https://www.computerweekly.com/news/366626981/How-GoTos-high-stakes-cloud-shift-is-powering-Its-AI-future
Indonesian technology giant GoTo is embarking on a “pragmatic and problem-driven” AI strategy, building on the back of what its group chief technology officer, William Xiong, described as “one of the most complex and challenging cloud migrations ever”.
Speaking at a customer panel at the Alibaba Cloud Global Summit 2025 in Singapore, Xiong detailed how the company migrated half of its infrastructure to Alibaba Cloud, a move that has set the stage for its next phase of AI-powered innovation.
The nine-month project involved moving GoTo Financial’s infrastructure onto Alibaba Cloud. This represented 50% of GoTo’s business and was critical due to the scalability and security needs of financial services. The migration covered tens of petabytes of data across 9,000 services, while ensuring zero downtime for the tens of millions of users, merchants and drivers who rely on the platform daily.
The company now uses Alibaba Cloud’s MaxCompute big data platform for all its services, from on-demand transport to e-commerce and financial services. For its growing lending business, which comprises over 500 microservices, GoTo has adopted the cloud-native PolarDB database to deliver high-performance, low-latency services. Previously, the company relied heavily on self-managed open-source databases such as MySQL.
“In the beginning, no one believed it would happen,” Xiong said. “The migration has enhanced the scalability and resilience of our data platform. By delivering cost efficiency, performance parity and operational continuity, this collaboration strengthens the technical foundation for GoTo’s ecosystem.”
Xiong attributed the success of the high-stakes migration to “extremely good organisation”, rapid decision-making, a strong internal team of thousands of engineers and a close partnership with Alibaba Cloud. “We worked as one team, which is very important,” he added.
With a more unified and scalable infrastructure in place, GoTo is now turning its focus to AI. However, it is taking a measured approach, based on the company’s guiding principle, which is to be “pragmatic and problem-driven”, Xiong said. “We want to have real business problem statements and use AI to excel and provide better user experiences and services to different type of clients,” he added.
Central to this strategy is Sahabat AI, GoTo’s own large language model, which it developed with local partners. Xiong said the goal is not to compete with other foundation models, but to create an open ecosystem that provides the best support in Indonesia, including its many local languages and dialects.
At the same time, GoTo remains open to using other models, including Alibaba Cloud’s Qwen, for specific use cases such as video generation that Sahabat AI does not yet support. “We are open to any suitable models to support our business,” Xiong said.
GoTo’s growing use of Alibaba Cloud came as the Chinese tech giant marked its 10th anniversary in Singapore by launching its first AI global competency centre, which aims to accelerate AI adoption by supporting more than 5,000 businesses and training 100,000 developers globally. The centre is part of a broader regional push that includes new datacentres in Malaysia and the Philippines to meet rising demand for AI and cloud services.
Besides Alibaba Cloud, GoTo is also using Tencent Cloud to run systems that power its on-demand services following a recent migration. Executed under cross-border and multicloud conditions, the migration involved more than 1,000 microservices. Despite its complexity, the migration was completed in around five hours – more than an hour ahead of schedule.
In September 2024, GoTo also teamed up with Google Cloud to tap the latter’s generative AI services to power Dira, GoTo’s AI-based voice assistant in the GoPay app. The voice assistant understands commands in Bahasa Indonesia, making it the first tool of its kind in Indonesia’s fast-growing fintech industry to support the country’s national language.
03 Jul 2025