'%3e%3cpath%20fill-rule='evenodd'%20clip-rule='evenodd'%20d='M91.1446%2016.9537L71.4456%2025.0803L25.8825%2075.569L19.8136%2095.996L3.32812%2081.1189L23.0271%2072.9923L68.5902%2022.5036L74.6592%202.07666L91.1446%2016.9537Z'%20fill='%23454545'/%3e%3cpath%20d='M23.1344%2011.5672C23.1344%2017.9555%2017.9556%2023.1343%2011.5672%2023.1343C5.17881%2023.1343%200%2017.9555%200%2011.5672C0%205.1788%205.17881%200%2011.5672%200C17.9556%200%2023.1344%205.1788%2023.1344%2011.5672Z'%20fill='%2318181B'/%3e%3cpath%20d='M57.4625%2011.5672C57.4625%2017.9555%2052.2836%2023.1343%2045.8953%2023.1343C39.5069%2023.1343%2034.3281%2017.9555%2034.3281%2011.5672C34.3281%205.1788%2039.5069%200%2045.8953%200C52.2836%200%2057.4625%205.1788%2057.4625%2011.5672Z'%20fill='%2318181B'/%3e%3cpath%20d='M92.5367%2011.5672C92.5367%2017.9555%2087.3579%2023.1343%2080.9695%2023.1343C74.5811%2023.1343%2069.4023%2017.9555%2069.4023%2011.5672C69.4023%205.1788%2074.5811%200%2080.9695%200C87.3579%200%2092.5367%205.1788%2092.5367%2011.5672Z'%20fill='%2318181B'/%3e%3cpath%20d='M57.4625%2049.6268C57.4625%2056.0151%2052.2836%2061.1939%2045.8953%2061.1939C39.5069%2061.1939%2034.3281%2056.0151%2034.3281%2049.6268C34.3281%2043.2384%2039.5069%2038.0596%2045.8953%2038.0596C52.2836%2038.0596%2057.4625%2043.2384%2057.4625%2049.6268Z'%20fill='%2318181B'/%3e%3cpath%20d='M92.5367%2049.6268C92.5367%2056.0151%2087.3579%2061.1939%2080.9695%2061.1939C74.5811%2061.1939%2069.4023%2056.0151%2069.4023%2049.6268C69.4023%2043.2384%2074.5811%2038.0596%2080.9695%2038.0596C87.3579%2038.0596%2092.5367%2043.2384%2092.5367%2049.6268Z'%20fill='%2318181B'/%3e%3cpath%20d='M23.1344%2088.4328C23.1344%2094.8212%2017.9556%20100%2011.5672%20100C5.17881%20100%200%2094.8212%200%2088.4328C0%2082.0445%205.17881%2076.8657%2011.5672%2076.8657C17.9556%2076.8657%2023.1344%2082.0445%2023.1344%2088.4328Z'%20fill='%2318181B'/%3e%3cpath%20d='M92.5367%2088.4328C92.5367%2094.8212%2087.3579%20100%2080.9695%20100C74.5811%20100%2069.4023%2094.8212%2069.4023%2088.4328C69.4023%2082.0445%2074.5811%2076.8657%2080.9695%2076.8657C87.3579%2076.8657%2092.5367%2082.0445%2092.5367%2088.4328Z'%20fill='%2318181B'/%3e%3c/g%3e%3cdefs%3e%3cclipPath%20id='clip0_1831_9300'%3e%3crect%20width='93'%20height='100'%20fill='white'/%3e%3c/clipPath%3e%3c/defs%3e%3c/svg%3e)
AI Insights Module · Governance
Insight Feed Configuration
AI Generated Summary
342 insights were generated across all AI models today, of which 12% (41 insights) were automatically suppressed by your current confidence and severity thresholds. Notification routing delivered 96 insights to department leads within SLA. Recommendation: the Fraud Detection v4.1 feed is generating high-confidence insights that are currently filtered out for the Finance team — consider lowering their severity threshold.
Explain This AI Summary
Model
Insight Feed Orchestrator v1.6 (rule-based filtering over 6 upstream models)
Data Sources
Churn Prediction v3.2, Fraud Detection v4.1, Revenue Forecasting v2.7, Demand Planning v2.3, Customer Segmentation v1.9 and Sentiment Analysis v1.4, aggregated over the last 24 hours.
Why 94.2% confidence
Suppression decisions were cross-checked against 30 days of analyst feedback; false-suppression rate over that window was 1.1%, giving the current threshold set high reliability.
Suggested Action Rationale
Fraud Detection v4.1 insights routed to Finance are filtered at an 80% confidence threshold, but 14 suppressed insights this week scored above 85% — lowering the threshold to 75% would surface them without materially increasing noise.
Insights Today
342
Suppressed
41
Active Rules
6
Feed Sources
6
Insights Surfaced vs. Suppressed
Last 14 days
Surfaced Today
301
Suppressed Today
41
False-Suppression Rate
1.1%
Insight Rules Engine
Rules are evaluated in order · toggle to enable or disable
Suppress insights below 70% confidence
Applies to all models · low-confidence insights are archived, not deleted
Notify #data-science-team for Critical severity anomalies
Triggers immediate notification · bypasses digest batching
Auto-archive resolved insights after 30 days
Keeps the active feed focused on unresolved items
Deduplicate near-identical insights within 24 hours
Merges repeated model outputs into a single feed item
Batch Medium severity insights into a daily digest
Delivered at 8:00 AM local time to department leads
Require dual model agreement for Fraud insights
Fraud Detection v4.1 insights must be corroborated before surfacing
Notification Routing
Sarah Chen · Data Science Lead
Fraud Detection, Anomaly & Churn insights
Marcus Lee · Revenue Operations
Revenue Forecasting & Demand Planning insights
Priya Patel · Finance
Fraud Detection insights above $10K exposure
Dana Ruiz · Customer Success
Customer Segmentation & Sentiment Analysis insights
Jordan Kim · Product Analytics
Low-severity product usage insights, weekly digest only
Ana Rossi · Marketing
Customer Segmentation shifts affecting campaign targeting
No routing rules match this severity filter.
Confidence & Severity Thresholds
Insights that fall below these thresholds are suppressed from the live feed
Insights scoring below this confidence are archived automatically
Insights below this severity level stay in the archive only
Caps how many insights appear in a single digest notification
Near-duplicate insights within this window are merged
Feed Sources
Models currently feeding the insight pipeline
Churn Prediction v3.2
62 insights today · 96.1% accuracy
Revenue Forecasting v2.7
48 insights today · 97.8% accuracy
Fraud Detection v4.1
57 insights today · 98.4% accuracy
Customer Segmentation v1.9
39 insights today · 94.6% accuracy
Demand Planning v2.3
44 insights today · 95.3% accuracy
Sentiment Analysis v1.4
Disabled · paused pending model review