Sample Gallery›Defense›D2 — Governance Design Specification

🛡Defense · D2 · Veterans Health ProgramAnonymized — Real Engagement

D2 — Governance Design Specification

Configuration-ready governance framework for military/veteran health readiness intelligence. 5 governance rules · 3-tier authority chain · NASA ERI/HEHI integration · DHA/IMCOM alignment · SHA-256 receipt chain · 6 system integrations. Engineering loads directly from this document.

📄 D2 Deliverable — Governance Design Specification · Engagement ENG-002 · PDIO Phase 2 of 4

3-Tier Authority Chain Design

Military governance requires explicit chain-of-command documentation for every AI-assisted decision. ATLAS-M implements a three-tier authority chain where tier selection is automatic based on readiness score severity:

Tier 1 — Routine Readiness

Primary Care Provider review

Receipt Type

Automated receipt generation

Tier 2 — Elevated Concern

Primary Care Provider + Behavioral Health consultation, dual-reviewer receipt

Receipt Type

Dual-reviewer receipt with consultation record

Tier 3 — Readiness Determination

Medical Officer of Record + Commanding Officer notification

Receipt Type

Full authority chain receipt with time-stamped acknowledgments at each level

Governance Rules

GOV-001

Readiness Assessment Governance

NON-NEGOTIABLE

Trigger Condition

Monthly readiness assessment cycle OR any biometric/behavioral signal crossing critical threshold

Trigger Signal(s)

VH-001 (Fitness Score) + VH-004 (PHQ-9) + VH-005 (PCL-5) + VH-010 (NASA ERI)

Threshold

Composite readiness score < 0.65 OR any single signal in critical range

Primary Authority

AUTH-001: Primary Care Provider

Escalation Authority

AUTH-002: Medical Officer of Record

Response Window

Within 48 hours of trigger

Action

Generate governed readiness assessment with confidence score. Route to Primary Care Provider (Tier 1). Flag for Medical Officer of Record notification if score < 0.50.

Regulatory Basis

DHA AI Policy 2023, IMCOM Vision POC Plan requirements, DoD Instruction 1332.18

Judge Criteria

Is the readiness score computed from all available signals (not partial data)?
Is the confidence level ≥ 0.75 before routing to provider?
Does the recommendation reference specific signal contributors?
Is the authority chain documented for Tier selection?

⚠ All AI-assisted readiness determinations must produce a sealed receipt. No ungoverned AI output permitted after go-live.

GOV-002

Environmental Exposure Integration

Trigger Condition

NASA ERI composite > 0.40 OR PACT Act qualifying exposure flag detected in service record

Trigger Signal(s)

VH-007 (Deployment AQI History) + VH-008 (PACT Act Flag) + VH-010 (NASA ERI Composite)

Threshold

ERI > 0.40 (warning) or ERI > 0.65 (critical) or confirmed PACT Act exposure

Primary Authority

AUTH-003: Environmental Health Specialist

Escalation Authority

AUTH-002: Medical Officer of Record

Response Window

Within 5 business days of trigger

Action

Initiate environmental exposure consultation referral. Update readiness score with ERI contribution. Generate PACT Act documentation package if qualifying exposure confirmed.

Regulatory Basis

PACT Act 2022 (PL 117-168), EPA National Ambient Air Quality Standards, NASA FIRMS data use agreement

Judge Criteria

Is the ERI composite derived from ≥ 2 confirmed data sources?
Is the PACT Act flag based on verified service record data (not self-report alone)?
Does the referral include the specific environmental signals that triggered it?

GOV-003

Behavioral Health Escalation Governance

NON-NEGOTIABLE

Trigger Condition

PHQ-9 ≥ 15 OR PCL-5 ≥ 50 OR AUDIT-C ≥ 8 OR AI-human disagreement flagged by provider

Trigger Signal(s)

VH-004 (PHQ-9) + VH-005 (PCL-5) + VH-006 (AUDIT-C)

Threshold

Any screen in critical range OR provider-flagged AI disagreement

Primary Authority

AUTH-004: Behavioral Health Provider

Escalation Authority

AUTH-005: Chief of Behavioral Health

Response Window

Within 24 hours

Action

Escalate to Behavioral Health Provider within 24 hours. Document AI-human disagreement resolution in governance receipt if applicable. Notify Primary Care Provider.

Regulatory Basis

Joint Commission Behavioral Health Standards, VA National Suicide Prevention Program, DoD MHS Clinical Practice Guidelines

Judge Criteria

Is the escalation triggered by a validated screening instrument (not proxy data)?
If AI-human disagreement: are both the AI recommendation and the clinical judgment documented?
Is the response proportionate (24-hour escalation for critical screens)?

⚠ AI-human disagreement resolution is required by DHA AI Policy 2023. Absence of this protocol was a P0 finding in D1.

GOV-004

Biometric Monitoring Coverage

Trigger Condition

Beneficiary without biometric monitoring data for > 30 days OR biometric signal dropout detected

Trigger Signal(s)

VH-003 (Resting Heart Rate) + VH-001 (Fitness Score)

Threshold

No biometric data for > 30 days OR signal dropout > 7 days

Primary Authority

AUTH-006: Primary Care Coordinator

Escalation Authority

AUTH-001: Primary Care Provider

Response Window

Within 3 business days

Action

Trigger care coordinator outreach for monitoring re-enrollment. Downgrade readiness score confidence rating. Flag for Primary Care Provider review.

Regulatory Basis

IMCOM Vision POC Plan biometric monitoring requirements, DoD Health Affairs guidance

Judge Criteria

Is the dropout attributed to equipment failure (not patient refusal)?
Is the confidence downgrade proportionate to the duration of missing data?

GOV-005

Command Readiness Reporting

Trigger Condition

Weekly reporting cycle OR readiness determination required for duty assignment

Trigger Signal(s)

Aggregate of all active signals — composite readiness score

Threshold

Any beneficiary with composite readiness < 0.65 flagged for command report

Primary Authority

AUTH-002: Medical Officer of Record

Escalation Authority

AUTH-007: Commanding Officer Liaison

Response Window

Weekly cycle (48 hours for urgent determinations)

Action

Generate weekly command readiness summary. Route individual readiness determinations to Medical Officer of Record for Tier 3 authorization. Document chain-of-command acknowledgments.

Regulatory Basis

DoD Instruction 1332.18 (Disability Evaluation System), DHA AI governance requirements for readiness determinations

Judge Criteria

Does the readiness report aggregate all available governance receipts for the period?
Is the Tier 3 authority chain documented for each readiness determination?
Are all unsealed receipts from the reporting period included?

Decision Routing Table

Rule	Tier	Primary Authority	Escalation	Response Window	Auto-Escalation
GOV-001	Tier 1 (routine)	AUTH-001: Primary Care Provider	AUTH-002: Medical Officer of Record	48 hrs	Yes — at 48 hrs
GOV-001 (critical)	Tier 3 (determination)	AUTH-002: Medical Officer of Record	AUTH-007: Commanding Officer Liaison	Immediate	Yes — immediate
GOV-002	Tier 1	AUTH-003: Environmental Health Specialist	AUTH-002: Medical Officer of Record	5 business days	Yes — at 5 days
GOV-003	Tier 2 (elevated)	AUTH-004: Behavioral Health Provider	AUTH-005: Chief of Behavioral Health	24 hrs	Yes — at 24 hrs
GOV-004	Tier 1	AUTH-006: Primary Care Coordinator	AUTH-001: Primary Care Provider	3 business days	Yes — at 3 days
GOV-005	Tier 2–3	AUTH-002: Medical Officer of Record	AUTH-007: Commanding Officer Liaison	Weekly / 48 hrs urgent	Yes — at cycle close

DHA / IMCOM Alignment Documentation

The governance architecture is documented against Defense Health Agency (DHA) AI governance requirements and IMCOM Vision alignment criteria:

✓

DHA Audit Requirements

Every AI output is traceable to source data, model version, and authorization chain — satisfying DHA requirements for auditable AI-assisted clinical decisions.

✓

Human-in-the-Loop Compliance

Confidence scoring with explicit uncertainty disclosure satisfies DHA requirements for human-in-the-loop AI. All recommendations below 0.75 confidence are escalated before reaching the authority.

✓

Receipt Standard Compatibility

The governance receipt format is compatible with DHA's emerging AI governance receipt standard. Receipt chain integrity (SHA-256) and immutability are implemented per DHA guidance.

✓

IMCOM Vision POC Alignment

The biometric monitoring coverage roadmap and readiness signal taxonomy align with IMCOM Vision POC Plan requirements for continuous readiness monitoring.

✓

PACT Act Documentation

GOV-002 generates PACT Act documentation packages for qualifying exposures — creating the structured record required for claims processing and retroactive benefit determination.

Integration Architecture

System	Direction	Protocol	Data	Auth	PHI Handling
EHR (DoD GENESIS / MHS)	Inbound	HL7 FHIR R4	Clinical encounters, labs, medications, screening results	CAC + OAuth 2.0	De-identified before LLM; raw stays in GENESIS
DoD DEERS	Inbound	REST (HTTPS)	Service record, exposure history, PACT Act flags	PKI certificate	Exposure flags only — no PHI in ATLAS layer
NASA Earthdata	Inbound	REST (HTTPS)	FIRMS fire data, environmental exposure indices (ERI/HEHI)	Earthdata token (DN-2469)	No PHI — geospatial data only
EPA AirNow	Inbound	REST (HTTPS)	Real-time and historical AQI readings	API key (free)	No PHI
Biometric Monitor API	Inbound	REST (HTTPS) / Bluetooth LE	Heart rate, activity, sleep — where enrolled	Device pairing + patient consent	De-identified stream
Command Reporting System	Outbound	REST / sFTP	Weekly readiness summaries, sealed receipts	Command-issued cert	De-identified summaries only

Receipt Specification (Defense Variant)

═══════════════════════════════════════════════════════════
  ATLAS-M GOVERNANCE RECEIPT
═══════════════════════════════════════════════════════════
  DECISION ID:     GR-{YYYYMMDD}-{PROGRAM}-{SEQ}
  TIMESTAMP:       {UTC ISO 8601}
  PROGRAM:         Veterans Health Program — {Facility/Zone}
  TRIGGER:         {GOV-XXX: Rule Name}
  READINESS TIER:  {1 / 2 / 3}
  ERI SCORE:       {NASA ERI composite: 0.XX}
  LPRM SCORE:      {Living Patient Risk: 0.XX}
  CONFIDENCE:      {0.XX}
  JUDGE RESULT:    {PASSED / BLOCKED — reason}
  RECOMMENDATION:  {ATLAS-M recommendation text}
  AUTHORITY:       {AUTH-XXX: Role — Title}
  AUTHORITY CHAIN: {Tier 1/2/3 — all acknowledgments}
  HUMAN ACTION:    {What the authority decided}
  RATIONALE:       {Authority's stated reason}
  SHA-256 HASH:    {computed from all fields above}
  RECEIPT CHAIN:   {SHA-256 of previous receipt}
  PATENT REF:      TPP96862
  NASA LICENSES:   DN-2469, DN-2604, DN-2610, DN-2611
  STATUS:          SEALED
═══════════════════════════════════════════════════════════

← D1 — Process Intelligence Brief D3 — Validation & Acceptance Report →

Anonymized · Real Engagement · CROMTEC.AI · Patent TPP96862

See what ATLAS would produce for your organization.

This D2 specification is configuration-ready — engineering loads directly from this document. Start a conversation.

Start a conversation →← Back to samples