ASDA scales mobile test coverage by 50% and cuts manual testing effort by 70% with Perfecto AI

Background

ASDA's engineering teams depended on physical, on-site mobile devices for testing, creating resource constraints and delays. Manual testing bottlenecks were slowing delivery of digital features for its ecommerce site, mobile app, and in-store scanning applications. The company needed to scale testing across more devices, reduce manual overhead, and improve release confidence without proportionally growing headcount.

What Was Implemented

• Perfecto AI cloud testing platform deployed in a hybrid cloud configuration (ASDA-hosted devices + Perfecto cloud)
• AI-driven automation covering ASDA's ecommerce website, mobile app, Scan & Go feature, and Rewards app
• Image validation automation — confirming promotions appear correctly and with accurate content
• Natural language testing enabling non-engineers to create and maintain test scenarios
• Integration with Zebra store/depot devices via DataWedge for barcode-based test automation
• Removal of on-premise physical device dependency; team members access test devices remotely

Results

After implementing Perfecto AI, ASDA reported: 50% increase in test coverage , 70% reduction in manual testing effort , 4x increase in device and OS coverage , and test execution time reduced from days to hours . Cost management was also improved through a transition from onshore to offshore QA resources while maintaining quality, enabled by Perfecto's automation capabilities. > "Perfecto enabled us to cut costs while enhancing test coverage across mobile apps. With the ability to run more tests across more devices, we streamlined operations, eliminated physical laptops, and removed reliance on our environments team—delivering a smarter, more reliable mobile testing solution." > — Matt Hatherly, Engineering Lead – Automation & NFT, ASDA

Lessons

• AI-powered cloud testing platforms can eliminate the physical hardware dependency that limits test scale in retail and enterprise environments
• Automating image validation and promotions-content testing — areas historically resistant to script-based automation — is a high-value early target for AI testing tools
• Hybrid cloud deployment (vendor cloud + self-hosted devices) addresses security and control concerns that block full cloud adoption in regulated enterprise environments
• Integrating AI testing with proprietary in-store hardware (e.g., Zebra scanners) requires vendor-specific support; validating this capability before platform selection is critical
• Natural-language testing interfaces that allow business users to create and maintain tests reduce dependency on engineering resources for test authorship