Test Engineer, Automation Engineer, QA Engineer, Developer in Test. After more than a decade in software testing, I’ve heard all of these job titles come and go. But with AI taking on a bigger part in software development than ever before, I’d like to propose a new job: AI Automation Engineer. Though this sounds like a highly competitive role, there’s good news for hiring managers: these employees are in infinite supply and you don’t need an elaborate screening process to find the right one.

Shifting automation left

Over the years, significant effort has been put into finding issues earlier in the software development lifecycle. The rise of automation approaches has shifted efforts from manual regression testing into automated regression suites that run with every build. This has been great for those people working in software testing—there’s nothing more boring than running the same test case manually for three, six, or even twelve months (or longer).

With a good automation suite, your manual test resources are able to focus on one thing: Does your product do what the customer wants? A regression test is great at telling you what behavior has changed in your product, but a human is required to tell you if the product is doing what your customer wants today, tomorrow, etc. I am a big believer in using exploratory testing to truly understand how a product is going to behave. The best testers may not be able to tell you why they found that interesting bug, just that their spidey sense took them in the right direction.

Bring in the AI Automation Engineer

How can your AI Automation Engineer help? Simply put, the big problem with having a comprehensive automation suite is the cost of creating the tests. Very few of us are lucky enough to work on completely greenfield projects with the time we want to devote to writing automating testing. This means we are always playing catch up.

Catching up on automation efforts is difficult and time consuming. Tasking engineers with increasing coverage of test suites is normally met with groans. To provide a quality test suite requires good engineers. This, coupled with a lack of desire to do the job, means that it is going to be expensive. There must be a better way.

With the progress made in AI in the last few years, AI can now take on the supporting role of AI Automation Engineer. Our AI tool, Diffblue Cover, is able to read your existing code and write a suite of unit tests that describe the current behavior of your application.

Companies like Goldman Sachs have been using this technology to save man-years of effort in bringing up their existing test suites to acceptable standards. This has allowed their valuable test resources to focus instead on tasks relating to code quality, and developers can enjoy getting earlier feedback on the quality of their code changes.