← Back to Demo Home
CRI Assessor — Deployment Architecture
“How do we use AI without exposing client data?”
Docker Container
Assessment Engine
Control Library
Requirement Library
Profile Engineering
Compatibility Reasoning
Coverage Assessment
Decision Records
Human Review Workflow
Interactive Report
Audit Trail
Local Knowledge Cache
Local AI Model (L1)
Gemma · Ornith · Future Local Models
Semantic reasoning · Profile generation · Compatibility analysis
Customer data never leaves the client environment
Controls · Risks · Policies · Procedures · Evidence
Knowledge
Synchronization
Standards
Taxonomies
Vocabulary
Methodology
Client Controls
Client Risks
Client Evidence
Client Reports
Standards & Frameworks
CRI Knowledge
NIST Mappings
SCF Alignment
Compatibility Rules
Methodology Assets
Controlled Vocabulary
Taxonomy Library
Synonym Library
Decision Tables
Specifications
Procedures
Templates
Prompt Library
Governance
Version Management
Knowledge Promotion
Model-Independent
Contains NO client data
Privacy First
Customer information never leaves the client environment
Model Independent
Any compatible local AI model can execute the methodology
Governed AI
Deterministic rules
↓ Semantic reasoning
↓ Human review
↓ Audit trail
Knowledge Reuse
Knowledge improves centrally. Client data remains local.