Task 1: Fixing the Tesla Model 3 Listing Problems

Overview

Our claims department frequently gets claims from dealers, who think they purchased a Tesla Model 3 Long Range (Big battery), but get delivered a Tesla Model 3 Short Range (Small battery). Our car listing page shows “Long Range”, so the claim is valid and we lose money. Our intern does a preliminary investigation, and identifies the following breaking points:

1. Information from the seller
2. Information ingestion by us
3. Internal quality check
4. Internal data transfer to listing page
5. Unclear wording of listing page

What would you do to fix this leakage?

Proposed Solutions

1. Information from the Seller

Sellers may provide incorrect or incomplete details about their vehicles, either unintentionally or to make their listing more appealing.

Options:

• Mandatory VIN Entry
  • Require sellers to input the Vehicle Identification Number (VIN) when creating a listing.
  • The VIN contains detailed information about the vehicle, including model, trim, and battery size; hence, the buyer will be able to validate the model before its arrival.
  • That also opens up the way to automate such validation on the purchase step.
• Standardise Seller Input Forms
  • Use dropdown menus and predefined options for make, model, trim, and battery size to reduce manual entry errors.
  • Improve frontend and Include validation rules to prevent submission of incomplete or inconsistent data.
• Seller Education
  • Provide clear instructions and tooltips explaining the importance of entering accurate information.
  • Include examples and a FAQ section to address common questions.

2. Information Ingestion by Us

Errors can occur during the data ingestion process, leading to incorrect information being stored in our system.

Options:

• Data Validation Rules
  • Implement validation scripts that check for inconsistencies or missing critical data during ingestion.
  • For example, if a Tesla Model 3 is listed, ensure the battery size corresponds appropriately.
  • Use tools like GreatExpectatiosn to validate the common sense of the data.
• Error Logging and Monitoring
  • Set up real-time (or at least postprocessing) monitoring to alert the team of any data anomalies.
• GenAI Validations
  • Schedule periodic reviews of ingested data against source information to ensure accuracy.
  • Use GenAI tools for deep and heuristic analysis.