What do you do? is the company that monitors and assures the health of AI models in production. Already used by top-tier organizations, monitors millions of predictions daily to eliminate the risks derived from these models’ black-box nature: bad decisions, unwanted bias, and compliance issues. Their AI assurance solution acts as the one source of truth for all stakeholders, and empowers data science and operational teams with the right insights to scale their use of AI by becoming more independent, agile, and gain confidence in their models’ operations.

Implemented use cases include Customer Lifetime Value (CLV) predictions, fraud detection, lead scoring, underwriting, credit risk, and more. Recognized for its innovative technology and approach, Gartner recently named Superwise as a 2020 Cool Vendor in Enterprise AI Governance.

How much does it cost?

Pricing is based on the number of predictions and the number of models monitored. A premium is applied for enterprise-wide deployments.
There is also a free trial option

What’s a sample use case? Where can I learn from?

Feature List

Our Documentation is available here

  1. Real-time automated anomaly detection for any feature or performance metrics at the segment level – without the need for manual thresholds definition
  2. Automated detection of problematic segments in the data’s sub-population 
  3. Ability to prioritize incidents amongst dozens of models
  4. Feature level and multi variant drift detection
  5. Model performance metrics (model agnostic) – Confusion matrix, Calibration charts, ROC curves
  6. Feature level descriptive statistics – Outliers, New values, Missing values
  7. Data Profiling
  8. Alert consumption via webhook
  9. Full UI to investigate the root cause of alerts with ease
  10. API-based management and consumption