LAHORE: Public sector productivity has long posed challenges for governments worldwide, but in Pakistan, service delivery is marked by extreme inefficiency and high costs. In contrast, developed economies, while incurring higher public sector costs, generally deliver services more efficiently.
To boost productivity through artificial intelligence (AI), UK government leaders have emphasised the need to tackle outdated technology, limited access to quality data and a shortfall in digital skills. Similarly, as the US seeks to redefine the role, scale and scope of its federal government, public sector productivity has returned to the spotlight. With US national debt swelling to $36 trillion, the renewed focus is timely.
Measuring the productivity potential of the public sector is far from straightforward. The absence of clearly defined economic outputs makes it difficult to calculate Total Factor Productivity (TFP), hindering meaningful comparisons with other sectors of the economy.
Both challenges -- the difficulty of measuring public sector productivity and the practical barriers to improving it via AI -- are acutely relevant to Pakistan, where structural and institutional inefficiencies are even more pronounced.
As in the UK, Pakistan’s public sector lacks clearly measurable economic outputs, making it difficult to assess performance using TFP or similar metrics. Many services -- such as education, policing, regulation and healthcare -- are assessed based on input measures (budgets, staffing levels) rather than outcomes (citizen satisfaction, case resolution times). Without a productivity baseline, measuring gains or justifying investment in technology remains nearly impossible.
The barriers to efficiency and AI adoption are arguably steeper in Pakistan than in the UK or US. Many government departments still rely on paper-based systems or outdated, standalone software, hampering integration and scalability for any digital or AI-driven solution.
A major challenge is the poor quality and accessibility of data. Even basic citizen databases -- covering land records, taxation and health -- are often fragmented, inaccurate or undigitised. This poses a significant obstacle for AI, which relies on clean, structured data.
Pakistan’s civil service also suffers from a pronounced digital and AI literacy gap. Most public officials lack the training needed to understand, procure or implement AI tools effectively. Unlike the private sector, the public sector struggles to attract AI talent due to rigid pay structures and limited career progression.
As in the UK, critical data in Pakistan is often trapped in obsolete or siloed systems -- if digitised at all. Integration between federal, provincial and local systems is weak, limiting opportunities for intelligent automation.
To enhance public sector productivity, experts say Pakistan must develop a clear strategy, setting measurable, output-based KPIs for departments to begin identifying productivity gaps and potential improvements. Foundational records and systems -- such as those held by NADRA, land revenue, and health authorities -- must be digitised in interoperable formats that support data analytics.
Investment in basic digital infrastructure and cloud-based platforms enabling real-time data sharing across departments is essential. The government should also introduce specialised AI and data-skilling programmes for civil servants, and consider special pay scales or public-private partnerships to attract AI specialists. Pilot projects in high-impact areas -- such as tax collection, health diagnostics and citizen grievance redressal -- could serve as a proof of concept for AI’s productivity gains.