Infrastructure and governance are key for GenAI adoption
Many generative AI projects are failing to scale because firms are overlooking foundational needs like network modernisation and governance frameworks, says a regional leader at NTT Data
The rush to embrace generative artificial intelligence (GenAI) is hitting a wall, with many proof-of-concept (POC) projects failing to scale due to inadequate infrastructure and a lack of robust governance, according to the head of NTT Data in Asia-Pacific.
Speaking to Computer Weekly in a recent interview, John Lombard, CEO of NTT Data Asia-Pacific, said that before organisations can truly harness the power of GenAI, they must first address their foundational technology debt and establish clear governance frameworks.
“When organisations started looking at GenAI, there was a rush to do these POCs,” said Lombard, adding that many POCs didn’t make the jump into fully fledged deployments because organisations found that they had to follow a process to deploy the technology in the right way and look at their underlying infrastructure.
Lombard pointed to recent research from NTT Data, which found that 90% of organisations are currently reviewing their existing infrastructure – including networking, storage and security – which must be modernised before enterprises can confidently deploy AI applications at scale.
NTT Data, formed through the integration of Dimension Data and other NTT companies, is positioning itself to address this with offerings from connectivity to advanced AI applications. The company now has a dedicated Asia-Pacific workforce of 15,000 employees, all of whom have completed a fundamental AI certification with an emphasis on governance and ethics.
From GenAI to AI agents
Lombard noted that the conversation with customers is shifting from generative AI to agentic AI, where autonomous agents can perform complex, multi-step tasks without human intervention. To help clients make this leap, NTT Data has developed “smart AI agents” capable of executing tasks autonomously in response to user instructions.
“With millions of RPA [robotic process automation] bots out there today, we’ve also developed plugins that allow organisations to convert RPA bots into agentic AI bots,” said Lombard, adding that this will help organisations avoid incurring technical debt from older technologies.
We’ve developed plugins that allow organisations to convert RPA bots into agentic AI bots
John Lombard, NTT Data Asia-Pacific
He also shared examples of where organisations are already deploying AI agents, going beyond simple rules-based automation.
In healthcare, one organisation is using autonomous agents to classify and prioritise insurance claims and appeals to ensure the most impactful cases are handled first. “Their near-term roadmap is very much looking at specialising around early interventions, medical compliance, payer validation and preventing fraud,” said Lombard.
In the automotive industry, another client is using agents to analyse regulatory warning letters and citations. “They’re also developing specialist automation and processes to look at the root cause analysis of vehicle defects,” he added, noting the potential to reduce the high costs of vehicle recalls.
‘Rubbish in, rubbish out’
Despite the potential of AI, Lombard repeatedly stressed the importance of putting governance before technology. He warned that without proper guardrails, AI systems can perpetuate and amplify existing biases.
“It’s the old story, you know, rubbish in, rubbish out,” he said. “It’s so important that the data organisations are using for training doesn’t have bias. You don’t want to have a situation where the algorithm is using data that was already flawed to make decisions on, for example, loan approvals.”
NTT Data’s approach is to begin engagements with advisory services focused on ethical, regulatory and compliance considerations. “We advise our clients to build a very strong and robust governance framework,” Lombard said. “It is really important to look at that before you jump into the technology.”
Underpinning NTT Data’s AI efforts is its investment in the region’s core infrastructure. The company is the third-largest datacentre provider in Asia-Pacific and is expanding capacity in India, Thailand, Indonesia and Malaysia to support high-demand AI workloads.
Lombard also pointed to the recent commissioning of the Mist submarine cable that connects Malaysia, India, Singapore and Thailand to the rest of the world. Spanning 8,100km, it carries over 200 terabits per second of bandwidth and is an “important connection for this part of the world to open up the use of AI to a broader cohort of people”.
Read more about AI in APAC
Microsoft co-founder Bill Gates underscores AI’s transformative potential in healthcare, education and agriculture while highlighting the need to ensure its benefits reach lower-income nations.
The AI Cloud Takeoff initiative, part of the Singapore government’s Enterprise Compute Initiative, will offer local firms up to S$500,000 in incentives to establish in-house AI centres of excellence and build AI capabilities.
