The testing community has to find a distinctive voice for the As-a-Service journey


One thing about testing services that continues to strike me is that the development is largely out of sync with the broader IT market. That is not to suggest that the testing community lacks sophistication or innovation, but we cannot just use the usual mindset, concepts and monikers without adaptions when we discuss testing services. Much of that has to do with the reluctance of buyers to invest in testing. For many organizations, testing services remain a secondary concern when setting strategic IT goals or embarking on transformation projects. Yet, as organizations journey toward to the As-a-Service Economy is accelerating, and in particular Intelligent Automation is fundamentally changing the way we deliver services, the discussions on testing have to move center stage. HfS had the opportunity to sit down with executives of Capgemini and TCS to discuss their strategies for test automation and how the notion of Intelligent Automation will shape the future of testing services.

Desperately seeking an organizational model for testing

Testing services have never fully mirrored the broader IT market in the way it was seeking to optimize its organizational models. Be it aiming to centralize large parts around the notion of shared services or be it by embracing large-scale outsourcing. The build out of Test Centers of Excellence (TCoE) has always been a litmus test for the progress with centralization efforts in testing. However, as executives at Capgemini put it:”TCoEs have flat lined”. The reasons for that a likely to be twofold. First, the lack of maturity on the buy-side. Second, the traction of Agile and DevOps methodologies. The latter has two direct consequences: On the one hand the requirement for more co-location, yet as Capgemini put it with more intelligent solutions than just aligning delivery teams. On the other hand, both executive teams agreed on the rise of Distributed Agile. While Agile is intrinsically aligned with the journey toward the As-a-Service Economy, the testing community has to articulate and demonstrate what the concept exactly means. Not least in the context of vastly varying buyer maturity, or in the exasperated words of a Capgemini executive:”99% of the market is still Waterfall.” As a result, both Capgemini and TCS see Distributed Agile as the next key development phase for testing services.

Deconstructing Test Automation

Distributed Agile is the logical evolution of testing services to support the journey of organizations toward the As-a-Service Economy. Yet, as we have suggested, we need clarity around the different methodologies and monikers compared to the broader market. Historically the notion of Test Automation was largely defined as test case automation, and to a lesser degree as the automatic provisioning of test environments. For Capgemini, the direction of travel is toward the notion of a Virtual Test Factory (VTF) that can be embedded in heterogeneous test factories through virtual delivery management and governance. Over time it will also be key for the alignment with Automation Drive suite of services that is aiming to leverage the disparate, broader automation skills as well as four CoEs across the traditional business units. Thus, progress with VTF is crucial for pushing competitiveness and distributed agile as well as moving toward the As-a-Service Economy. The two key building blocks for that journey are Smart QA, an end-to-end ecosystem that includes smart assets, zero touch testing, smart environment provisioning, 360 degree view insights and smart analytics to drive down cost of delivery and cycle times significantly while improving customer experience. Furthermore, the Intelligent Test Automation Platform (ITAP), comprising of intelligent frameworks and robotic agents that underpin analytics and rules driven smart test strategy, quality gates, job chains with no manual intervention leading to continuous testing and delivery. Both platforms are centered around end-of-end life cycle automation along with smart dashboards to offer a service catalogue as well as a heat map of critical issues. Consequently, these platforms are evolving toward notions of self-remediation. However, in this context self-remediation means more providing knowledge-based solutions for business agents than self-remediating engines.

TCS is echoing many of the sentiments on Distributed Agile. To adapt its existing client relationships to this methodology the company is creating “virtual rooms” on the account level. Thus, assuring that the “distributed” work streams are being aggregated to support business processes. Therefore, the vision for its 360 Degree Assurance Platform is to evolve into an Adaptive Assurance Ecosystem. To progress toward the notion of “adaptive”, TCS is aiming to leverage Machine Learning and AI to evolve into self-healing capabilities. This is predominantly done by leveraging a set of neural networks (i.e. beyond its flagship platform ignio). Use cases are test suite optimization, automated defect analysis and the prediction of outcomes through linear regression algorithms. TCS executives were pointing to the fact that customers are starting to look for value chain execution rather than just test case execution. Another reference point that testing services are starting to move up in the value chain.

What are the testing strategies for Intelligent Automation?

While the two discussed approaches indicate that the testing community is closing the gap to the broader IT market both in terms of development as well as maturity, one obvious question has still not been answered: What are the testing strategies for Intelligent Automation? When we put this question to the supply side, the standard answer tends to be pointing to the broad portfolio of existing testing services. Yet, are these service sufficient to test Deep Learning and broad scale Cognitive Computing? Traditional approaches can look at outcomes or deal with user acceptance testing, but how should we deal with the ever more sophisticated algorithms that underpin those concepts and how can we assure that those algorithms are crunching the right data sets? If the testing community wants to be included in the decision-making for the large transformational projects it has to find answers to those questions.

Bottom-line: The testing community has to find its voice – one that is being understood by the business

The more the market is moving toward the As-a-Service Economy and outcome based models, the more the testing community has to find solutions as to how to support those strategies. While we see a clear maturation in testing services, the community has to change its mindset and embrace business-led discussions. Thus, the supply side has to move beyond conversing in jargon around function and features to align itself with the broader IT stakeholders. In Q4 we will launch a Blueprint on Application Testing for the As-a-Service Economy and look forward to exploring these themes with stakeholders.

Posted in : Cognitive Computing, Digital Transformation, Robotic Process Automation


Leave a Reply

Your email address will not be published. Required fields are marked *

    Continue Reading