Report: AI is key to modern software testing

A new report reveals how traditional test automation tools and practices can no longer scale to meet the complexity of modern software delivery without AI

Organisations leveraging traditional software testing tools often fail to scale to the needs of today’s digital demands, finds a new report published by EMA Research and Applitools

The report, "Disrupting the Economics of Software Testing Through AI", examined the impact of AI and automated software testing on large enterprises. It found that Visual AI has the highest impact on software testing as compared to other available applications of AI technology in the market today.

Respondents cited escalating quality control costs and release velocities as the top factors hindering engineering and DevOps efforts, as well as the increasing number of target devices, operating systems, and app programming languages.

"Business's ability to accelerate the delivery of customer value through software innovation, at lower cost, has become critical for achieving competitive advantages," said Torsten Volk, Enterprise Management Associates (EMA), Managing Research Director. "AI-based test automation technologies can deliver real ROI today and come with the potential of addressing, and ultimately eliminating, today's critical automation bottlenecks that stifle modern software delivery."

 

Using AI to solve problems in software delivery 

The accelerated adoption of apps that reside on cloud services, up 225% since 2015,  is further compounding the complexity of software delivery, according to EMA Research and Applitools. The report shows that there’s been nearly a 100% uptick in the number of test automation-related questions posted to Stack Overflow, a popular Q&A website for programmers, over the past year. 

The EMA Research and Applitools paper outlines five areas where AI and machine learning could solve major blockers in software delivery: test creation, self-healing, visual inspection, coverage detection, and anomaly detection. Taken together, these technologies have the potential to streamline and automate parts of the software testing workflow while enhancing productivity, according to Volk.

Test creation automates the discovery of new and changed test requirements by analysing changes in apps and documentation. Self-healing fixes broken test workflows while visual inspection trains models to audit apps through the eyes of end-users. Meanwhile, coverage detection and anomaly detection identify the different paths that end-users can take through apps and report gaps in code coverage or anomalous behaviour. 

“AI-based test automation technologies can deliver real return on investment today and come with the potential of addressing, and ultimately eliminating, today’s critical automation bottlenecks that stifle modern software delivery,” Volk said.

 

Share

Featured Articles

Microsoft: building robust AI strategies in manufacturing

Manufacturing leaders have a duty to understand AI strategies, says a Microsoft Thought Leader, and data could be the key to unlocking AI opportunities

Uniphore: supporting customer service with AI innovation

Balaji Raghavan, Uniphore’s CTO discusses AI, the customer service industry and how his company’s software for customer conversation optimisation works

Catching up with Sophia: gender bias in AI

Gender bias in AI is discussed often. Here, Hanson Robotics’ robot, Sophia, shares how this bias is experienced by humans and robots alike

Artificial intelligence could be a stroke of genius for golf

AI Applications

Are we ready to hand humanity’s future over to AI?

AI Strategy

A watershed moment: feeding the world with AgriTech

Data & Analytics