How Visa, DP Architects and Ocean Network Express are building scalable AI solutions with Singapore’s AI Trailblazers Initiative

Singapore’s AI Trailblazers Initiative is designed to help businesses bridge gaps in talent and skills and in data and infrastructure through gaining access to a conducive testbed to experiment, prototype, and scale AI solutions quickly and confidently. 

A joint initiative between the Ministry of Digital Development and Information (MDDI)​​, Digital Industry Singapore (DISG), Smart Nation Group (SNG) and Google Cloud, Singapore’s AI Trailblazers Initiative has supported up to 84 organisations and 100 use cases since its inception in 2023. The programme offers a safe, collaborative environment for companies to prototype, test, and deploy real-world solutions, with full access to Google Cloud’s suite of AI tools and infrastructure. Discover the journeys of three companies that used the sandbox provided by the AI Trailblazers Initiative to innovate and create real-world AI applications:

 

AI Trailblazers Case Study 1: Visa’s personalised itinerary planner for savvy travellers

Visa, an international payments network company, assembled a team of experts across Innovation, Design, Data Science, and Technology to take a holistic approach to product creation. As Sonia Yap, Director of Innovation at Visa, noted, this initiative is built upon the company’s 30 years of pioneering AI leadership and over US$3 billion (S$3.87B) in AI investments.

AI MISSION / PROBLEM STATEMENT

Travel planning is often tedious, with users juggling multiple platforms and generic recommendations based on popularity. As travellers increasingly seek personalised experiences, Visa aimed to create end-to-end itineraries tailored to individual preferences. 

Although generative AI has become a popular tool for itinerary creation, Yap noted risks around “poor-quality or biased data”. For Visa, the goal was not only to personalise travel, but to do so responsibly.

THE AI TRAILBLAZER DIFFERENCE

Visa gained access to a suite of powerful tools—including the Gemini Large Language Model (LLM)—which accelerated prototype development and reduced time and cost. With direct, hands-on support from Google Cloud experts - the team quickly upskilled and learned to use tools that were previously unfamiliar to them, like Vertex AI, which helped streamline workflows. The tool provides companies with a unified, end-to-end platform that consolidates Machine Learning process from data preparation, training, and deployment. Teams can now move from experimentation to production faster and more reliably.  

THE OUTCOME

By synergising Visa’s robust data capabilities with Vertex AI, the team built a highly personalised itinerary planner grounded in real user behaviour. With consent, it draws on cardholder spending data to suggest tailored destinations, flights, accommodations, and restaurants.

Working with legal and privacy teams, Visa developed clear terms and transparent processes to “ensure that we implemented proper risk management and governance”, Yap notes. Since building the prototype, Visa has expanded its features, Yap explains: “We are continuing our efforts to gather more feedback on our solution, both internally and from our clients.”

AI Trailblazers Case Study 2: Optimising the smart city design process at DP Architects

DP Architects (DPA), a global multidisciplinary design firm, brought together its Digital Transformation team to explore real-world architectural use cases through architectural and digital innovation.

AI MISSION/PROBLEM

Today, 55 per cent of the world’s population lives in urban areas; this proportion is expected to increase to 68 per cent by 2050¹. This upward trajectory means DPA needs a solution for its challenge of optimising limited real estate within shorter design cycles. With existing design processes still largely manual, speed and quality vary by architect. To improve consistency and output, DPA set out to develop an AI model that could accelerate and standardise design iterations.

“We quickly realised that the output we want to achieve is unobtainable with the models accessible to us,” said Chan Hui Min, Head of DPA’s Smart Sustainability Unit: “Hence, we had to develop and train our own AI model.”

THE AI TRAILBLAZER DIFFERENCE

With access to Gemini LLM and AI expert support, DPA gained valuable clarity in “demystifying development of AI products”. This has also enabled rapid prototyping and training of internal teams without needing to manage the costs of setting up test environments.

THE OUTCOME

Within eight weeks, DPA developed a model that boosted productivity by over 50 per cent and generated 10 times more design iterations in the same amount of time. Built with curated architectural data and synthetic data generation techniques, the tool also enhanced transparency, allowing outcomes to be traced back to original templates for easier audit and review. 

Now integrated into multiple digital tools, the model automates technical drawings and streamlines reviews. More importantly, it marked a broader shift within the firm. Chan said: “[The programme] has made us more open to exploring and implementing AI-driven solutions in various aspects of our work.”

AI Trailblazers Case Study 3: Faster access to shipping data at Ocean Network Express

Ocean Network Express (ONE), one of the world’s largest container shipping companies, brought together seasoned professionals and emerging talent to explore AI solutions grounded in real-world business needs.

AI MISSION/PROBLEM STATEMENT

ONE’s sales representatives needed a faster way to retrieve booking information to provide customers with timely logistical information, previously only accessible through a complex, tedious manual process. 

Guided by ONE’s commitment to customer satisfaction, Data Scientist Lim Zheng Sen said: “Our decision to develop a chatbot was driven by a need to provide timely, accurate information [to our customers].” 

To create the chatbot, ONE refined a Text-to-Structured Query Language (SQL) chatbot that translates plain English prompts into a language that databases understand, eliminating the need for sales representatives to know technical syntax. 

For example, a prompt like “Give me the SC RFA NO (short for Service Contract and Rate File Activity number) for shipper customer name NIPPON” is instantly converted into SQL, a language that databases understand. 

The team ensured that the chatbot could also be simultaneously used by multiple users while ensuring seamless performance despite the heavy workload. 

THE AI TRAILBLAZER DIFFERENCE

With access to Google Cloud’s advanced AI infrastructure and expert support, the team ensured the chatbot could scale to thousands of concurrent users. Guided by Google experts, ONE adopted AlloyDB, a reliable, high-performance smart database that simplifies data management processes.

THE OUTCOME

ONE is rolling out a faster, more accurate, and updated chatbot to serve potentially thousands of sales personnel.

The initiative also reshaped how the company approaches AI. “Singapore’s AI Trailblazers Initiative has transformed our view of AI from an individual pursuit to a collaborative effort,” shared Lim. Now, AI is a company-wide priority. “We actively explore diverse applications—from forecasting vessel capacities to querying organisational charts,” added Kentaro Yuasa.

Team member, Justin Siow reflected that the experience “affirms Singapore’s role as a global innovation hub” and that his team “is grateful to operate in a country that actively fosters innovation”.

Access our one-stop online resource for developments in the AI sector in Singapore and the region.

Follow us on LinkedIn to get more updates or stay in the loop with our newsletter.
 

---

Footnote:

¹ United Nations. World Urbanization Prospects: The 2018 Revision. United Nations Department of Economic and Social Affairs, Population Division, 2018, p. 1.