Industry executives and experts share their predictions for 2021. Read them in this 13th annual VMblog.com series exclusive.
AI as the Path to Autonomous Testing
By Daniel Levy, Senior Director of Product Marketing at Applitools
For several years, the software testing
industry has been navigating the best ways to implement AI technology to solve
the complexity of delivering app quality. Where is the complexity coming from
and why is quality critical in the digital age?
According to the Applitools State of Automated Testing Report,
many large companies have anywhere from 10 to 100 different applications, and
each of these applications need to be displayed on multiple devices, browsers,
screen sizes, and resolution across multiple languages. Visual bugs are
regularly escaping into production, costing organizations millions annually.
With brands increasingly moving to digital,
the functional and visual quality of an app is now quintessential to revenue
generation. Test automation is the solution to this problem and AI is playing a
big role in making it successful. While technology such as Visual AI has
matured through years of deep learning, its use cases across the software
delivery lifecycle continue to expand. In 2021, we will see the level of AI
sophistication in test automation grow and lay the foundation for the future of
quality software delivery.
The AI
Boom Continues
According to Gartner's Innovation Insight for Autonomous
Testing (October 2020), "84% of recipients responded that AI
and ML features are more important than any other features in software testing
tools." This will drive software testing companies who have yet to invest
in AI or ML to do so, and those that already have AI capabilities will likely
double down on those efforts. However, it will be important to implement AI in software
testing in a way that is meaningful, sophisticated and drives value. Data will
be the key to achieving this.
Data,
AI and Autonomous Testing
AI is naturally driven by data, and the way to
continue to refine AI is with more data. As an example, Applitools recently
passed the one billion images analyzed mark and thanks to all of this data
collected from our AI engine, we are able to deliver one false positive out of
a million assertions. Not only does data allow for a move towards more
sophistication and accuracy in AI, but the data collected also provides a key
to moving towards autonomous testing.
Currently, AI in software testing tools is
able to make recommendations based on statistical patterns, but aren't
automatically implementing these recommendations yet. The data collected from
AI engines will be the key to achieving autonomous testing, freeing testers up
for more mission-critical tasks and innovation while driving even faster
time-to-market. End-to-end Autonomous testing may not be fully realized in
2021, but as AI in software testing matures, we may begin to see the beginnings
of this pattern take form.
##
About the Author
Daniel Levy is a Senior Director at Applitools working remotely from his home in Portland, OR. Daniel has a passion for technology, a vision for product, and a story to tell. Outside of work, you’ll likely find him hiking, biking, behind a camera at sunrise, enjoying an espresso or craft beer, or gaming.