OEMRA / Maintenance Integrity
Operational & Entropy Model Risk Assessment for stack STK-2024-MRC-001
Medium Risk
View AISL Impact →
OEMRA SNAPSHOT — 12/28/2024
MII
78/100
Maintenance Integrity Index
SAR
72.0%
System Autonomy Ratio
EDF
0.910
Entropy Decay Factor
Entropy Impact
9.0%
Drift/Decay Signal
Maintenance Burden
24 hrs/mo
Autonomy Ratio
72.0%
Overall Risk
MEDIUM
78
MII Score
Good standing
OEMRA-Adjusted KOV
$2.58M
vs $2.84M base
OEMRA FORMULA
OEMRA_KOV = KOV × (MII/100) × EDF
KOV
Base Stack Liquidity Value
$2.84M
MII
Maintenance Integrity Index ÷ 100
0.78
EDF
Entropy Decay Factor
0.910
OEMRA-Adjusted KOV
$2.58M
Discount: $256.0K (9.0% reduction from base KOV)
COMPONENT HEALTH
Primary Inference Engine
OpenAI GPT-4o
Complex Reasoning Engine
Anthropic Claude 3.5
Cost-Optimized Classifier
Mistral Mixtral-8x22B
Knowledge Retrieval Layer
Pinecone p2 Index
Data & Analytics Platform
Snowflake Cortex
Code Intelligence Layer
GitHub Copilot
Revenue Ground Truth
Stripe Revenue API
DRIFT FLAGS
3 flags detected
Cost Inflation — openai/gpt-4o
Token cost increased 18% vs baseline; no model change — likely prompt expansion in Revenue Ops Copilot
medium
Detected: 12/20/2024
Error Spike — wf-003/support-automation
Error rate increased from 2.1% to 3.1% on support workflow; likely ticket volume surge
low
Detected: 12/22/2024
Latency Degradation — pinecone/p2-index
p99 latency increased from 38ms to 45ms; vector index size growth — re-indexing recommended
low
Detected: 12/18/2024
RISK FACTORS & MAINTENANCE BREAKDOWN
Model Drift Risk
18%
Cost inflation detected on primary inference engine
Latency Degradation
12%
P99 latency increasing on vector retrieval layer
Error Rate Trend
9%
Error rate within acceptable bounds; minor spike on support workflow
Dependency Concentration
68%
OpenAI concentration at 92 dep score — single-vendor risk
Autonomy Override Risk
28%
2 of 5 workflows without human oversight
Evidence Freshness
4%
96% of evidence records verified; all within 24h