Eighty-two per cent of public-sector organisations are already using autonomous AI agents, and 71 per cent plan to expand deployments in 2026–27. IDC research, highlighted by ZDNet, says budget pressure, citizen demand for faster, personalised services and legal requirements around data sovereignty and transparency are driving the shift — and that public bodies may lead the next wave of operational automation.

Why the public sector is racing ahead

Governments face a different set of incentives to commercial firms, and those incentives are pushing them toward agentic AI sooner. IDC research, cited by ZDNet, finds many agencies treat agentic systems as a leadership priority rather than an experimental toy. Tight budgets mean public organisations are searching for tools that can deliver efficiency gains across departments without large increases in headcount.

But fiscal pressure is only part of the story. Citizen expectations are changing fast. People want quicker responses to queries, services tailored to their circumstances and simpler end-to-end journeys. Governments see AI agents — autonomous digital workers that can reason and take action — as a way to meet those demands at scale, providing context-aware interactions and proactive outreach.

And there are legal and strategic drivers unique to the public sector. Compliance obligations around data protection, algorithmic transparency and accountability make governments prioritise control and traceability. That has pushed them to adopt agentic architectures that can be audited and governed centrally, even if the systems themselves operate across many touchpoints.

Operational gains come from stitching together multi-step workflows. An agent that coordinates tasks across benefits, housing and employment teams can reduce hand-offs and delays. Agencies are beginning to see that kind of orchestration as low-hanging fruit for automation.

Mastheads

Where agents are already being used

IDC’s work highlights three focus areas for government application: operational orchestration, citizen service delivery and decision support for policy and planning. In practice, that looks like chat and voice agents fielding routine enquiries, back-office agents routing and prioritising cases, and planning agents simulating policy scenarios using synthetic data.

Those uses are complementary. Frontline agents take pressure off contact centres. Back-office agents speed up approvals and document handling. Planning agents give officials richer context when designing programmes, by modelling likely outcomes under different assumptions. Together, they build a more responsive public service.

Many deployments are still pilots. But the research shows a clear path from proof-of-concept to scaled operation: identify high-impact workflows, prepare the data architecture, ensure data quality and set up an operations and governance model for agents. Where agencies have followed those steps, they’ve been able to move faster.

Data foundations and governance remain the bottleneck

Scaling agentic AI depends on solid data work. Agencies must know where their data sits, how to link it across silos and how to keep it clean. IDC warns that without a strong data foundation, agents will be brittle and inconsistent. That matters more in government than in many private firms because public services must meet legal tests of fairness and transparency.

Governments are also wrestling with new staffing needs. The study notes gaps in cybersecurity skills and machine-learning operations among public-sector workforces. Agencies are hiring and retraining, but the pace of change means workforce disruption is a real issue. Agents can free staff from repetitive tasks, but they also require new roles — people who can tune, monitor and explain automated decisions.

Addressing these data, skills and governance gaps is essential before agencies can safely and reliably scale agentic AI across services and workflows.

Related Articles

IDC found that 82 per cent of government organisations have already adopted AI agents, and 71 per cent plan to increase their use in 2026–27.

This article was created with AI assistance.