Computer Systems Design and Related Services2025Generative AIMachine Learning (classification)NLPB2B
Amazon Web Services (AWS) / Schaeffler

AWS Virtual Engineering Workbench cuts automotive test case creation time by up to 80% using generative AI

AWS built a generative AI extension of its Virtual Engineering Workbench that automates test case creation from automotive software requirements — reducing test case creation time by up to 80% while maintaining a human-in-the-loop review step. In a real-world deployment with Schaeffler, the system cut an experienced test engineer's preparation time from 820 hours to 265 hours for 837 requirements.

Test case creation time reductionUp to 80%
Hours saved (Schaeffler)555 hrs per 837 requirements
Deployment time4 weeks to production
5 min read

Background

Automotive software complexity has grown to a point where manual test case creation is a significant bottleneck. For a single complex electronic control unit with tens of thousands of requirements, and a ratio of three to five test cases per requirement, the manual effort reaches into the hundreds of thousands of hours for a full vehicle program. AWS Professional Services built the VEW AI extension specifically to address this bottleneck, targeting the classify-and-generate steps that consume the most engineering time.

What Was Implemented

  • Generative AI extension integrated into the Virtual Engineering Workbench (VEW) cloud framework
  • Four-step automated workflow: requirements upload → AI classification of requirement type → AI generation of test conditions → AI derivation of test cases
  • Built on Amazon Bedrock with Anthropic Claude (Instant for classification, 2.0 for test case generation)
  • Human-in-the-loop review and approval at each step before progression
  • LiteLLM-based AI gateway for multi-model access, cost tracking, and rate limiting
  • Session state preservation so engineers can pause and resume generation jobs
  • Collaborative sessions allowing multiple team members to work on a single test generation run
  • Deployed to Schaeffler in production within four weeks; architecture uses API Gateway, Lambda, S3, and DynamoDB

Results

The general VEW capability reduces test case creation time by up to 80% , per the January 2025 AWS blog. In the Schaeffler deployment specifically, the system accelerated test case generation by up to 60% : an experienced test engineer now prepares 837 system requirements' worth of test cases in 265 hours rather than 820 hours (reduced from ~1.02 hours per test case to ~0.32 hours). The solution was implemented in production in four weeks. The human-in-the-loop design also reduces the risk of system requirements remaining unvalidated due to human error.

Lessons

  • A human-in-the-loop design is essential in safety-critical automotive testing — it preserves AI efficiency while maintaining engineer accountability
  • Prompt engineering (role-playing as a test engineer, structured output formatting, precision instructions) is a critical factor in output quality
  • Session state preservation and collaborative workflow support are required for production adoption at scale
  • Even with strong AI assistance, an initial context investment (prompt templates, classification criteria) is required before benefits materialize

Ready to implement AI in your commerce operations?

McFadyen Digital helps teams move from case study to live implementation.

Talk to an expert →