Machine learning (ML), robotic process automation (RPA), and artificial intelligence (AI) technologies are running the show these days and will continue to mold the future business operations. According to Forrester, “RPA and AI will join forces to create digital workers for more than 40% of enterprises.” But here comes the big question: are we ready for the changes?
In the World Quality Report 2018-19, analysts state that businesses understand the importance of AI and wish to leverage the technology, but 51% experience difficulties integrating it with their working apps, while 55% are still trying to define where to use it. The research also finds that many companies are experimenting with the applications of artificial intelligence and how it can be leveraged to improve their test automation strategies. One of the main applications of AI being explored is how it can be used to optimize test cycles by making them shorter and more effective.
Breaking through the buzz in automated testing
Unfortunately, after our own internal market research and trial of the existing artificial intelligence assistants for quality assurance, I should note that this is one of those times when hype outweighs reality. And although there are many testing solutions claiming to incorporate machine learning or AI, when it comes to actual use, they are often limited to analyzing static code or providing record-and-playback features. We can hardly consider these limited functionalities a true machine learning automation platform. Additionally, most of these “AI-ish” startups are valued at less than USD 5 million and lack enterprise-scale solutions, which can lead to additional risks and problems for large organizations.
We’re certainly not proclaiming that machine learning test automation software is a myth, as there are a few tools moving in that direction. But, don’t get overpromised – they need more time. When working with large enterprises, we constantly analyze the market and explore new tools and technologies to help our partners deliver better products within shorter timeframes. Understanding that test automation using machine learning is the future, we’ve asked ourselves: what can we do today?
In addition to knowledge sharing within the company and QA courses at our Top Gun Lab, we’ve decided to use our test automation best practices and use cases to guarantee the maximum efficiency of QA teams within the projects. Our department has built Java and .NET test automation frameworks that consist of tools for test management, reporting, continuous test cycles, and much more. This platform is aimed at reducing the time and cost required to integrate and maintain automated testing on a typical web project.
Our initiatives related to custom test automation frameworks have already produced good results, including:
- Reducing the average time needed to build infrastructure for a test automation project from 1-2 months to just 1-2 weeks;
- Making it possible for mid-level QA engineers to implement automated tests within the existing framework instead of having to rely on senior-level engineers to design and build it from scratch;
- Ensuring rapid growth of QA expertise on the project by learning from our best-practice test automation frameworks.
What does the future of AI and automated testing look like?
Even though artificial intelligence and automated testing aren’t new, their combination is still in the preliminary stages. In upcoming years, more products will break into the market and there will, of course, be trials, errors, and failures due to inflated expectations. At our QA Center of Excellence, we’re striving to take a long-term view of the technology and what kind of impact it will have on our skills and workflows.
The World Quality Report forecasts the emergence of new QA and testing roles, including:
- AI QA strategists: These professionals will need to have a solid grasp of technical and business aspects and understand how AI is applicable to business;
- Data scientists: Future data scientists will need to be experienced in data analysis methods and be able to use machine-learning test data, mathematics, statistics, and predictive analytics to create models;
- AI test specialists: These specialists will need an extensive background in testing and be able to understand natural language handling methods, machine learning algorithms, among other advanced skills to take part in testing AI apps.
Today, it’s difficult to find experts with a least some of the above-mentioned skills and it will become even harder as companies start introducing machine learning automation testing techniques. Forrester states that two-thirds of executives wrestle with recruiting AI talent, while 83% are fully committed to their retention. Currently, there is a huge opportunity for automated testing engineers to develop their skillset and grow into system trainers.
The Bottom Line
Industry buzz around artificial intelligence and machine learning in testing automation continues to set up inflated expectations for these technologies. Often times, organizations hop on the bandwagon without a clear understanding of the capabilities and limitations of the technology or tool. We recommend that you start by analyzing your weaknesses and needs and don’t shift the entire organization or project all at once. If you decide to experiment, start small, consolidate your gains, analyze mistakes, and keep learning.
If you have questions or doubts during your exploration of AI and ML for testing automation, feel free to contact TEAM’s QA and Testing Center of Excellence. We will be happy to answer any questions you have about implementing automated testing as well as provide guidance on best practices for QA and testing for your specific software project.
Delivery Manager – Ukraine,
Head of QA Center of Excellence at TEAM International