Agentic AI in Southeast Asia: Balancing Innovation and Trust in the Age of Automation

In the rapidly evolving world of artificial intelligence, the term "Agentic AI" has emerged as a key player, especially in the context of Southeast Asia. This innovative technology promises to shape various sectors, yet it walks a tightrope between excitement and skepticism. Research from the Capgemini Research Institute suggests that Agentic AI could contribute a staggering USD 450 billion to the global economy by 2028. But despite such potential, penetration in enterprises remains low, with only 2% of organizations having successfully integrated it into their systems. As trust in AI-based agents begins to wane, it becomes clear that redefining its role is essential.

The nuances around Agentic AI are intriguing, particularly when you consider the findings from a recent survey of 1,500 executives across 14 countries, including Singapore. It’s evident that trust and oversight are pivotal in leveraging the real economic benefits of this technology. Surprisingly, almost 75% of executives affirmed that human involvement in AI workflows results in greater advantages compared to some costs that may incur. Remarkably, 90% regarded oversight as beneficial, if not at least neutral, pointing to a crucial insight: AI agents flourish when working alongside human collaborators rather than being set on autopilot.

Slowly Getting There

While a noteworthy 25% of businesses have initiated pilot projects, only 14% have moved forward with implementation. The prevalent trend is a focus on planning rather than execution. Such a gap between ambition and readiness remains one of the foremost hurdles in capturing the anticipated economic benefits. Recent applications of Agentic AI, like personal shopping assistants capable of searching, detailing product features, and managing carts through voice or text commands, showcase its evolving utility. Even though financial transactions are often withheld for security purposes, these assistants mimic the functions of a human counterpart.

This reality nudges us to ponder: with AI handling tasks typically requiring human intervention, like product reviews and comparisons, will consumers continue to prefer traditional online shopping? For individuals daunted by complex websites, an AI-driven interface could provide a breath of fresh air, simplifying their shopping journey.

What Exactly is Agentic AI?

To clarify what Agentic AI entails, we consulted Jason Hardy, Chief Technology Officer for Artificial Intelligence at Hitachi Vantara. “Agentic AI refers to software capable of making decisions, taking actions, and refining its strategies independently,” Hardy explained. It operates more like a collaborative team of experts, continually learning and adapting, unlike generative AI that simply reacts to instructions. This distinction truly encapsulates the essence of Agentic AI for technological applications in enterprise settings.

Driving Adoption

Hardy attributes the growing interest in Agentic AI to the increasing complexity of enterprise environments. Organizations today face substantial complexity, risk, and scale. Agentic AI addresses these issues effectively—it doesn’t just analyze data but optimizes various operations in real-time, from governance to security threats. This shift from mere insight to executing autonomous actions is pivotal in fueling quicker adoption.

Supporting this trajectory, Capgemini’s survey revealed that while attitudes toward Agentic AI may vary, initial deployments have proven beneficial, particularly for essential IT-related tasks.

Emerging Value

IT operations currently represent the most promising applications, according to Hardy. This includes automated data classification, proactive storage optimization, and comprehensive compliance reporting that significantly reduce workloads. Predictive maintenance tied with real-time responses to cybersecurity events can mitigate risks and improve uptime, showcasing the value of Agentic AI beyond just efficiency.

For enterprises eyeing this technology, getting data processes right is a non-negotiable first step. Proper classification, security, and governance of enterprise data are essential for realizing any substantial benefits from Agentic AI.

Transforming Workflows

As Hardy envisions it, Agentic AI has the potential to redefine workflows across information technology, supply chain dynamics, and customer service realms. In IT operations, it might foresee capacity requirements, balance workloads, and readjust resources dynamically. Similarly, its application in cybersecurity can expedite anomaly detection and timely responses, safeguarding assets effectively.

The ripple effects of these capabilities transcends mere proof-of-concept trials. Early deployments demonstrate how Agentic AI strengthens reliability amid complex ecosystems.

The Skills Landscape

This transition calls for a refreshed skill set among the workforce. “The role of people must pivot from executing tasks to overseeing processes and systems,” Hardy pointed out. It’s about setting boundaries and maintaining vigilance over autonomous systems to keep them within ethical lines.

For management, this translates to devoting less time to routine tasks and instead focusing on mentorship, strategic planning, and innovation. Human Resources will also need to cultivate governance skills and integrate structured frameworks for effectively implementing Agentic AI technologies.

The implications for job markets are significant, with the World Economic Forum estimating that AI may create 11 million positions while displacing roughly nine million in Southeast Asia by 2030. Vulnerable demographics, particularly women and younger generations, may experience more substantial disruptions, suggesting an urgent need for reskilling initiatives.

The Road Ahead

Looking ahead, Hardy urges consideration that many leaders may miscalculate the speed of changes. Rapid automation processes underway in IT—like data classification and cybersecurity responses—are already easing operational burdens. But perhaps the more profound changes lie in how strategies are crafted at broader organizational levels.

For Southeast Asian enterprises, the pivotal consideration isn’t simply whether Agentic AI will become integrated, but how smoothly enterprises can juggle independence and supervision as technology takes the reins on various business fronts. In this transformative space, infrastructure, data quality, and effective oversight will be the backbones supporting not just operational success but the ethical deployment of AI as well.