KPMG’s new report paints a picture of governments worldwide moving fast toward digital transformation. Almost half—48%—of public sector organizations plan to launch full-scale AI projects in the next year. The excitement’s obvious, but there’s a big catch: a “capability gap” is slowing things down.
Let’s look at the numbers. Agencies are ramping up funding, with 59% boosting their budgets for data and analytics (many by over 10%). Leaders expect that managing AI agents will become a must-have skill for government workers soon—86% say so. Cybersecurity’s also top of mind. Sixty-four percent plan to put more money into cyber defense as AI rolls out.
That said, actually getting AI up and running isn’t simple. 43% percent admit to hitting serious roadblocks.
The main sticking points?
• Data Maturity & Quality: Many agencies lack the high-quality, structured data necessary to train and run reliable AI systems.
• Legacy Infrastructure: Fragmented enterprise systems and outdated cloud foundations are struggling to support the high compute demands of modern AI.
• Talent Constraints: There is a severe shortage of personnel who understand both public sector policy and the technical nuances of AI orchestration.
• Organizational Silos: AI requires cross-departmental integration, yet most governments still operate in isolated silos, leading to inconsistent governance and “fragmented” use cases.
The challenges go beyond technology. Government leaders worry about ethics and regulation—31% cite unclear rules as a major hurdle. Nearly half are most concerned about privacy and cyber threats. And without a strong, unified approach to oversight, these new AI systems risk missing the mark on fairness and transparency.
KPMG’s advice? Stop patching up old systems and start thinking about full, mission-driven change.
1. Strengthening Core Foundations: Prioritizing cloud platforms and enterprise data management.
2. Human-Centric Design: Ensuring AI tools are built to improve citizen services rather than just automate tasks.
3. Orchestration: Moving away from isolated experiments toward a coordinated, enterprise-wide strategy where AI agents are integrated into daily workflows.
Governments are ready to act, but now comes the real test. Over the next year, they’ll have to prove they can turn their big investments and ambitions into actual results.