“The effortless configurability of digna enables its use as a self-service data quality platform across the entire social insurance sector.”
Thomas Schauer
Head of Data Analytics
The Challenge: 9000 Rules. No Control.
Before implementing digna, ITSV — the IT backbone of Austria’s social insurance — was managing data quality through 9000 hand-crafted rules in its data warehouse. But this traditional approach wasn’t keeping up with the pace or complexity of modern data:
The growing volume (50GB/day from 30+ sources in 500+ structures) made rule maintenance unsustainable
Knowledge was lost due to staff turnover and insufficient documentation
Rules had to be constantly adjusted, but no team had ownership
140+ daily alerts were flooding inboxes — most ignored due to unclear meaning
Only 25% of relevant data quality cases were actually covered
In short: data quality was out of control, and trust in the system was at risk.
The Solution: AI-Powered Anomaly & Timeliness Monitoring
As early as September 2021, we started the Proof of Value with ITSV and began replacing its entire rule-based framework with digna Data Anomalies and digna Data Timeliness — making a full shift to AI-driven observability and trust.
All analysis happened securely inside ITSV's infrastructure — ensuring full data privacy compliance
No tuning, no threshold configuration, no maintenance overhead
digna Data Timeliness used machine learning to learn arrival patterns and alert when data was late or missing
digna Data Anomalies automatically learned normal patterns for 50GB/day across 30 sources, then flagged deviations — eliminating manual thresholds
Use Case: Nationwide Social Insurance Data
ITSV's warehouse integrates data from five health insurance providers, serving millions of Austrians. This includes highly heterogeneous, sensitive datasets used in billing, medical claims, patient insights, and more.
The challenge wasn’t just data quality — it was data volume, velocity, and accountability. digna made it possible to:
✦ Monitor all incoming tables without writing or rewriting a single rule
✦ Ensure data freshness across regions and pipelines — even without defined schedules
✦ Provide self-service anomaly detection across teams — boosting adoption and coverage
The Results: From Chaos to Confidence
Metric | Pre-digna (2021) | With digna (2025) | Improvement |
|---|---|---|---|
Data Quality Rules | 9,000 | 0 | 100% reduction |
Case Coverage | 25% | 90% | 3.6x increase |
Daily Alerts | 140 (mostly ignored) | 20 (actionable) | 86% fewer false positives |
Rule Maintenance | Constant adjustments | Fully automated | 0 hours/month |
Learn More from ITSV
ADV Data Excellence Conference Report
Big Data Minds Europe: ITSV x digna (YouTube)
One Year Without Technical Rules – digna.ai
