How AI can Assist Script-Based Automation
summary
Mobile Appium scripts break when locators change; AI ingests baseline Appium runs (inspect-generated xpaths), scans the XML, pins UI elements in XML and isolates object-to-object metadata, then suggests or auto-applies the next-best technical identifier. Claim: maintenance drops from an average of two weeks per release to about two days. Implementation notes: works via a CLI/API with Appium, trains continuously (TensorFlow-style ML), supports self-healing scripts and reduces fragile xpath churn while keeping programmatic control.
AI-written summary · about
chapters (20)
- 0:00Introduction
- 0:44What is Mobile Application Automation
- 1:40ScriptBased Automation
- 2:52Reaching for the Dream
- 4:27Mobile App Automation
- 4:43WebBased Automation
- 6:22WebBased vs Mobile
- 8:18Pros
- 9:22Cons
- 12:16Scriptless Automation
- 14:22Why ScriptBased Automation Fails
- 17:54How AI Assists ScriptBased Automation
- 19:08Inspect Element
- 22:30Map Element
- 24:54Twopronged approach
- 25:54Antiflake automation
- 27:54Theoretical demo
- 28:13Demo
- 33:13Recap
- 35:21QA
related talks
57mAppium: Automation for Apps
39mQ & A With The Appium Committee Panel
15mAutomate Any App with Appium #AAAA
52mBuilding an Advanced Automation Solution Based on Appium
43mMobile Automation State in 2019
41mHow AI is transforming software testing
40mBeyond DOM: Leveraging Open-Source LLMs in an AI-Powered Appium Plugin
43mSmart and Stable Automated Testing Using AI Based Dynamic Locators