Automation is easy. It's the Testabality That's Hard.
There's no shortage of front-end and backend tools to create automated tests. There are even AI-aided tools and self-healing tests. If so, then why do so many automation projects fail or lose credibility? The big challenge with automation is not making button click through a script. That's the easy part. The challenge with automation is creating an environment that is deterministic, decoupled from dependencies that make test results unpredictable. One of the toughest dependency issues revolved around data, creating it, consuming it and discarding it. At Transamerica, we developed a tool Test Data as A Service, or TDAAS to provision 401k data for testing, with an API. This concepts allows for data tp be always the same as it is backed up into a read-only repo, and then cloned through the API. Shlomo will discuss the problem and the solutions Transamerica considered before arriving on a solution, and how it greatly stabilized their automation.