Diffblue attended this year’s Quest for Quality conference in Dublin, Ireland on the 5th of November to learn about the wider market for applying AI to testing tools and talk to testers about their needs. We got the chance to speak with testers at the forefront of the industry and attend interesting and relevant sessions on topics ranging from what the future of testing with AI will look like to how testing can be shifted left.
Here are some of our key takeaways from the event:
AI is augmenting testing
Tariq King, Head of Quality at Ultimate Software, gave the opening keynote on the subject of “AI-Driven Testing: A New Era of Test Automation,” where he gave an overview of the main players in the AI for testing field. Using AI, many of the tools currently in development learn how users are testing and are augmenting the process with reinforcement learning, with a lot of potential for automating UI testing. Innovative AI for testing tools mentioned in his talk included Diffblue, mabl, test.ai, testim, eggplant, sofy.ai, AppTest.Ai, Functionize, testrigor, retest, autonom, and applitools (among others).
Tariq King’s opening keynote at the Quest 4 Quality Conference 2019
Diffblue’s CEO, Mathew Lodge, presented on the subject of using AI to create regression suites of unit tests to allow developers to get instant verification of their code changes on every commit, which is currently the only application of AI to testing in the early phases of the software development lifecycle (SDLC), compared to at the UX level for end-to-end testing on the interface.
Diffblue’s CEO, Mathew Lodge, presenting at the conference
Topics also touched on the ethics of AI, including data privacy and how to get the datasets that are needed for testing—challenges that will continue to be addressed as AI tools advance.
QA testers can be empowered to produce unit tests
Ilya Sakharov, Director of QA at HelloFresh SE, made the point that quality is everyone’s responsibility and outlined suggestions for how to actualize this goal in the organization.
One way for engineering teams to share responsibility for quality is to bring unit testing to everyone. Typically, QA testers don’t write unit tests, but they still rely on them in order to produce their integration and end-to-end tests and validate the overall quality of the code. With the right tools, QA testers can create unit tests themselves, which is one way to share the responsibility for quality.
Manual testing is key to understanding business logic and user problems
The third main takeaway from this event was that manual testing isn’t going anywhere any time soon. The function of a tester may adapt to changing needs, but human intuition is still an important part of understanding the bugs caused by other humans, and people are best suited to understanding user problems and the needs of the business.
For this reason, manual testers should be engaged earlier into the customer journey, even working with sales to see what users need. With new testing tools to take some of the more repetitive work off their plates, QA can shift left and create better software for the business.