Subject Matter Expert Quality Control
​
Our automotive SMEs work hand-in-hand with advanced automated testing frameworks to ensure consistent accuracy and production-ready output.
Subject Matter Expert (SME) Quality Control is a critical component in ensuring the accuracy of Predii’s automated processes. While our system handles large volumes of data and feature extraction, outputs are continuously validated by automotive experts to guarantee production-level accuracy.
Automated Testing Framerwork
Predii employs automated testing frameworks to perform initial validations on AI-generated outputs. These frameworks simulate real-world scenarios to assess the accuracy and relevance of the data.
SME Quality Control & Continuous Improvement
Following automated testing, automotive Subject Matter Experts (SMEs) review the outputs to ensure they align with industry standards and real-world practices. SMEs provide feedback and corrections, which are then integrated into the system.
Predii Review Management System
​
The Predii Review Management System is a UI based review tool that allows our Subject Matter Experts to review the quality and accuracy of our Feature Extraction.
​
This review process helps ensure that our AI models are producing reliable, accurate results, allowing our customers to trust the insights provided. It enables continual refinement of the system, addressing gaps and improving coverage over time, ultimately leading to more accurate and actionable data for decision-making.
​​
Predii's SME review process is divided into three steps: Generic Review, Feature-specific Review, and Undiscovered Lines Review, each designed to ensure processing accuracy and identify any gaps. Reviews are prioritized based on the most impactful input-output combinations.
Generic Review: A general check of processed data, reviewing approximately 1,000 lines to identify known gaps and required changes.
Feature-specific Review: Focuses on key components or repairs with high impact on the output, reviewing up to 85% of unique input-output combinations.
Undiscovered Lines Review: Identifies missed features and blind spots, continuously expanding coverage by adding new symptoms, components, and repairs.
​