Architecture Overview · Enterprise AI Delivery

Enterprise Cloud Migration Analysis Pipeline

Multi-agent AI pipeline that processes an enterprise application portfolio through 7 specialized agents, each governed by deterministic validation rules, to produce auditable 6R cloud migration recommendations.

● Live Deployment · Healthcare Client AWS Bedrock · Claude Sonnet 4.6 High-Governance Design

Click any component to explore it in the full case study.

Input A
RVTools vCenter Export
VM inventory: CPU, RAM, storage, power state, OS across 408 virtual machines across 36 clinical applications
CSV · VMware Telemetry
Input B
Application Assignment Sheet
App owner, business criticality, life-safety classification, regulatory scope, and known dependency mapping
Excel · Structured Metadata
Agent Pipeline — 7 Specialized Agents per Application
Step 0
Telemetry Enrichment
Joins vCenter export with app assignment sheet
›
Agent 1
Telemetry
CPU/RAM utilization profile + right-sizing heuristics
›
Agent 2
Dependency
Migration risk, hardware deps, life-safety classification
›
Agent 3
Procurement
SaaS alternatives, vendor intelligence, contract signals
›
Agent 4
Provisioning
LLM classifies workload; Python engine does all math
›
Agent 5
Synthesizer
Final 6R recommendation with confidence gate
›
Agent 6
Confidence Advisor
Prioritized backlog to close evidence gaps
›
Agent 7
Portfolio Narrative
Executive-grade prose summary across all apps
AWS Bedrock · Claude Sonnet 4.6 · Parallel execution across 36 apps
4.9h total runtime · ~$24 total AI compute
🔒
Deterministic Governance Layer — Applied after every LLM call
No LLM output reaches the next agent without passing all relevant rules. Every recommendation is traceable to its source data by design.
22 validation rules HIPAA scope enforcement Life-safety veto gates Schema validation Confidence scoring Provenance tracking Confidence < 0.7 → escalation Audit trail per app
Output A
Per-App 6R Decisions
Rehost, Replatform, Repurchase, Refactor, Retain, or Retire — each with confidence score, evidence chain, and gap backlog
36 apps · Ranked by confidence
Output B
Executive Summary Report
Portfolio-level narrative synthesized across all applications. Migration complexity distribution, risk heatmap, and prioritized initiative roadmap
HTML · Stakeholder-ready
Output C
Full Analysis Report
App-by-app deep dives with agent reasoning, VM telemetry tables, procurement alternatives, and audit trail for every recommendation
Live Demo Available
36
Clinical apps analyzed
408
VMs in scope
$24
Total AI compute cost
22
Governance rules enforced
Stack AWS Bedrock Claude Sonnet 4.6 Python boto3 Pydantic validation S3 static hosting Parallel execution