Governments are struggling to adopt AI across the Asia Pacific region. It's not because the technology isn't ready, but because their operating models aren't.
The Australian state of Victoria has declared an intent to be its country’s leader in AI. Singapore's Smart Nation program has made AI-powered public service delivery a national priority, positioning the city-state as one of the region's most advanced digital governments.
Across the region, ministers and secretaries are setting bold government AI targets. Yet, when you sit down with Asia Pacific technology leaders tasked with delivering on those targets, a more nuanced story emerges.
At the March ServiceNow AI Summit in Melbourne, I heard the same conversation repeatedly across government agencies. Proofs of concept (PoCs) demonstrated meaningful value but stalled when rolling out to scale.
Australia’s Digital Transformation Agency (DTA) has a name for it: PoC hell.1 DTA Deputy CEO Lucy Poole flagged it publicly in 2025 as an emerging risk across the Australian Public Service. She described agencies that cycle through PoC after PoC without the governance mechanisms to move beyond them and realize real benefits.
What’s wrong is that most government agencies don't have a single, unified view of their data. They operate across dozens of overlapping systems, many of which can't connect with each other and weren't built with AI in mind. Running a small pilot on top of that architecture might be straightforward, but scaling it isn't.
Governance can compound the problem. In most agencies, governance is applied as a set of approvals that must be cleared before anything moves forward. That makes it a brake on speed.
The agencies that can scale AI treat governance differently. They build it into the platform itself, with audit trails, permissions, and escalation boundaries embedded in the workflow rather than layered on top of it. Governance in this framework makes faster AI adoption safe, not something that slows it down.
A 2026 Center for Data Innovation study of more than 3,000 public servants across 10 countries found that while more than 70% now use AI, only 18% say their governments are using it effectively.
A government study by ServiceNow and ThoughtLab across 1,248 global government leaders identified a cohort we call Pacesetters. These are the top 22% of government organizations by AI and digital transformation maturity.
These agencies are already past the PoC stage. They're reporting:
- 1.5 times faster returns on digital investments than peers
- 47% higher citizen self-service success
- 59% new cross-functional workflows through human and AI collaboration
What sets them apart isn't budget; it's approach. They prioritize platform consolidation over point solutions. They measure outcomes from day 1. And they treat AI as an operating model transformation, not a technology project.
The Tony Blair Institute's analysis of 20,000 public sector tasks found that more than 40% of administrative work is at least partly automatable. For most government agencies, that means IT service management, HR operations, and finance, which are among the highest-volume, most process-intensive back-office functions across public sector organizations globally.
That potential doesn't unlock itself. It unlocks when the data, workflows, and governance are connected on a single platform.
The agencies making the most progress aren't attempting whole government transformation in a single program. They're moving through three stages:
- Building the platform foundation and enabling self-service
- Introducing generative AI to co-handle and proactively resolve cases
- Realizing the benefits of agentic AI where autonomous workflows handle structured process work end to end
Each stage generates measurable value that funds the next stage. Each also helps build governance confidence, allowing ministers, auditors, and the public to see the benefits and support continued AI investment.
Across Australia, New Zealand, and Singapore, agencies are beginning that journey now. And they're doing more than just improving efficiency metrics; they're building the operational foundation for governments that can deliver the expectations of their citizens.
Find out more about how leading public agencies are scaling PoCs into public sector AI transformation.
1 Justin Hendry, InnovationAus.com, DTA flags AI 'PoC hell' as emerging APS risk, 15 Aug. 2025
