Cognitive QA – How to Predict the Unpredictable
As digital technologies, IoT connectivity, business intelligence and SMACT continue to develop apace, John Naisbitt’s cautionary observation that “We are drowning in information, but starved for knowledge” has never seemed more apt. As applications and products grow increasingly smarter and users and customers demand more functionality as quickly as possible, businesses seeking a competitive advantage need to drive their test automation journey towards Cognitive QA. This means deploying AI and robotics to implement intelligent test automation and smart analytics that enable insights-driven decision making, rapid validation and self-adaptive test suites.
Challenging Times in Testing & QA
This year’s World Quality Report revealed that 99% of respondents say they are having difficulty overcoming the challenges of testing in Agile development and overall the 3 biggest challenges are optimising test environments, test data and virtualisation. Only 16% are automating their most common test activities such as functional test execution, functional test case design, test data generation and testing end-to-end business scenarios. As a rule of thumb, around 30% of all test cases in every test set don’t reveal different information or defects from the other 70%, and are therefore redundant. One of the main issues is that a lot of test organisations utilise a disjointed approach to test automation, while the data they rely on is similarly siloed and not properly analysed. We need to be working towards a more centralised and intelligent model.
Predicting the Unpredictable
Cognitive QA has a variety of capabilities including automatically identifying the most effective test set in your regression suite, determining which test cases should be automated, supporting the creation of test cases, and testing physical products as well as software. For example, In the medical sector, one of our clients was running over 14,000 regression tests on pacemakers and insulin pumps. Applying our cutting-edge Cognitive QA model, we determined that only 4,000 of these tests were of genuine value, resulting in significant time savings and cost reductions.
At Sogeti HQ in France, engineers have used intelligent QA automation, with cognitive elements based around natural language processing, to create a self-learning robotic arm that physically tests aircraft cockpit equipment.
Cognitive QA connects quality to desired business outcomes; minimising risk and rapidly delivering high quality software and products to market with optimal cost savings for an enhanced customer experience and a stronger brand reputation.
So the business benefits of Cognitive QA are clear, but how do we get there?
4 Elements of Cognitive QA
The 4 main elements of a successful Cognitive QA strategy are:
- A Customisable Predictive QA Dashboard that pulls together real-time, relevant information on the quality of your application landscape across development, operations and testing, making it accessible to all stakeholders and moving away from a development view and towards a delivery view.
- Smart QA Analytics which remove unconscious human bias and emotion for smarter, faster, more comprehensive data analysis. This achieves a more focused test effort, higher productivity, optimised environment provisioning, testing aligned with genuine user behaviour and a reduction in the effort required for test preparation.
- Intelligent QA Automation that facilitates automatically generated test scripts, predictive environment configuration and robotic process automation so you can focus more accurately on what to test and when, and improve the return on your automation investment.
- Cognitive QA Platforms which address the growing challenges of test environments, data, and virtualisation by automatically provisioning self-aware and self-adaptive environments to support QA and testing for the complete application lifecycle.
Navigating the Cognitive QA Journey
At Sogeti we recommend you begin your Cognitive QA journey with an enterprise wide data, test and automation maturity and quality assessment that looks at the interplay of functions, analytics and platforms. This enables the creation of a Cognitive QA Roadmap, detailing the best route to achieving the 4 elements outlined above.
We then collaborate with you to build a customised dashboard that gives you visibility that aligns with your desired business outcomes. To implement smart analytics, we run a solution build workshop with your stakeholders to define the scope of the Proof of Value. This is an 8-week process in which we confirm analytics readiness, collect and prepare your data, conduct Cognitive QA rule modelling and define the deployment approach. The next stage is utilising the insights from the smart analytics to implement intelligent QA automation for accelerated testing.
Finally, there is the introduction of Cognitive QA platforms with self-learning capabilities that make huge inroads into future proofing your QA and Test capabilities.
For a deeper dive into the benefits of Cognitive QA and how you can start your journey to more intelligent testing, take a look at our Cognitive QA microsite here.
- Paul SaundersGlobal Marketing & Communications Director
Paul SaundersGlobal Marketing & Communications Director