Defining Agentic AI
The buzz around agentic AI is growing, yet enterprises still debate what it truly means. Capgemini estimates the technology could unlock US$450 billion in value by 2028, but adoption remains low. Only 2% of organisations have scaled its use. Capgemini’s findings highlight that oversight and human involvement remain essential. Nearly three-quarters of executives believe human guidance outweighs the costs, showing that AI agents perform best when paired with people.
From Pilots to Real Applications
About a quarter of firms have launched pilots, while only 14% are implementing systems. Examples are emerging, such as AI shopping assistants that can search, generate product descriptions, and place items in carts. These tools stop short of transactions for security reasons, but they show how agentic AI can replicate many assistant-like functions. Such cases raise bigger questions about the future of traditional websites, as users may prefer AI-driven interfaces that simplify browsing.
Distinguishing Agentic from Generative
Jason Hardy, CTO for AI at Hitachi Vantara, describes agentic AI as software that can decide, act, and refine strategies on its own. Unlike generative AI, which produces content reactively, agentic AI pursues objectives and adapts in real time. Hardy calls it “a team of domain experts that can learn from experience, coordinate tasks, and operate dynamically.” This distinction highlights why enterprises see potential beyond content creation.
Read Also
Why Your SEO Isn’t Working
AI in Marketing
AI Security Wars – Google Cloud
Where Value Emerges
Hardy points to IT operations as the strongest use case so far. Automated data classification, storage optimisation, and compliance reporting reduce manual workloads. Predictive maintenance and real-time cybersecurity responses improve reliability and lower risk. These capabilities allow systems to anticipate issues and act before problems escalate. For enterprises, IT offers a practical entry point that delivers measurable results.
Southeast Asia’s Starting Point
For Southeast Asian businesses, Hardy stresses the importance of strong data foundations. Agentic AI only works effectively when data is classified, secured, and governed. Infrastructure readiness is also vital, requiring systems that support orchestration, persistent memory, and dynamic allocation. Many firms will begin with IT operations, where the gains in performance and reliability are immediate, before moving into broader functions.
Reshaping Workflows
Hardy expects agentic AI to reshape IT, supply chains, and customer service. In IT, it can rebalance workloads and automate predictive maintenance. In cybersecurity, it isolates compromised systems and deploys immutable backups in seconds. Early deployments show how businesses can strengthen resilience while reducing downtime. The workforce impact, however, will be uneven. The World Economic Forum predicts 11 million new jobs in Southeast Asia by 2030, but nine million could be displaced, affecting women and Gen Z disproportionately.
Balancing Autonomy with Oversight
Capgemini’s findings echo Hardy’s view: autonomy without accountability is risky. Oversight ensures AI agents act ethically and within organisational goals. Leaders must reskill teams, establish governance frameworks, and prepare infrastructure to fully capture value. With Microsoft investing $1.7 billion in Indonesia and launching training in Malaysia, capacity building is accelerating. Still, Southeast Asia faces a race between innovation and readiness.
Looking Ahead
The promise of agentic AI lies in its ability to move from insight to autonomous action. IDC projects that AI could add US$120 billion to ASEAN-6 GDP by 2027. Yet success will depend on how enterprises balance speed with responsibility. The technology can reshape workflows, lower costs, and enhance resilience. But without governance and human oversight, progress could stall. For Southeast Asia, the challenge is no longer if agentic AI will take hold, but how quickly and responsibly it will transform the region’s economy.
