Agentic AI is being positioned as the next major leap in artificial intelligence, with Capgemini Research Institute estimating it could unlock US$450 billion in economic value by 2028. Yet reality paints a different picture: adoption remains in its infancy. Only 2% of organizations have scaled agentic AI, and trust in AI agents is already showing signs of erosion.
This paradox — sky-high potential vs. slow deployment — sits at the heart of the debate. For Southeast Asia, the implications are particularly pressing. The region’s enterprises are eager to harness AI for efficiency and growth, but they must also contend with governance, infrastructure readiness, and workforce disruption.
Defining Agentic AI
To understand the hype, it helps to distinguish agentic AI from generative AI:
- Generative AI: Produces content when prompted (text, images, code). It is reactive.
- Agentic AI: Acts like a team of autonomous experts — deciding, learning, refining strategies, and taking actions in dynamic environments.
As Jason Hardy, CTO for AI at Hitachi Vantara, explains:
“Agentic AI may use GenAI inside it, but its role is to pursue objectives, coordinate tasks, and respond in real time.”
This distinction between outputs (GenAI) and outcomes (Agentic AI) is what makes it transformative for enterprises.
Early Steps, Limited Progress
Capgemini’s April 2025 survey of 1,500 executives across 14 countries (including Singapore) reveals a cautious start:
- 25% of organizations have launched pilot programs.
- 14% have moved to implementation.
- The majority remain stuck in the planning stage.
Executives are not rejecting the technology, but rather grappling with readiness. The gap between intent and execution is now one of the biggest barriers to unlocking economic value.
Emerging Use Cases: From Theory to Practice
While large-scale deployment is limited, real-world applications are already surfacing:
- Personal shopping assistants capable of searching, comparing, and preparing carts using voice or text.
- IT operations assistants that optimize storage, automate compliance, and predict system failures.
- Cybersecurity responders that detect anomalies and isolate systems before breaches spread.
These use cases reveal a shift from insight generation to autonomous action.
Why Enterprises Are Paying Attention
Complexity and Scale
Modern enterprises face overwhelming data complexity, risk, and scale. Agentic AI helps by:
- Optimizing resources in real time.
- Automating governance and compliance.
- Anticipating and responding to threats.
Value Beyond Efficiency
Agentic AI doesn’t just save time — it prevents failures, reduces downtime, and strengthens resilience. Early adopters in IT report measurable improvements in reliability, cost control, and performance.
IT Operations: The Practical Starting Point
According to Hardy, IT is the most mature entry point for agentic AI:
- Automated data classification frees hours of manual labor.
- Proactive storage optimization reduces waste.
- Compliance reporting becomes streamlined and continuous.
- Predictive maintenance cuts downtime.
- Real-time cybersecurity response limits damage.
By proving its worth in IT, agentic AI sets the stage for wider adoption across supply chain management, customer service, and risk governance.
Southeast Asia’s Readiness: Promise and Pitfalls
Data Foundations
Hardy stresses: “Agentic AI delivers value only when enterprise data is properly classified, secured, and governed.”
Without strong data pipelines, adoption risks failure.
Infrastructure Gaps
Agentic AI requires:
- Multi-agent orchestration systems.
- Persistent memory.
- Dynamic resource allocation.
Many Southeast Asian firms are still modernizing IT foundations, creating a readiness gap.
Regional Opportunities
Despite these challenges, Southeast Asia is primed for growth:
- Microsoft’s $1.7 billion investment in Indonesia reflects growing confidence.
- Training programs in Malaysia and across ASEAN aim to build local AI skills.
- Governments are pushing national AI strategies, signaling public-private momentum.
Workforce Impacts: Creation, Displacement, Reskilling
The World Economic Forum forecasts:
- 11 million jobs created in Southeast Asia by 2030.
- 9 million displaced.
Sectors like IT, retail, and manufacturing will be reshaped.
- Women and Gen Z are expected to face the sharpest disruption, with up to 76% of younger workers in vulnerable roles.
This underscores the urgency of inclusive reskilling. HR teams will need to prepare for:
- New roles in AI governance, auditing, and orchestration.
- Less administrative execution, more strategic oversight.
- Rebalancing workforce diversity to avoid widening inequality.
Oversight and Trust: The Balancing Act
Capgemini’s findings are unambiguous:
- 73% of executives believe human involvement in AI workflows outweighs its costs.
- 90% describe oversight as positive or neutral.
The message: agentic AI cannot run on autopilot.
Governance is not optional. Oversight ensures:
- Systems operate within ethical and organizational limits.
- AI actions align with business strategy.
- Accountability remains with people, not machines.
The Road Ahead for Southeast Asia
Hardy predicts the first visible transformation will be in IT operations, but the larger surprise will come at the economic level.
- IDC estimates AI and GenAI could add $120 billion to ASEAN-6 GDP by 2027.
- In Indonesia, 57% of jobs are expected to be disrupted or augmented.
- AI will reshape business models, risk management, and value creation.
This suggests agentic AI will move faster and have more material impact than many Southeast Asian leaders anticipate.
Conclusion: Promise with Guardrails
Agentic AI holds extraordinary promise — from preventing IT outages to reshaping customer experiences and unlocking GDP growth.
But Southeast Asia’s path to adoption hinges on three imperatives:
- Build the foundation – Secure, classified, governed data and robust infrastructure.
- Reskill inclusively – Invest in workforce adaptation to ensure no demographic is left behind.
- Balance autonomy with oversight – Treat human oversight as a value multiplier, not a cost.
The technology will not replace enterprise decision-makers — it will redefine their role. For Southeast Asia, the challenge is not whether to adopt agentic AI, but how quickly and responsibly it can be deployed.
