How AI can Assist Script-Based Automation
Mobile application automation is a formidable approach to decreasing testing time and cost while increasing confidence in testing coverage. However, mobile application automation is complicated given the lack of open source tools, device fragmentation, complicated native application development, and loss of time-to-develop compared to web-based automation. However, we are fortunate to bridge the years of mobile application automation from its current infancy to further maturity via AI & machine learning assistance; thus, creating codeless (scriptless automation). AI assistance can put forth even further value to mobile application automation by also offering a solution to assist script-based automation in finding and generating correct technical identifiers. This helps alleviate the major pain point of flaky & unreliable test scripts, so that testers may focus more time on the important things, like the never-ending checklist of testing, and less time on maintaining & tweaking unreliable scripts, all while keeping the traditional, programmatic approach to automation.
Outline/Structure of the Demonstration
1. presentation 0:00 - 25:00mins: In this presentation, I will go over mobile application automation: what it is, in what perspectives can mobile automation be achieved (programmatic perspective/ script-based automation compared to codeless perspective/ scriptless automation), the advantages and disadvantages of each, how AI has assisted scriptless automation to alleviate disadvantages, and how AI can also assist script-based automation to also alleviate the large disadvantage that is unable to find elements on page.
2. Demonstration 25:00 - 35:00: Following the presentation, I will demonstrate how AI can assist in finding elements to help maintain script-based automation by first showing an example of a failed execution due to 'elements on page not found' using Appium. After, I will implement AI on the same run and show how AI's 2-pronged approach - ability to scan XML hierarchy object by object for best technical identifier for element & ability to map elements across the page from a baseline manual session - can report correct technical identifier to be used in the Appium script. I will then rerun the test and show the passing of the execution.
3. Q&A 35:00-45:00: After the demonstration, I will take a moment to answer any questions regarding the presentation and/or demo and open up for discussion with participants of Appium Conf.
The outcome of this presentation is to share information with other quality engineers and testers on how AI can assist his/her automation efforts.
Quality Engineers (jrs & srs), Software Developers in Test
Prerequisites for Attendees
Background knowledge of script-based automation for mobile applications