OpenText-Startseite.
Customer stories

pick n pay logoPick n Pay

End-to-end payment testing driven by OpenText™ Functional Testing

pick n pay logo

About Pick n Pay

With more than 2,000 stores across South Africa and seven other countries, Pick n Pay Group is one of the leading retailers of consumer goods in Africa. Trading under two main brands—Pick n Pay and Boxer—the company traces its history back to 1967.
See how Pick n Pay enabled end-to-end POS testing with OpenText test automation.
  • Stores:
    2238
  • Headquarters:
    Cape Town, South Africa
  • Employees:
    90,000
  • Main retail lines:
    Grocery, home, clothing

Summary

Challenges

  • Difficulty validating multiple payment methods consistently.
  • Errors or fatigue when mimicking real customer actions.
  • Adding new payment modes without process redesigning.

Solution

  • Enabled end-to-end POS testing using a robotic arm and software automation.
  • Unified validation across payment types and third-party services.
  • Scalable, high-volume POS automation.

Results

  • 50,000 automated test scripts weekly
  • Regression cycle reduced to 3.5 days
  • ROI achieved in 5 months with nearly 99.9% automation

Challenges

  • Difficulty validating multiple payment methods consistently
  • Errors or fatigue when mimicking real customer actions
  • Adding new payment modes without process redesigning

Pick n Pay, like other retailers, operates in a world where payments, fulfillment, and customer expectations evolve faster than manual testing can keep up. Every tap to pay, swipe, QR code, loyalty card transaction, and every item scanned on Zebra devices for online orders or store pickups, must work flawlessly, every time. But validating all these journeys across hundreds of stores, devices, and payment types is complex, repetitive, and prone to human error. Pick n Pay needed a way to reliably recreate real customer interactions, automate them end to end, and adapt quickly as new payment methods or fulfillment workflows were introduced.

Robotic arms assembling circuit board in an automated factory

We moved from quarterly to monthly releases—and can now release every 12 days—thanks to 24/7 automation and near 100% regression coverage. The robotic arm paid for itself in just five months!

Leon van Niekerk
Head of Testing Center of Excellence, Pick n Pay

Solution

Pick n Pay created a cost‑effective automated testing setup using a robotic arm with OpenText Functional Testing, delivering scalable, accurate, and repeatable validation of all payment methods without affecting real transactions.

Products deployed

Enabled end-to-end POS testing using a robotic arm combined with OpenText Software

In physical retail stores, customers can pay using multiple methods—cash, card (tap, insert, or swipe), loyalty cards, or QR‑code payments through bank wallets. In parallel, especially in markets like South Africa where on‑demand shopping and delivery are essential, Zebra devices are used in‑store to scan and process online orders. Because these devices are critical to the order‑fulfillment process, Pick n Pay needed to include them as part of their test automation coverage.

The process revolves around automating end-to-end testing for point-of-sale transactions using a robotic arm (ARM) and OpenText Functional Testing as the control center.

The ARM interacts directly with devices by picking up cards, tapping NFC, or scanning codes. It connects to OpenText Functional Testing through an API exposed by a custom Windows application developed for the ARM. OpenText Functional Testing sends commands to the ARM and waits for confirmation before moving to the next step, ensuring accurate sequencing and removing human error.

To avoid costly hardware investments, the team implemented two mobile devices mounted on stands in the server room, positioned face to face. One device displays a dynamic QR code for each transaction, while the other scans it, accurately simulating real-world payment scenarios.

This setup is fully integrated with OpenText Functional Testing and OpenText Functional Testing Lab for Mobile and Web, enabling end-to-end testing of customer mobile payment journeys and in-store operations, including scenarios where employees use Zebra Android devices to scan products for online orders. Native integration with OpenText Functional Testing enables scripts to be executed directly on OpenText Functional Testing Lab for Mobile and Web managed environments without script modifications, simplifying execution and maintenance.

By leveraging OpenText Functional Testing together with OpenText Functional Testing Lab for Mobile and Web’s parallel execution capabilities, test suites are distributed across multiple devices simultaneously, significantly reducing regression cycles from weeks to days.

The result is a fully automated, repeatable, and reliable testing environment that mirrors production scenarios without impacting real financial data. It supports various payment methods—credit, debit, loyalty, and QR—and ensures consistent outcomes across different mobile OS versions.

By combining ARM automation, OpenText Functional Testing orchestration, and multi-tier integration, the team achieved a cost-effective, scalable solution that simplifies complex testing while maintaining accuracy and efficiency.

Unified validation across payment types and third-party services

Pick n Pay’s POS now supports added value services like airtime, electricity, fine payments, and banking‑like functions. To test these reliably, it uses OpenText Service Virtualization to simulate third‑party providers via APIs, avoiding reliance on physical devices or unavailable external services.

A single platform for scalable, end-to-end POS automation

OpenText Functional Testing is designed to work across any architecture and development language, enabling enterprise-wide automation without complexity. Teams do not need to manage integrations or data handoffs between multiple tools. A single automation platform is used consistently across the environment, with shared capabilities and reusable functions. The solution supports all major enterprise technologies, including SAP, Oracle, Java, and Python, making it suitable for even the most heterogeneous IT landscapes.

Automation can be created with minimal effort, and end-to-end business processes can be automated without additional overhead. This significantly reduces time to value and operational friction while requiring only one core skill set across teams.

Finally, OpenText Functional Testing offers flexibility in how teams work. It can operate in a fully codeless mode for broader adoption, while still allowing experienced automation engineers to extend and customize functionality quickly and efficiently when needed.

OpenText Functional Testing platform workflow and full end-to-end process

OpenText Functional Testing supports all environments and architectures, enabling seamless end-to-end automation with minimal effort. With one tool and one skill set, teams can optimize processes, create scripts easily.

Leon van Niekerk
Head of Testing Center of Excellence, Pick n Pay

Results

Pick n Pay’s small team of testers and automation engineers delivered high-volume, efficient POS testing. With nearly 100% automation coverage, they reduced regression cycles to 3.5 days, enabled monthly releases, and cut defects drastically.

A small testing team executing 50,000 automated scripts

The team started with seven or eight members but now consists of one senior software tester and two junior testers focused on point-of-sale testing. In addition, there are seven dedicated automation engineers managing robotic arm (ARM) automation. The team operates with about 17 runtime licenses and 13 full OpenText Functional Testing licenses, enabling each automation engineer to work on up to four virtual machines simultaneously. Despite its small size, the team delivers significant output, executing close to 50,000 automated scripts per week. This efficiency is also supported using OpenText Functional Testing Lab for Mobile and Web, which ensures smooth device management and eliminates latency issues, allowing the team to maintain high performance and scalability.

Nearly 100% automation coverage

The point-of-sale (POS) system now achieves full regression testing within about three and a half days per cycle, which has significantly improved release efficiency. Previously, releases occurred quarterly, averaging four per year, but the process has evolved to allow monthly releases, about twelve per year, and even the capability to release every twelve days if needed.

This improvement is largely due to the robotic arm and automation framework, which runs 24/7 and delivers nearly 100% automation coverage for regression testing. The automation team for POS is now larger than the functional testing team, enabling high-volume execution and reducing manual effort, ultimately driving faster, more reliable deployments into production.

The ROI analysis showed that the robotic arm paid for itself within five months. The total cost for the ARM included engineering, on-site setup, calibration, and installation.

Next steps

Pick n Pay’s approach for improvements is to plan thoroughly, run a small proof of concept to confirm functionality, and once validated, roll it out easily for broader use. Currently they are running a set of proof of concepts:

  • First, OpenText Functional Testing for Developer is being used to test APIs and build end-to-end processes that start with mobile app transactions and validate results through backend API calls, with plans to roll it out for the Pick n Pay One app.
  • Second, Postman is being leveraged for API calls because developers already use it, and its collections are integrated into OpenText Functional Testing for Developer for fully automated scenarios.
  • Third, the ARM is being upgraded to become more dynamic by using OCR to analyze screenshots of the POS screen and adjust its actions, accordingly, supported by a webcam for real-time feedback.
  • Finally, the team is exploring AI-driven object recognition in OpenText Functional Testing to identify screen elements from captured images and dynamically click specific areas, adding intelligence to automation for greater flexibility and adaptability. It will reduce maintenance and increase cross-platform efficiency.

How can we help?