Implementing AI-powered Resourcing Decisions to Improve Response Time and Reduce Waste


Performance assurance with software continuously making resource decisions.



Improvement in application response time after implementing AI-powered resourcing decisions


  • Africa’s largest supermarket retailer
  • More than 2,892 stores and 20M shoppers across the African continent
  • Largest private sector employer in South Africa with over 140,000 employees
  • 2020 revenues: 41.7 billion ZAR
  • Hybrid cloud environment including Azure and AWS footprint


Understanding Shoprite

With over 2,892 stores and over 20 million loyal shoppers, Shoprite is Africa’s largest grocery retailer providing communities with a variety of food products, household goods, home appliances and value-added services at the lowest prices. Their customer-oriented business strategy has allowed Shoprite to grow exponentially since first opening their stores in 1979, create jobs for 142,000 workers and generate approximately 41.7 billion ZAR through December 2020.

Reducing Overprovisioning to Improve Application Performance

As Africa’s largest supermarket retailer, Shoprite is always looking for new ways to deliver an exceptional experience to shoppers. Whether they are expanding their virtual voucher program or iterating on their order delivery system, Shoprite is an organization that prides itself on efficiency and high-quality service. Assuring application performance while minimizing spend is critical to Shoprite’s growth and success.

When Avinash Rajkumar, Data Center Team Lead at Shoprite, first took over Shoprite’s hybrid cloud environment, he immediately saw an opportunity to reduce overprovisioning and improve application performance. In the past, when developing new services, application owners requested more resources than necessary because there was no way to determine precisely what each new application required in the long run. Moreover, there was a widespread misconception that allocating more resources to a given application would yield better performance. Despite this tendency to overprovision, the teams often observed performance issues. Avinash knew they needed to reallocate their resources and increase density in order to achieve a faster response time—but needed the right evidence to support this hypothesis. Additionally, the application and virtualization teams needed a single source of truth to help them better understand the state of their environment on an ongoing basis and get new services to market faster.

Historically, application teams were required to consult with virtualization teams each time a new system was deployed. This process required extensive manual work on the part of the virtualization team which needed to calculate the remaining resources in the cluster and allocate resources to other VMs. This time-consuming and highly manual approach could not continue indefinitely.

In addition to supporting the deployment of new services, Shoprite’s IT organization needed to scale out existing services such as the order delivery system and virtual voucher system which offers customers additional savings based on their shopping habits. Adoption of those services had grown increasingly and Avinash and his team needed to make sure they had the capacity to meet demand during the peak holiday season.

In summary, to keep up with their rapid pace of innovation, the Shoprite IT organization needed a single source of truth to understand their hybrid cloud environment and they needed trustworthy resource insights in order to improve cross-functional collaboration and assure application performance.

“That was the defining moment where we got the most skeptical team in the company to accept that Turbonomic will bring value, performance and cluster management tools better than any human can.”

Avinash Rajkumar
Data Center Team Lead

Building Trust in AI-Powered Automation

Avinash started by approaching the application owners with a plan to reallocate resources. Using Turbonomic AI-powered resourcing recommendations, he advised them to implement Turbonomic’s recommended actions and check back in a day to see how response time was impacted by this set of decisions. One day after they implemented these recommendations manually, they observed application response time was 35% faster. Encouraged by this result, the team chose to implement automated resourcing decisions. One week later they observed their applications running 85% faster compared to when they began.

Expanding Automation and Shortening Development Timeline

With concrete evidence that trustworthy resourcing decisions could improve application response time, the team began a series of rightsizing actions across their environment. This yielded amazing results for application performance and consistency, evidenced by a reduction in performance-related support tickets being submitted. It also ensured that they would no longer need to rely on overprovisioning because application owners could trust that their applications would have the resources they need when they needed them thanks to Turbonomic’s automated resourcing actions. In fact, Turbonomic Application Resource Management has allowed the Shoprite team to keep their workloads optimally placed since 2018.

Avinash and his team were also able to shorten their timeline for deploying new solutions once they turned to Turbonomic. Using Turbonomic’s planning functionality, he and his team greatly reduced the time it took to produce capacity requirements. They also were able to determine the optimal configuration upfront thereby reducing their hardware costs in the long run.

Finally, Turbonomic’s full-stack visibility has helped Avinash stay up to date on resource utilization and possible future constraints. This visibility also gave application owners confidence that their services were performing at their optimal level and would continue to do so in the future.

“Turbonomic has assisted us in realizing our goals of precision retail by ensuring our workloads are correctly sized and placed to assure application performance.”

Avinash Rajkumar
Data Center Team Lead

Looking Ahead

As they move forward, Avinash and his teams are planning to manage containerized platform on-premises for easier application portability across their hybrid cloud environment. They are also aiming to expand their cloud footprint in both Azure and AWS.

What Can Turbonomic Application Resource Management Do for You?

See for yourself with a personalized demo.