Get Started

An Ideal AI Use Case

Regression Testing is perfect for AI

Artificial Intelligence is all the rage in 2018. Indeed, every software vendor worth their venture capital is touting how AI is infused into their products. Such peak-hype masks the fact that productive uses of AI require appropriate use cases, including the availability of relevant big data from which the AI can learn.

An ideal AI use case is one where humans currently perform routine work tasks, tasks that require more persistence than creativity. Such mundane work is far from the best use of human capabilities. In fact, it often falls prey to the very human tendency towards boredom, which in turn leads to spotty execution. That can lead to disaster, since routine work is still important work, sometimes even mission-critical work.

Then there is the cost component, as in-house labor is expensive, whether applied to routine tasks or not.

It is for these reasons that so many work processes comprised of routine tasks have been outsourced and then off-shored. While off-shoring reduces the labor cost of routine work, it also inhibits collaboration and agility, and doesn’t remotely assure high quality execution.

That is why business processes comprised of routine work tasks are better automated than outsourced. A goodly chunk of the valid enthusiasm for AI is because it’s seen as the key to automating routine work processes. That enthusiasm is not misplaced, assuming that a current and comprehensive big data source is available from which the AI can learn.

Regression Testing Loves AI

Now let’s consider software regression testing, the QA process for finding defects that have crept into previously released software as a side effect of seemingly unrelated fixes or new features.

Regression testing is often the bane of testing departments. Necessary but mundane, exacting but boring, labor intensive but unrewarding, it gets little respect and is often shortchanged in an effort to address other items in the QA backlog.

Said another way, regression testing is a quintessential work process comprised of routine tasks that require more persistence than creativity. Not surprisingly, it is an oft outsourced process. So in those regards, it is well suited for automation, and likely a good candidate for AI driven automation.

What about a big data source from which the AI can learn? Here too, regression testing is well suited for AI, since regression testing’s purpose is to assure that real users won’t experience bugs when traversing existing functionality. Fortunately, their activity is logged in production logs, whether plain Apache or W3C logs, or using log managers like Splunk or SumoLogic. These logs have previously proven a boon for sysadmins, and now can be further exploited for QA, in particular as a big data source for AI driven regression testing.

Add up the routine tasks involved in regression testing with the ready availability of an ideal big data set from which an AI driven regression testing system can learn, and it becomes clear that regression testing loves AI.

Appvance IQ’s Automatic Regression Testing

Appvance IQ seizes on all of the above to enable it’s AI driven testing, coupling that with Continuous Testing capabilities to create an Automatic Regression Testing system. The results is a breakthrough in efficiency, agility and quality improvement. And there’s nothing artificial about that.

Want to see the ideal AI use case in action? Request a personal live demo here.