As we work with more and more customers of various sizes and scale, we consistently hear that companies aren’t satisfied with their ability to have high velocity sprints without sacrificing quality. Though automation is nothing new, the stakes have never been this high, nor the IT landscape this complicated. As we continue to bring autonomous testing to new customers in different verticals, the same problems persist across customer environments. We want to reflect upon two recurring themes we see in customer discussions.
People Have Burnt by AI
Over the past few years, IT has been seduced by AI solutions that have promised the world but delivered minimal value. Companies have been burnt by vendors trying to deliver value with immature technology, vendors who frankly don’t understand the daily headaches of managing large IT teams. At the same time, AI-enabled technology solutions are rapidly becoming a reality due to the rise in ecosystems like Tensorflow, the cheapness and scalability of cloud computing, and the apparent bottomless pockets of large technology companies—and the venture capital community—making it harder to ignore the writing on the wall. We often find that customers are keenly aware of this reality and know they need to take action but find it difficult to separate fact from fiction. Moreover, teams are just now adopting the right vocabulary to assess the answers to their tough questions.
So how can companies cope with the divide between the need to innovate without leaning on the tips of their skis? By going into each project with eyes wide open on the value provided by these new solutions. Companies should be keenly aware of their limitations. They must understand what they’re getting (and not getting) from vendors, and looking for the skill sets needed to use the product. They must understand the scalability and maintainability of the solution, and the ease to bring it into (and take it out of) the existing tool stack.
The Rationality of a Low Test Automation Environment
We talk with three main types of companies: those without automation, those just starting to fidget with automation, and those behemoth enterprises who have bought everything under the sun. The most interesting companies to talk with are those without automation, especially since test automation is nothing new, and yet they’ve chosen not to adopt.
So why do companies consistently not adopt automation? They aren’t afraid to invest the capital into automation projects. Instead, they have come to realize automation isn’t maintainable. It would force them to spend more time reworking the automation than testing, draining resources from precious development time. Moreover, they can’t afford to put their value-added resources on constant maintenance. They prefer to keep those teams as productive as possible, even as they pile up on manual QA people. In this context, the decision to not adopt automation is perfectly rational. Furthermore, the only way to actually solve this problem is to find a lateral solution to scale in their environments.
As we continue to work with companies across various scales, industries, and technological maturities, we find autonomous testing gets companies over these hurdles, though the work is far from over. With an increasing number of customers to learn from, we hope to drill down deeper into the ailments of poor quality and apply the concepts underlying autonomous testing across the software development cycle. For now, we hope companies can use autonomous testing to eliminate their quality bottlenecks so they can focus on what matters most—developing great solutions to their customer needs.