Designing a Smart QA Strategy
New technology is bringing unforeseen solutions to the market by the day, many of which are capable of complex decision-making. Quality as a function is also leveraging technology to offer smart, effective, and reliable solutions that further take us down the path of automation, freeing up some of our bandwidth for bigger and better activities that need intrinsic human reasoning.
However, we need to remember that with increasingly complex and sophisticated technologies being used in testing, quality has to become a smart activity. Test automation is only as smart as we design it to be.
Although one of the success criteria for automation is minimal maintenance, ongoing maintenance is inevitable, so planning for it is essential. This also holds true for new technology solutions we will use in testing. For example, artificial intelligence-based models that we build to recognize objects, tags, and varied other elements will only be as smart as we design them to be. And even a well-designed and trained model is usable as is for only a short period of time—there is no one-shot, comprehensive solution that will hold up indefinitely. Ongoing maintenance becomes imperative, and the best effort is to build a smart model that needs minimal maintenance.
To understand this better, we should examine what attributes are truly required to make these models—and, for that matter, any automation—“smart” by today’s requirements. Matching them as much as possible to the human attributes of an effective tester is the key.
Of these attributes, curiosity is the core. If the models can be built to be curious, questioning, and interactive in nature, we will enable educated decision-making along the way.
Understandably, though, it’s difficult to simulate all elements of a human tester. This is where the distinctive nature of a human tester comes in—traits such as empathy are difficult to accommodate, but elements such as curiosity can be built in through a logical questioning mode. A model that is able to emulate a human tester as much as possible and be curious enough to ask questions and evolve is what we need for smart and scalable quality.
This will be true for all new technology solutions, whether it’s products built on this technology or that will use this technology when carrying out testing functions. Though models and smart automation solutions are becoming fairly prevalent in areas of testing that are objective, such as functional testing, performance, and security, even areas that need human reasoning and subjective choices, such as accessibility, are seeing an intervention from smart models.
A smart QA strategy is both about building smart models and making smart choices. We must keep in mind the balance between manual and automated solutions and which are needed to accomplish our quality goals.