APAC's AI Reality Check

For a growing number of APAC IT leaders, the conversation about artificial intelligence has moved past experimentation into a more uncomfortable territory. According to Confluent's Data Streaming Report, 75% are now deploying or piloting agentic AI systems, yet 78% say insufficient real-time data infrastructure is slowing AI growth. The gap between ambition and execution has become the defining tension of enterprise AI across Asia Pacific, and a new chorus of regional industry leaders is warning that adoption is outrunning the governance, trust and operational foundations required to make AI work at scale.

The numbers tell a stark story. Kyndryl's 2026 People Readiness Report found that while 57% of organizations say AI is now embedded in core business processes or deployed broadly across the enterprise, only 32% have achieved at least one of their top two AI goals. Adoption is moving faster than execution, and that gap becomes more urgent as AI agents move from supporting tasks to acting across IT, finance, operations, supply chains and customer service.

Adoption Accelerates, Trust Lags

The pace of daily AI use across the region is striking. Globally, around half of workers now use AI multiple times a week, while one in five use it almost daily. In Singapore, nearly one in four workers use AI almost every day, and more than half engage with it multiple times a week. Yet the same organizations racing to embed AI are expressing deep reservations about letting it operate without supervision. Kyndryl's research found that 81% of organizations expect AI agents to make impactful decisions within the next year, yet only 25% completely trust AI systems operating without human oversight.

That trust deficit is shaping how regional leaders are approaching deployment. Gartner predicts that by 2028, over half of enterprises will move beyond assistive AI in favour of platforms that deliver workflow outcomes. Across Asia Pacific and Japan, that shift is already visible in industries as varied as automotive manufacturing in Japan, semiconductor production in Malaysia and food traceability in Australia, where AI is being embedded into operational workflows rather than treated as a standalone productivity tool.

Governance, Security and the Shadow AI Problem

The governance gap is particularly acute in software supply chains. JFrog's 2026 Software Supply Chain Security State of the Union report found that 83% of organizations in APAC have checks for AI inputs and outputs, twelve percentage points above the global average. Yet only 55% use automated Shadow AI detection. Across the region, 68% of organizations surveyed self-host AI models. In Singapore, 18% have Shadow AI policies without enforcement, while in India, 34% accept AI-suggested fixes after only a cursory review.

That gap between policy and enforcement is where risk accumulates. AI agents can generate code, identify vulnerabilities, recommend fixes, and introduce packages, models and tools into the development process. Without a trusted system of record for every software and AI asset, including packages, models, agent skills and MCP servers, governance cannot keep pace with agentic development. The same trust deficit is visible in customer engagement. Braze's 2026 Customer Engagement Review found that 93% of marketers say AI helps them understand customers better, but only 53% of customers feel brands are accurately predicting their wants and needs. In APAC, where consumers are digitally sophisticated and quick to disengage from experiences that feel intrusive or impersonal, that gap is particularly difficult to ignore.

Voices from the Ground

Regional executives are increasingly framing AI as an operating model change rather than a technology rollout.

For the past two years, enterprises have treated AI as a race to adopt. The next phase will be a race to control.

That sentiment is echoed across industries. At ADP, leaders are reframing the conversation around what they call "The Great Job Unbundling," a new approach to work design that involves identifying which tasks are best suited for AI and which require uniquely human capabilities.

The true value of artificial intelligence lies not in the technology itself, but in its ability to empower people to focus on more meaningful, strategic and high-value work.

In critical systems, the bar is even higher. AI for aviation, defense, public safety and critical infrastructure must work within systems built around resilience, validation, redundancy, traceability and human oversight. Its outputs need to be explainable, verifiable and trusted in operational conditions. In Singapore, this approach is reflected through cortAIx SG, Thales' AI accelerator for critical systems in Asia.

Strategic Outlook: Building the Foundations for Trusted AI

The playbook emerging from regional leaders is consistent. Organizations pulling ahead will be those that treat AI as an operating model change, modernizing the infrastructure beneath it, redesigning work around people and agents, and orchestrating AI with trust, control and accountability. That means defining clear roles, permissions, escalation paths and auditability before agents are deployed at scale.

For manufacturers, the path forward runs through trusted industrial data and transparent AI governance. Singapore's Budget 2026 announced the AI Mission for Advanced Manufacturing, aiming to leverage AI to establish best-in-class manufacturing facilities, create first-in-world solutions and drive broad-based deployment across the sector. Realizing that vision will require manufacturers to move beyond isolated AI deployments towards scaling AI across engineering and manufacturing operations, embedding trusted AI into the workflows that design, simulate, build and operate products.

For software teams, the next phase of AI maturity will be measured by automated policy enforcement, end-to-end asset traceability, and the ability to demonstrate that AI-assisted development improves outcomes without increasing security risk, operational cost or audit exposure.

AI deserves appreciation. But in the enterprise, it also requires evidence.

The infrastructure layer is equally critical. As AI moves from training to production, enterprises are looking beyond where their data resides to how the intelligence and institutional knowledge generated through AI is governed, secured and retained. Singapore's Digital Infrastructure Bill reflects this broader shift, recognizing the strategic role resilient digital infrastructure plays in supporting the next generation of AI applications and long-term economic resilience. The organizations that will realize the greatest value from AI are not necessarily those that deploy the most models. They are the ones that invest in the data foundations, governance frameworks, human oversight and operational redesign required to make AI work reliably, explainably and securely across the systems people depend on.