Governments urged to streamline AI approvals to unlock public value
Public sector organisations are facing growing pressure to move beyond high-level principles in the risk frameworks that guide AI adoption, and to simplify the complex, duplicative approval processes.
A recent report released by Boston Consulting Group (BCG), Salesforce, and the Centre for Public Impact highlights a major opportunity for governments to accelerate AI delivery while continuing to meet high expectations for trust and safety. Though many governments have already established strong foundations with ethical guidelines, the report indicates that the next phase of maturity requires streamlining day-to-day operational workflows.
The report suggests that governments should ensure assurance frameworks do not unnecessarily slow progress, as failing to keep pace with societal adoption risks eroding public trust. A phased, lifecycle-based approach can help the public sector boost service delivery improvements while carefully managing risk.
The capabilities of AI models are moving faster than frameworks, guidance, and operation processes can keep up with. Addressing operational bottlenecks comes at a critical time, as global data from BCG shows that weekly AI usage among citizens has risen by more than 25% since 2024. This growing familiarity drives demand for public sector organisations to incorporate AI into their operations.
“Applying the principles and frameworks in practise has proven more challenging than necessary for real-world practitioners in the public service” said Miguel Carrasco, managing director and senior partner at Boston Consulting Group.
“AI projects are delayed or stalled, because frameworks are out-of-date, unclear, duplicative or the process is too hard to navigate. AI could generate significant public value, but without a simpler, more practical approach, many of those benefits risk being left on the table.”
Smarter governance for the agentic era
The report introduces a new operating model that separates organisational roles and clarifies how guardrails are enforced. This structure enables public sector entities to clear operational gridlocks faster and achieve measurable progress. Moving beyond traditional automation, these smarter controls are designed to support emerging technologies without trapping valuable tools in endless pilot stages.
To make assurance more effective and responsive, the report outlines nine practical suggestions. These recommendations focus on simplifying risk triage, clarifying roles for clearer accountability, and fast-tracking low-risk productivity tools within updated frameworks.
The report also emphasises adopting a staged lifecycle approach to reduce duplication through reusable artifacts, building baseline AI literacy across all staff, and shifting away from passive compliance forms toward active lifecycle monitoring.
“The question is no longer whether to govern AI responsibly, but whether our assurance processes are actually helping the adoption of technology that could transform public services,” said report co-author Gisele Kapterian, public sector strategy lead at Salesforce.
From policy frameworks to practical execution
International examples illustrate how governments are successfully turning abstract principles into streamlined pathways that eliminate repetitive tasks. For instance, in New South Wales, which the report singles out for leading governance maturity, officials successfully compressed a specialist-heavy review process that previously took roughly 40 hours down to an intuitive 15-minute inherent-risk triage.
Globally, other nations demonstrate similar progress in eliminating unnecessary administrative hurdles. Japan, for example, has replaced rigid preapproval mandates with a flexible, tiered report-and-review pathway tailored to project risk levels.
Meanwhile, the United Kingdom uses a mandatory central transparency standard to clear a predictable path for AI usage across public departments. By deploying these practical methods, public sector organisations can safely scale technology to unlock substantial value, including potential productivity gains estimated at $1.75 trillion annually on a global scale.
“The aim is to make AI governance better and more efficient, so that more AI projects make it from proof-of-concept, to pilot, to production more quickly,” concluded Carrasco.
“If we do that, then we will all benefit from the enormous innovation, productivity and value that AI – done well and responsibly – can bring to policy, regulation and service delivery.”
