Enterprises are eager to reap the benefits that artificial intelligence (AI) promises to deliver, but many remain in the early stage of adoption and are mulling over how to get things moving.
Organisations still are trying to figure things out, such as where to start and what to do, said Rondy Ng, Oracle’s executive vice president of applications development.
They do not want to miss out on the AI bandwagon and are eager to tap the potential the technology promises to deliver, said Ng, who was speaking to FutureCIO on the sidelines of Oracle AI World Tour in Singapore.
There is a lot of pressure from management to tap the key business benefits, including improved productivity and operational efficiency, he said.
“But, most of the time, we see people trying to figure out a strategy,” he noted, adding that most companies would have small AI projects running to meet demands from management and assess the impact of these initiatives. “They want to see what actually works for them.”
In fact, some organisations in Asia-Pacific have hit a wall and are facing difficulties advancing their AI adoption.
While almost 90% have taken the first steps in rolling out AI initiatives, 71% are stuck in the “builder” phase, where they struggle to scale their pilot projects into actual production environments capable of delivery measurable returns on investment (ROI).
Just 17% of organisations are deemed “future-ready” with their AI efforts, running scalable infrastructure and mature data governance and operational expertise, according to research commissioned by ST Telemedia Global Data Centres. Conducted by Ecosystm, the study surveyed more than 600 enterprises across nine Asian markets, including India, Japan, Singapore, South Korea, and Thailand.
The report revealed that respondents prioritise baseline requirements, such as security and reliability, while citing operational expertise, scalability, and cost efficiency as their top challenges.
In Singapore, at 40%, more organisations than the regional average (17%) have moved beyond early-stage AI pilots. However, just 3% of Singapore enterprises are considered in the “leader” stage of AI infrastructure maturity.
"Across Asia, organisations are moving quickly from experimentation to implementation, but many are discovering that AI success now depends less on training models and more on foundations," said Chris Street, group chief revenue officer, ST Telemedia Global Data Centres. “Without scalable infrastructure and operational readiness in place, it becomes difficult to convert early AI ambition into consistent business value.”
Tracking potential faults with AI
One Singapore company is taking steps to extract measurable gains from its AI initiatives, specifically, in predicting potential system failures and improving workplace safety.
Public transport operator SMRT designed its Joint Automated Repository for Various Information Systems, or Jarvis, to pull data from across its rail network, spanning sensors and monitoring systems, and power advanced visualisation and analytics capabilities.
The AI platform taps large language models (LLMs) to power insights and generative AI (GenAI) chatbots, providing a range of information, such as how trains are performing and where faults might surface.
Developed by SMRT’s tech and engineering arm Strides Technologies, Jarvis is touted to provide relevant data that will enable SMRT to better predict system issues and plan maintenance schedules. This then will lead to faster fault resolution, according to SMRT.
The transport operator operates a rail network that supports more than 2 million passenger journeys daily across Singapore.
SMRT this past week announced a new pilot with Oracle, in which Jarvis will run on Oracle Cloud Infrastructure (OCI) Enterprise AI and Oracle Autonomous AI Database. Jarvis will run on the latter as its core data platform, analysing maintenance data across the rail network, including sensor readings, train performance, and asset lifecycle information.
Jarvis also was developed and tested via the Oracle AI Customer Excellence Center in Singapore.
SMRT assesses technology, including AI, on how it can enable staff to do their jobs better. This includes employees in the front-end who support commuters, and in operations room or depot, where robots and other heavy equipment are deployed, said Ngien Hoon Ping, SMRT’s group CEO, who was speaking at Oracle AI.
Technology could allow SMRT engineers to better focus on their tasks, without having to rely on physical attributes, he said. “We still need engineers [to do the work]. It’s really about value adding,” Ngien said.
And with SMRT already sitting on a raft of data, including system failure patterns and encompassing flow charts -- most of which are digitised -- Jarvis will allow its engineers to more quickly analyse the data and identify the cause of failure in the event of an incident, he said.
“We want to be able to detect faults and issues before they pop up,” he said.
With Jarvis still in phase one of development, Ngien noted that SMRT operates a complex system and hopes to continue improving its operations, while tapping the tools available.
More than 50 engineers currently are participating in various tests on Jarvis, including coding of AI agents and chatbots, and evaluating, for instance, how closely AI-powered user experience can match human-assisted experience.
Ngien added that he wants all SMRT engineers to be familiar and comfortable with AI.
And it seems 64% of business leaders in Singapore are applying AI across their daily workflows, with 18% using AI agents, according to research by HubSpot. The study polled 736 respondents in Singapore.
Some 43% say trust and reliability are the top barrier to scaling their AI deployments, while 37% point to data quality and integration as their key challenge.
Amongst those already using AI agents, 42% cite legacy system limitations as a challenge and 41% point to data integration. Another 39% highlight skills gap as their top struggle.
“The key challenge amongst Singapore businesses is no longer whether they are using AI. It is whether they have the knowledge of customers, market trends, and operations needed to scale the business reliably,” said Megan Hughes, HubSpot’s Asia-Pacific Japan managing director and vice president.
"Only by powering AI with customer data and process understanding can businesses consistently transform generic outputs into tangible results,” Hughes said in the report. “The most successful businesses will be those with a context advantage, combining leading-edge AI models with deep context to deliver highly impactful outcomes and create reliable digital teammates for sustained growth.”
Ng further pitches an embedded AI infrastructure as critical to provide organisations with a cohesive software environment, from which to access seamless AI workflows and optimised outcomes.
He pointed to Oracle’s Fusion Cloud Applications, with embedded AI agents and GenAI capabilities, that he said enabled customers to turn on AI features directly within the applications, including supply chain and finance management software.
