Appearance
Government & Defense Demo
GhostAI for federal agencies, defense contractors, and government IT teams. This demo follows Colonel James Mitchell, a program manager overseeing classified and CUI-bearing programs, as he uses AI for operational summaries, ITAR export review, document classification, personnel provisioning, and FedRAMP compliance posture — all with CUI and classified content automatically detected and protected.
Pre-Demo Checklist
- [ ] GhostAI instance running
- [ ] Industry vertical set to Government
- [ ] CUI and ITAR detection profiles enabled
- [ ] SCIM provisioning panel accessible
- [ ] FedRAMP compliance module active
Demo Script
Chapter 1: CUI Detection
Duration: 3 minutes Story: Colonel Mitchell summarizes a program update for the weekly brief. The content is marked CUI//SP-CTI and contains radar cross-section data, propulsion specs, clearance levels, badge numbers, budget figures, and COMSEC references.
Input:
Summarize this program update for the weekly brief:
CUI//SP-CTI
Project THUNDERBOLT phase 3 testing at Edwards AFB shows radar cross-section of 0.0015 m² for the Block II prototype, exceeding threshold requirement of 0.005 m². Propulsion integration on track — GE F414-EPE delivering 26,000 lbf in afterburner config. Program lead: Lt. Col. Sarah Mitchell, clearance TS/SCI, badge #JM-28471. Next milestone: IOT&E Phase 1 at Nellis AFB, scheduled 15 SEP. Budget execution at 94.2% ($847M of $899M obligated). COMSEC material transfer pending with NSA — reference EKMS account #4847.Expected Output:
- CUI markings detected: CUI//SP-CTI banner
- Sensitive data flagged: RCS values, propulsion specifications, clearance level, badge number, EKMS account
- Personnel PII detected: name, clearance level, badge number
- Summary generated with sensitive specifics tokenized
Talking Points:
"CUI markings, radar cross-section data, clearance levels, EKMS account numbers — GhostAI detected the CUI banner and every sensitive element within it. The program summary stays useful for the weekly brief, but the specifics that make it controlled are never exposed to a commercial AI model. That's NIST SP 800-171 in practice."
Chapter 2: ITAR Export Control
Duration: 3 minutes Story: Colonel Mitchell evaluates export control implications for sharing a technical data package with a NATO ally. The content references ITAR Category XI military electronics, radar specifications, and ECCM capabilities.
Input:
We need to evaluate export control implications for sharing a technical data package with our NATO partner. The package contains specifications for the AN/APG-83 AESA radar system including waveform parameters, detection ranges, and ECCM capabilities. This falls under ITAR Category XI (Military Electronics). The receiving country is a NATO Tier 1 ally with an existing FMFA. What approvals do we need and what must be redacted from the TDP?Expected Output:
- ITAR content detected: Category XI reference, military system nomenclature, technical specifications
- Export control analysis: DDTC licensing requirements, TAA vs DSP-5 pathway
- Redaction guidance: ECCM parameters, waveform specifics, detection range envelopes
Talking Points:
"ITAR Category XI, AN/APG-83 nomenclature, ECCM capabilities — GhostAI flagged every export-controlled element. The analysis maps the approval pathway (TAA with DDTC, not DSP-5 for technical data) and identifies exactly what must be redacted from the TDP. This is the kind of guidance that normally requires a three-week review from your export compliance team."
Chapter 3: Classification Detection
Duration: 3 minutes Story: Colonel Mitchell reviews an operations memo for classified or sensitive content. The memo contains unit designations, deployment details, SIGINT references, ROE information, and SIPR email addresses.
Input:
Review this operations memo and flag any classified or sensitive content:
MEMORANDUM FOR RECORD
Subject: Deployment logistics — Operation SENTINEL REACH
3rd Battalion, 75th Ranger Regiment will deploy from Hunter AAF to FOB Cobra, AO North, NLT 0300Z 18 MAR. Advance party of 24 PAX under CPT Rodriguez departs 14 MAR via C-17 (Tail #04-4136). Main body follows in two chalks.
SIGINT assets from 780th MI Brigade will be pre-positioned. ROE is IAW CJCS Standing ROE with theater-specific modifications per FRAGO 0847. Personnel list attached (PII redaction required before distribution). POC: MAJ Tara Williams, DSN 312-555-0847, SIPR: t.williams@socom.smil.milExpected Output:
- Classification indicators: SIGINT reference, unit movement details, ROE specifics, SIPR domain
- Personnel PII: names, DSN phone number, SIPR email
- Operational security flags: deployment dates, aircraft tail number, troop strength
Talking Points:
"Unit movements, SIGINT assets, ROE references, SIPR email addresses — any one of these could constitute a classification event. GhostAI flagged every element that could push this memo above the unclassified level. For a program manager, this is the difference between a routine email and an OPSEC violation."
Chapter 4: Personnel Provisioning
Duration: 2 minutes Story: Colonel Mitchell reviews the SCIM provisioning interface — automated user lifecycle management integrated with government identity systems.
Input:
(No chat input — this chapter navigates Users and SCIM admin panels)Expected Output:
- User management interface with role-based access controls
- SCIM integration status: connected identity provider, sync status
- Personnel lifecycle events: onboarding, role changes, offboarding
Talking Points:
"In government environments, user provisioning is non-negotiable — SCIM integration means personnel changes flow automatically from your IdP. When someone's clearance changes or they PCS, their GhostAI access adjusts without a helpdesk ticket. That's NIST 800-53 AC-2 in action."
Chapter 5: FedRAMP Control Mapping
Duration: 2 minutes Story: Colonel Mitchell opens the compliance dashboard to review FedRAMP Moderate control mappings and the audit log for all CUI-related interactions.
Input:
(No chat input — this chapter navigates Compliance and Audit Log admin panels)Expected Output:
- FedRAMP Moderate control mapping: AC, AU, SC, SI control families
- CMMC Level 2 alignment indicators
- Audit trail: CUI detection events, ITAR flags, classification warnings
Talking Points:
"FedRAMP Moderate requires 325 controls across 17 families. This dashboard maps every GhostAI capability to the specific controls it satisfies — access control, audit, system protection, information integrity. When your AO asks for your control inheritance matrix, this is where it lives."