Singapore Airlines boosts software delivery with AI approach

Singapore Airlines has collaborated with PALO IT to implement an AI-driven software engineering approach intended to improve productivity and efficiency within its digital product teams.

The engagement involved the deployment of PALO IT's Gen-e2 methodology, which makes use of Generative AI to automate and streamline software development processes. The initial phase was focused on a proof of concept (PoC) with one of Singapore Airlines' digital product squads and resulted in notable improvements across key development metrics.

The Gen-e2 methodology, described as Generative AI Enhanced Engineering, reframes enterprise software development to be AI-first, integrating tools such as GitHub Copilot throughout the project. This approach aims to transform the traditional Software Development Life Cycle into a more agile and cohesive Product Development Life Cycle.

According to PALO IT, during the five-week PoC, GitHub Copilot was instrumental in producing the majority of the software code and supporting materials, with AI responsible for generating 95 percent of the code, documentation, architecture diagrams, and infrastructure as code. The resulting output enabled the team to deliver a new website feature for Singapore Airlines in significantly less time than predicted with standard methods.

The PoC demonstrated that the feature could be prepared for user acceptance testing in five weeks, compared with a forecast of nine weeks using traditional development approaches. This equated to a 44 percent improvement in project delivery time. The initiative also contributed to a sharp increase in build success rates, reaching over 90 percent by the third week of development.

The enhanced approach impacted not only productivity and quality but also the development team's overall experience. Feedback gathered from the team's weekly surveys highlighted a positive sentiment towards the methodology and indicated improvements in software engineering practices.

PALO IT and Singapore Airlines stated that these results have laid the foundation for further exploration of the Gen-e2 model, with the intention to expand AI-driven development methodologies across other IT teams within the airline.

The PoC arose from a collaboration built over two years between PALO IT and GitHub. The companies have worked together to implement Copilot's co-programming capabilities and to further integrate AI into the entirety of the software development lifecycle.

Sharryn Napier, Vice President, APAC at GitHub, commented on the partnership:

"Strategic partnerships are key as we extend the reach of GitHub's Copilot-powered platform. We're excited to deepen our partnership with PALO IT and enable more companies to unlock innovation at scale through AI-driven software development, just as we've seen with Singapore Airlines."

The collaboration sought to examine not only technical effectiveness but also how development teams adapt to AI-centred methodologies. Team feedback indicated that the Gen-e2 approach was easy to adopt and effective in enabling developers to achieve higher standards of speed, coverage, and accuracy from early in the project timeline.

Automation supported by Generative AI during the PoC was credited for the higher build success index and for allowing the squad to move multiple phases of development forward in parallel, thereby reducing bottlenecks commonly associated with linear workflows.

The team's experience formed a central part of the evaluation criteria for the PoC. According to the companies, weekly sentiment surveys and project analytics together demonstrated both measurable engineering gains and increased engagement among the developers using the AI-powered workflow.

PALO IT and Singapore Airlines indicated that the productivity gains and positive internal response from the initial squad are encouraging further rollout of the methodology across wider digital and IT teams within Singapore Airlines. The organisations are now planning to scale up the initiative and continue assessment as more teams adopt the model.