CapitaLand’s CapitaStar app embarks on AI-driven product development with Azure AI platform

CapitaLand Investment (CLI), headquartered in Singapore, operates across more than 45 countries and 270 cities worldwide. A leading global real asset manager, CLI focuses on fee income-related businesses, comprising listed and private funds management, lodging management and commercial management, and its real estate investment business. CLI uses digital technology to drive business outcomes and deliver greater value to its customers and partners.

One such value-adding initiative is its CapitaStar program, a lifestyle loyalty rewards platform designed to encourage repeat patronage of its retail services. Running for more than 13 years and currently having 1.8 million members, the program has seen many changes, including the transition of vouchers from paper to digital. The CapitaStar app today includes features like eCapitaVoucher, a gamified rewards system, and personalized promotions. Through these, CapitaStar strengthens the customer base, while bridging the physical retail mall presence with digital initiatives.

Defining goals of modernization

As CapitaStar program’s capabilities expanded, its manually managed DevOps processes were no longer sufficient. CLI needed to rethink its computing resources to meet business demands. Automation was the clear way forward to ensure continued scalability and efficiency.

CLI’s top priority was modernizing its legacy infrastructure while maintaining its robust security controls to protect customer data on the loyalty platform. Leveraging generative artificial intelligence (AI) to meet platform growth goals and drive significant bottom-line impact was another key priority.

“The decision to use Microsoft solutions came down to a few factors,” explains Darrence Tan, Assistant Vice President, Digital International (retail and workspace systems) at CLI. “We weren’t building something from scratch, so we couldn’t afford to use third-party products that didn’t integrate seamlessly with our existing tech stack.”

Ease of use was also key.  Tan notes, “When introducing tools, the product team had to be able to use them easily without requiring lengthy training.” With these needs in mind, CLI chose Microsoft solutions that aligned with its ecosystem, ensuring a streamlined modernization journey.

Moving to scalable architecture

The first step was addressing CLI’s legacy infrastructure by transitioning from virtual machines (VMs) to a managed service. “Traditionally, we relied on a large number of manual VMs. This setup posed challenges during peak retail seasons, such as in December and during Chinese New Year. During festive periods, we had to manually scale the infrastructure, which required time and effort. Everything was reactive and done manually,” Tan adds.

To overcome these limitations, CLI adopted Azure Kubernetes Service (AKS). Over the last three years, CLI shut down all its VMs and moved 90% of its tech stack into Azure Kubernetes. This shift not only enabled CLI to handle business fluctuations more effectively but helped achieve greater scalability and reliability. Today, platform uptime is at a 99.8% standard. There are now fewer production issues, and even if one is detected, recovery time is much faster.

“With Azure Kubernetes Service, we reduced critical priority incidents by 60%,” Tan shares. “The transition from VM-based architecture to AKS also resulted in a 50% cost reduction.”

Automating and streamlining development

The next step was addressing its development processes. Previously, CLI was hindered by siloed data, and coordination between its in-house and outsourced development teams was inadequate. "Our teams faced long feedback loops which slowed down our delivery cycles,” explains Tan.

With Azure DevOps, CLI automated the entire release management process. This also included code scanning capabilities with GitHub Advanced Security for Azure DevOps. With these solutions, CLI streamlined workflows and introduced early-stage security measures into their Software Development Lifecycle.

CLI’s security scanning, integrated directly into the development pipeline, ensures vulnerabilities such as hardcoded secrets are flagged early. “The entire cycle is shortened because fewer issues arise at a late stage of development,” Tan adds. This proactive approach minimizes costly fixes during production and improves compliance.

“With Azure DevOps, we measured a 25% increase in productivity through full end-to-end automation of release management,” shares Tan. Code quality has also improved, with Azure DevOps preventing untested code from being pushed to production.

Finally, CLI used GitHub Copilot to transform how its developers operate. “We started using it early in 2024 and since then, GitHub Copilot has been the biggest contributor to our productivity improvement,” says Tan.

Traditionally, CLI’s developers spent significant time writing and debugging code, which distracted them from more value-adding activities. “With the adoption of GitHub CoPilot, 25% of the code is written by AI and there is still further room for improvement with increased adoption,” Tan adds. The integration of GitHub Copilot also accelerated CLI’s ability to deliver product features. “We saw a 50% increase in product release frequency, enabling faster delivery of app updates or new features,” Tan shares.

Enhanced telemetry and monitoring now provide CLI’s developers with actionable insights. “Through Azure DevOps and GitHub Copilot, we can track how many lines of code have been generated, how many releases have been done, and how many code check-ins developers have performed,” Tan explains. “This visibility and predictability allow us to confidently commit to delivery timelines while reducing transaction failure rates by 70%.”

For the development team, the impact extended beyond productivity. Replacing manual processes like email-based consultations and release management with automated tools has increased job satisfaction and employee retention.

Scaling success across teams

Looking ahead, CLI is preparing to replicate its achievements across other business verticals. “What we have done is act as an innovation lab for the organization. Our leadership team is very excited about the improvements we’ve been able to realize with Microsoft solutions so far,” says Tan. “If we replicate this modernization process across other business verticals, the benefits will be even more substantial. That’s why one of our key priorities is to institutionalize this knowledge and create a playbook.”

The company plans to use this playbook to guide other teams in adopting Azure DevOps, GitHub Advanced Security for Azure DevOps, and GitHub Copilot. Future plans also include introducing Microsoft Dev Box to further optimize costs and improve productivity.

Reflecting on CLI’s modernization journey for CapitaStar so far, Tan emphasizes, “What made our journey successful is the fact that we started small, proved the value, and involved key stakeholders. There are many teams involved, such as the architectural board review team, cybersecurity, product managers, developers, operations, and quality assurance team. When all these people are equipped with the right tools to do their jobs and aligned on the process to do it, then you can move the needle and drive significant bottom-line impact.”

Discover more about CLI on Facebook, Instagram, LinkedIn, X/Twitter, and YouTube.