Agentic AI in Southeast Asia: Striking the Balance Between Autonomy and Human Oversight

A Technological Frontier with Human at the Helm

Agentic AI—systems that act, learn, and strategize autonomously—are generating buzz as the next frontier of artificial intelligence. In Southeast Asia, this emerging wave carries the promise of transformative economic gains, but also the sobering reminder that human oversight remains essential. Capgemini’s latest research places this tension squarely at the center of the conversation.

The Promise Meets the Reality

Capgemini Research Institute estimates agentic AI could unlock a staggering $450 billion in economic value by 2028. Yet, adoption remains cautious. Only 2% of organizations have scaled deployment, with most still navigating pilot phases or early planning. This cautious approach highlights a broader dynamic—enormous potential restrained by the need for trust, infrastructure, and clarity on outcomes.

What Defines Agentic AI?

Jason Hardy, Chief Technology Officer for AI at Hitachi Vantara, explains that agentic AI goes beyond the passive responsiveness of generative models. These systems operate more like a coordinated team of domain experts. They act proactively, learn from experience, strategize in real time, and execute tasks toward defined objectives in dynamic environments. The leap from generating text to driving real-world results marks a fundamental shift in how enterprises can leverage artificial intelligence.

Early Use Cases Show Promise

Initial applications of agentic AI in Southeast Asia are already delivering concrete results, particularly in IT operations and cybersecurity. In IT, agentic systems are automating essential tasks such as data classification, storage optimization, compliance reporting, and predictive maintenance. In cybersecurity, they’re proving invaluable by identifying anomalies, isolating threats, and initiating fail-safes like backup recovery—often in real time. These outcomes highlight the technology’s ability to elevate both routine and high-stakes workflows across industries.

Laying the Groundwork for Adoption

Hardy emphasizes that successful adoption depends on more than enthusiasm. Companies must first ensure their data is well-classified, secure, and governed. Beyond data hygiene, core technologies like multi-agent orchestration, persistent memory, and real-time resource allocation must be in place to make agentic AI scalable. For most businesses, a strategic rollout begins with IT operations, where the risks are manageable and the return on investment can be proven quickly before expanding into more customer-facing or supply chain systems.

Economic and Workforce Impacts

The broader economic implications for Southeast Asia are profound. IDC projects that AI—including generative AI—could contribute as much as $120 billion to the GDP of ASEAN-6 countries by 2027. Hardy believes that agentic AI, in particular, may accelerate that trajectory even faster than current models suggest. However, this growth won’t come without workforce disruption. In Indonesia, more than 57% of job roles may be altered or displaced by AI. The World Economic Forum predicts AI could create 11 million new jobs in Southeast Asia by 2030 while displacing 9 million, with disproportionate effects on women and younger workers.

Companies are already responding. Microsoft has pledged $1.7 billion toward AI infrastructure and education in Indonesia and launched significant reskilling programs across Malaysia and other regional markets. These investments are designed not just to build technological capacity but also to soften the social and economic friction that rapid AI deployment inevitably brings.

The Human Imperative

A consistent theme across Capgemini’s findings is the irreplaceable role of human oversight. Nearly three-quarters of executives believe the cost of human involvement in AI workflows is justified, and nine in ten say oversight is either cost-neutral or beneficial. Rather than eliminating human jobs, agentic AI changes the nature of work itself. The human role shifts from task execution to strategy, governance, and coordination—roles that require new skills in auditing, ethics, and operational oversight.

According to Hardy, the future of agentic AI hinges not just on its technical capacity, but on how organizations prepare their leaders, HR teams, and governance bodies to manage it. The era of “set-and-forget” AI is over; today’s agentic systems demand continuous human engagement, not just in development but in day-to-day operations and decision-making.

A Vision Forward

To successfully navigate the era of agentic AI, Southeast Asia must prioritize foundational readiness. That means building strong data and orchestration infrastructure, deploying AI strategically in low-risk environments like IT, and investing heavily in workforce development. But it also means constructing a culture of trust and transparency around AI. Governance frameworks will need to be robust and enforceable, balancing innovation with accountability.

Autonomy with Accountability

Agentic AI in Southeast Asia represents more than a technological leap—it is a social, economic, and ethical inflection point. The region stands poised to harness significant value from AI-driven autonomy, but only if it keeps human agency at the core of that transformation. The real story isn’t about machines acting independently—it’s about how we guide those machines to act wisely, responsibly, and in harmony with the needs of the societies they’re meant to serve.