Skip to Content

The Causal AI conference is back in San Francisco for 2024, bigger and better than ever.

Register Interest

$50bn Global Electronics Company chooses causaLens to revolutionize its early warning system for faulty parts with Causal AI


$50bn Global Electronics Company



Use Case

Early warning of failure and defects in products


Reduction of product recalls and risk to customers

The Challenge

A key business goal for this leading electronics manufacturer is to ensure the safety and quality of its products. Product defects lead not only to higher substantial costs but also pose significant safety hazards to its customers.

To mitigate these risks, they recognized the need for an advanced early warning system that could accurately detect potential failures. The key requirement was not only accuracy but also explainability – they needed insights into the system’s behavior that could identify root causes and provide transparent justifications for alerts.

In their search for a solution, the customer had explored the realm of traditional machine learning models, working previously with 6 different vendors and trying an in-house solution. However, they encountered a significant roadblock when these models, which showed promise in controlled lab environments, faltered when exposed to the real world. The challenge lay in generalizing the patterns learned in the lab to unpredictable real-world scenarios.

In essence, the customer’s quest for an explainable, accurate, and adaptable early warning system stemmed from the need to avoid disastrous consequences, enhance operational efficiency, and uphold safety standards in the face of complex, real-world challenges.

  • 1 Icon
    The outcomes of failure are both costly and hazardous

    Failures would impact their reputation and the safety of individuals using their products

  • 2 Icon
    The struggles with traditional approaches

    The customer had worked with 6 different vendors and tried an in-house solution without success

  • 3 Icon
    Lack of value from Existing Approaches

    Machine learning approaches fail to understand physical processes leading to lack of generalization and explainability

The solution

Given the lack of success in-house and with other vendors, the customer turned to causaLens for a solution.

  • causaLens offered a causal approach that generalized to the real world. The models, built using causalNet, our proprietary structural causal model, do not require labels and respect the physics of the system and the unique nature of each product
  • Through causal inference, features were extracted with direct physical interpretations, bolstering both explainability and generalization
  • The solution, built leveraging causaLens’ Human-Guided Causal Discovery tool, included a collection of prioritized features harmonizing with domain expertise or automation, a prototype model for defect identification complete with relevant metrics, and a visual application
  • causaLens not only tackled the complexities inherent in real-world scenarios but also enabled the extraction of meaningful insights, identification of defects, and facilitation of transparent model understanding—a comprehensive solution that resonated with their challenges

Results and Benefits

The causaLens solution allows the customer to significantly shorten the costly R&D cycle, with fewer tests that brings better and much more trustworthy results. The model and reality were a complete match to the real physical properties of the product. 

This project is the first step in a long-term vision for an early warning system for all their product range, envision a fully automated and explainable early warning system that:

  • Identifies defects ahead of time
  • Explains the root cause of such defects
  • A system that generalizes from the lab to the real world 
  • Extend the solution to many other use cases that can solve their problem better and faster

Start your Causal AI Journey

Request a Demo