Banner Image

Business and Technology Insights

Adopting Robotics in Quality Engineering

 
November 16, 2017

When we talk about the use of robotics in retail, we usually imagine warehouse or cargo operations in the background, with robotic vehicles, smart shelves and the like. With advances in machine learning, sensors and computational powers though, robotics is set to play a more visible role in retail and become an integral part of the customer experience. Best Buy’s robot ‘Chloe’, for instance, uses its extensible arm to retrieve DVDs, video games, and other items from store shelves. Walmart plans to use a patented, intelligent shopping cart to ensure in-store shopping convenience. The cart connects to customer mobiles and follows them across the store, while avoiding other shoppers. Shoppers can thus have both hands free while walking across store floors. Target’s robot ‘Tally’ tracks inventory on store shelves, while Lowe’s is testing a customer assistant robot to address the easier questions posed by customers.

These examples make it clear that the larger retailers are already leveraging the new generation of smart robots to foster the next wave of business improvements. With business innovating at such a fast pace, quality engineering (QE) and testing practices around retail solutions and devices too need to catch up. Let’s look at some specific examples of how robotics can enhance QE in retail.

Physical and Software Robotic Automation

While automation of mundane test processes is a default expectation, the true potential of automation lies in embracing hardware to enhance automation.

Retail Points of Sale (POS) comprise of peripherals that need to be tested. Often, simulators are used to automate testing, but devices such as Pin pad and Card readers cannot be virtualized as they are encrypted. Using custom built robots to perform pin pad tests can reduce the testing cycle from three weeks to few days.

The retail supply chain depends extensively on components such as Hand Held Terminals (HHTs), voice picking devices or scanners used with logistics solutions. The current windows or web automation solutions cannot access such hardware, making it difficult to devise automation solutions for these hardware.  A Cartesian physical robot (robot with degree of freedom on three dimension) integrated with test automation software such as HPE Unified Function Testing (UFT) or Selenium, can significantly improve the automation quotient in such scenarios.

Next comes the more laborious tests. Hardware devices as scanners and printers are typically tested by scanning one or maximum ten Stock Keeping Units (SKUs), due to the manual labor involved. But in production, millions of items are scanned. Such limited testing introduces risks such as missing loop tests, where errors can go undetected when business scenarios are run in loops. By going through the loop once, the uninitialized variables and loop repetition issues can be identified. Physical robots can be used in such data intensive laborious tests that cannot be performed with human hands.

Testing for Customer Connect

The increasing use of robotic solutions for customer support and service solutions has opened up a whole new set of challenges for QE and testing engineers.  Chat bots on e-commerce sites and personal assistance apps like Apple’s Siri are fast becoming the norm for the aggressively competitive e-commerce space. Accessibility and usability testing is now an imperative, but not really up to the task. While testing tools to verify visual, physical, hearing and cognitive impairments are available, decline in interpretation of the results from these tools demands cognitive software solutions. For example, if there is a description tagged to an image, the current automation can confirm the availability of a text, but it cannot determine whether the description is sufficient. Also, usability tests such as searching for known patterns, and identifying page elements and dimensions deliver precise results when a software robot performs the tests across devices and platforms. That is, the accuracy with which an image is read or a text box is positioned in iOS, Android platforms across browsers is more precise with software robots than manual or current automated tests. Current automated tests cannot compare images or compare texts and provide results.

With product improvements being made across websites, functional tests involve image and text comparisons in varying formats. This too demands a cognitive solution to compare texts and 3D images across different document formats like PDF or MS Word. Also, while such solutions are a must in pre-production test environments, they can also be powerful enablers in production sanity tests. Pushing promotions and product uploads with images during peak seasons can be doubly assured with a software robot running sanity tests.

Beyond Default Automation

Besides revolutionizing the test process, the past decade has introduced powerful tools and practices, pushing QE & testing capabilities towards greater coverage with shorter time to market. The introduction of AI and machine learning based physical and software robots can further broaden the role of automation in testing, playing an enabling role in scenarios where traditional testing techniques can’t work.

It’s time to harness the power of smart robots in pacing with the digital revolution.

How do you see intelligent robots changing QE in your business?

Shilpa Chandrasekhar is a solutions architect for the Quality Engineering (QE) practice in TCS' Retail Unit, and provides strategic test solutions for retailers across US, UK and Europe geographies. She has held varied roles, from being a mainframe developer to delivering projects in the QE space covering e-commerce, logistics, point of sale, warehouse management, PCI compliance, digital projects, etc. In her current role, apart from building test solutions for specific customer requirements, she works on new age QE solutions spanning across mobility and robotics.