Is a world where robots can learn to do the work of a store manager too far out of reach? Or what about a world where self-aware, autonomous drones can help data scientists validate complex data models related to traffic patterns and management, for example, in a matter of hours rather than weeks? Or a world where application repair is automatic – where users can submit a defect to a smart management system that could drive the self-healing process for repairing this application, without involving any humans. Is it possible to develop a smart system where managers are given real-time updates on software performance and provided with predictions on potential risks and mitigation strategies? Will we see these solutions come to life during our lifetimes? Perhaps. Many of these scenarios are quickly moving from the dominion of science fiction into reality thanks to the integration of Robotic Process Automation (RPA), Data Science and Cognitive methods called Cognitive Robotic Process Automation (CRPA).
Testing services, as an industry, is at the cusp of a transformational revolution that is being brought about by the adoption of CRPA. Many large IT companies are currently working on building platforms or solutions to prepare for this new trend. Why are these systems so desirable? Because there are bound to be anomalies in the application development process, whether large or small; human or machine. Not only that, but identifying and remedying an anomaly within an application is a time consuming process for all who are involved – it can take days, sometimes hours, to locate, isolate, or even fix a given defect. However, with new advancements that the industry is seeing in the CRPA field, solutions and systems – instead of people – can now predict and prevent these anomalies, and drastically shorten, or even eliminate, the time, effort and cost needed to fix them.
Automation is not a new process to those in the IT industry – various automated processes have been around for years and have plenty of time to mature. As these processes have matured, and been fine-tuned, the natural next step is to start looking into how machine learning and intelligence can be incorporated into these processes. With cognitive computing, systems will be able to contextualise and use reasoning to assess problems, instead of just solving the “usual” repetitive problems that IT faces.
Cognitive computing can also help businesses shift their focus towards prioritising “IT wellness” – placing more value on product quality (rather than just process quality). By being able to effectively process businesses’ data, coupled with the ability to learn from past experiences, these systems can learn and eventually master the tasks they are given. In short: continuous improvement. In the Quality Assurance (QA) space, there are two facets of AI technology that could successfully solve this equation:
1. Cognitive Automation: Data analytics forming the foundation of algorithms, pattern recognition and methodologies
2. Autonomic Healing: A one two punch of Natural Language Processing (NLP) and Machine Learning
While not deployed adequately in the QA process yet, both have the ability to inject dynamic intelligence into strong, existing systems and with Continuous Improvement, make them reliable over decades. The following are the key areas where CPA can make a significant difference in the QA process.
1. Proactively Monitor Application Health: Using CPA, application health can be monitored by a variety of BOTS (Software Robots) during application development. These BOTS observe data patterns , discover trends and use models to predict the impact of a given change on the application with allied risks and vulnerabilities.
2. Optimise Testing: Traditional optimisation techniques become inadequate when changes to applications are frequent and where large assets already exist. Manual intervention is ineffective and the QA process could be compromised. Cognitive computing, using continuous learning systems optimise and automate Test Assets to maintain a lean, dynamic risk based Test Suite that self-maintains throughout the application life- cycle.
3. Self-Healing: Identification of anomalies in an application and the subsequent effort required to accurately identify, isolate and fix them is time consuming and tedious. A learning based system coupled with the development of self- service BOTS (which can be invoked when a defect is detected by other BOTS or a human) can make the Defect Management process intelligent. This results in significant effort elimination that current QA systems necessarily entail.
Whether for monitoring application health, self-healing or optimising the testing process itself, there is extensive applicability for artificial intelligence – or Cognitive Robotic Process Automation – in testing. As a result, we will start to see its widespread adoption across the IT industry.