60/90
(30) Map programs/models; (60) Connect ISR/sustainment feeds; (90) Publish mission‑readiness Impact Scores and COA governance.