How AI will Address the Tester Shortage and Software Quality

Despite limited attempts at automation, almost every aspect of the software test life cycle remains a manual process. Add to this a global shortage in the number of available skilled testers. CIO’s can develop workaround strategies, but only by developing autonomous test capabilities powered by Artificial Intelligence (AI) will meaningful impact occur. AI will not only address the shortage of testers, it will improve software quality and the speed of software development.

AI will not be eliminating tester jobs. Instead, AI will address a tester shortage and allow IT to test more applications in less time, with higher quality.

Even if a CIO could staff at an ideal 1:1 developer-to-tester ratio, they’d never have peak demand testing capacity when the DevOps is in full production mode or when Quality Assurance (QA) has a backlog. Human testers take lunch breaks, sick days and vacation days, and they like to go home to their families every night and on weekends.

The numbers don’t lie. We estimate there are roughly 21 million developers and about 4 million testors globally. That means the developer-to-tester ratio in most companies is 1:4, or worse.

We can never have enough dedicated human testers to cover the scope of all testing a CIO might need done in agile development environments.

Why?

Because in our estimation, over 80% of the test automation cycle remains in manual mode:

So, how can the CIO deal with the tester shortage

1.Let There Be Bugs!

Lower the quality of software. Let the bugs flow straight into production and onto user screens. Let the user complaints that are logged in Support be your backstop.

We don’t advise this strategy, but it’s something you could do.

2. Serve up the Dog Food

There’s also the “Alpo” approach, where developers “dog food” a version of the software, pretending to be a user or working closely with people who do use it. Dog Fooding (that’s now a verb) is fine for general apps like spreadsheets and word processing, where user interaction is somewhat predictable. A niche, bespoke application will be much harder to Dog Food because you really have no idea how a customer will use it.

3. Train the CEO to be a Tester

That may sound like a ludicrous strategy, but it’s not all that different from deputizing developers, business analysts and other human proxies to step into the test suite.

One workaround to the tester shortage has been the Software Development Engineer in Testing (SDET), but that is an expensive option that reduces your pool of available developers.

4. Know the Flow

Look closely at testing workflows. Build-in more logging so you can quickly identify where the logjams are occurring and clustering.

5. Momentum from Machine Testing

A fifth strategy is to develop autonomous testing capabilities. By autonomous, we mean end-to-end testing directed by a machine without the need for human control or intervention. Advances in AI will result in roughly 80% of of the software testing cycle becoming automated, allowing skilled testers to oversee quality across the board without halting development to write Selenium scripts.

Two other advantages to establishing an autonomous testing environment is development speed and software quality. IT spend less time fixing bugs and there will be fewer bugs living in the code all the way into production. And best, no one has to ask the CEO to learn Selenium.