Agentic AI's Potential and Debate: What It Holds for Southeast Asia

Agentic AI is widely regarded as the next major frontier in artificial intelligence, though its practical implications for businesses are still taking shape. According to the Capgemini Research Institute, agentic AI could generate up to US$450 billion in economic value by 2028. Despite this promise, adoption remains low: a mere 2% of organizations have scaled its use, and trust in AI agents is beginning to weaken.

This gap—high potential but slow deployment—is the focus of Capgemini’s latest research. Drawing from an April 2025 survey of 1,500 executives from major organizations across 14 countries, including Singapore, the report emphasizes trust and oversight as critical factors for delivering value. Almost three-quarters of executives said human involvement in AI workflows delivers benefits that outweigh the costs. Nine out of ten described oversight as positive or at least cost-neutral.

The conclusion is evident: AI agents perform most effectively when working alongside people, not operating entirely on their own.

Early steps, slow progress

Around one-quarter of organizations have launched agentic AI pilot programs, while just 14% have advanced to full implementation. For the rest, deployment is still in the planning phase. The report frames this as a widening gap between ambition and readiness—now considered a key barrier to unlocking economic value.

The technology is far from theoretical—real-world applications are beginning to surface. One example is a personal shopping assistant that can find products based on specific requests, generate descriptions, answer customer questions, and add items to a cart using voice or text commands. Although such tools generally do not complete transactions for security reasons, they replicate many of the functions of a human shopping assistant.

This trend raises broader questions about the future of traditional websites. If AI can handle tasks such as searching, comparing products, and preparing purchases, will users still need to browse online stores directly? For those who find complex websites confusing or difficult to navigate, an AI-driven interface could offer a more streamlined and user-friendly alternative.

Defining agentic AI

To clarify the term, AI News spoke with Jason Hardy, Chief Technology Officer for Artificial Intelligence at Hitachi Vantara, about how enterprises across the Asia-Pacific should approach this emerging technology.

“Agentic AI is software capable of making decisions, taking action, and refining its approach autonomously,” Hardy explained. “Picture it as a team of domain specialists that learns from experience, coordinates tasks, and operates in real time. While generative AI creates content based on prompts, agentic AI may incorporate GenAI but focuses on achieving goals and acting within dynamic environments.”

This distinction—between generating outputs and driving outcomes—lies at the heart of what agentic AI means for enterprise IT.

Why adoption is accelerating

Hardy points to scale and complexity as the primary drivers of adoption. “Today’s enterprises are struggling under the weight of complexity, risk, and scale. Agentic AI is gaining traction because it goes beyond analysis—it optimizes storage and capacity dynamically, automates governance and compliance, predicts failures before they happen, and responds instantly to security threats. This leap from ‘insight’ to ‘autonomous action’ is fueling the acceleration,” he noted.

Capgemini’s research echoes this view. It found that although confidence in agentic AI varies, early implementations are showing value when the technology handles routine yet essential IT functions.

Where value is emerging

According to Hardy, the most established use case so far is in IT operations. “Automated data classification, proactive storage optimization, and compliance reporting save teams valuable hours each day, while predictive maintenance and real-time cybersecurity measures cut downtime and reduce risk,” he said.

The benefits go beyond efficiency. These capabilities enable systems to detect issues before they become critical, allocate resources intelligently, and contain security breaches faster. “Early adopters are already using agentic AI to resolve potential problems before they escalate—boosting reliability and performance in hybrid environments,” Hardy added.

For now, IT remains the most practical entry point: its deployment delivers measurable outcomes and plays a central role in how enterprises control costs and risk, demonstrating the operational significance of agentic AI.

Southeast Asia’s starting point

For organizations in Southeast Asia, Hardy says the top priority is data readiness. “Agentic AI only delivers value when an organization’s data is correctly classified, secured, and governed,” he emphasized.

Infrastructure is equally critical. Agentic AI depends on systems that support multi-agent orchestration, persistent memory, and dynamic resource allocation. Without this foundation, adoption will remain constrained.

Many businesses may prefer starting with IT operations, where agentic AI can prevent outages and boost performance before expanding into broader business functions.

Reshaping core workflows

Hardy anticipates agentic AI will redefine workflows in IT, supply chain management, and customer service. “In IT operations, it can forecast capacity needs, rebalance workloads, and reallocate resources in real time. It can also automate predictive maintenance, stopping hardware issues before they interrupt operations,” he said.

Cybersecurity is another promising area. “In this domain, agentic AI can spot anomalies, isolate compromised systems, and trigger secure backups in seconds—slashing response time and minimizing harm,” Hardy observed.

These capabilities aren’t confined to experimental trials. Early deployments are already showing how agentic AI can reinforce reliability and resilience in hybrid IT setups.

Skills and leadership

Adopting agentic AI will also require new human skills. “It will shift the human role from hands-on execution to supervision and coordination,” Hardy remarked. Leaders will need to define boundaries and monitor autonomous systems to keep them within ethical and operational guidelines.

For managers, this means less time on administrative duties and more focus on coaching, innovation, and strategy. HR teams will need to cultivate governance expertise—such as audit preparedness—and design new frameworks for integrating agentic AI smoothly.

Workforce effects will not be uniform. The World Economic Forum forecasts that AI could create 11 million new jobs in Southeast Asia by 2030 while displacing 9 million. Women and Generation Z are likely to experience the most disruption, with over 70% of women and as many as 76% of younger workers in roles susceptible to AI automation.

This underscores the need for rapid reskilling. Major investments are already happening—Microsoft has pledged $1.7 billion in Indonesia and launched training initiatives in Malaysia and the wider region. Hardy underlined that skill-building efforts must be inclusive, swift, and strategic.

What comes next

Looking ahead three years, Hardy believes many leaders will underestimate the speed of transformation. “The initial wave of benefits is already visible in IT: agentic AI is automating functions such as data classification, storage optimization, predictive maintenance, and cybersecurity response—freeing teams for more strategic tasks,” he stated.

However, the bigger surprise may occur at the economic and business model level. IDC forecasts that AI and generative AI could contribute roughly US$120 billion to the GDP of the ASEAN-6 by 2027. Hardy sees the implications as wider and faster than most anticipate. “This suggests the impact will be more rapid and tangible than many leaders currently assume,” he said.

In Indonesia, more than 57% of jobs are expected to be altered or influenced by AI—a sign that transformation won’t be limited to IT. It will reshape business structures, risk management, and value creation.

Balancing autonomy with oversight

Capgemini’s findings and Hardy’s observations point to a common conclusion: agentic AI offers tremendous potential, but its real-world impact hinges on striking the right balance between autonomy, trust, and human supervision.

The technology can help businesses cut costs, enhance reliability, and open up new revenue opportunities. But without a deliberate focus on governance, upskilling, and infrastructure readiness, adoption progress may stall.

In Southeast Asia, the key questions are not if agentic AI will be adopted, but how fast—and whether companies can blend autonomy with accountability as machines take on greater responsibility for business decisions.

See also: Beyond acceleration: the rise of agentic AI

Interested in learning more about AI and big data from top industry experts? Attend the AI & Big Data Expo in Amsterdam, California, or London. This comprehensive event, held as part of TechEx and co-located with other leading tech gatherings, has all the details you need. Click here for additional information.

AI News is brought to you by TechForge Media. Discover more upcoming enterprise technology events and webinars here.