AI for Transforming Public Service

Samisa Abeysinghe
Sep 30
3 min read

Updated: Oct 2

Generative AI (Gen AI) is reshaping conversations in every sector—but in government, its potential is uniquely profound. The public sector sits at the intersection of policy, service delivery, and citizen trust. Done right, Gen AI can make governments more responsive, empathetic, and efficient. Done wrong, it can erode confidence and reinforce inequities.

Over the past year, I have had the privilege of leading workshops across ministries and agencies, training nearly 6,000 public sector executives. These sessions were not just about technology—they were about reframing how leaders imagine the future of public service.

**Public sector executives empowered with AI**

From AI to Generative AI

To ground discussions, I often start with fundamentals. Artificial Intelligence (AI) is a broad field covering problem-solving, prediction, and decision-making. Generative AI, however, is different: it creates. It learns from patterns in data to produce new text, images, code, or even music.

This distinction matters for governments. Traditional AI can optimize tax collection or traffic management. Generative AI can draft policy briefs, translate legal texts, or produce multilingual citizen communications—tasks that bridge governance and creativity.

LLMs, GPTs, and the Natural Language Revolution

Much of Gen AI’s power today comes from Large Language Models (LLMs) and GPTs. These models are trained on vast text datasets, allowing them to understand and generate human-like language.

For government officers, this means a shift: no longer do they need to rely on specialized coding skills. Natural language itself becomes the interface. Officers can now ask questions in plain English (or Sinhala, Tamil, Hindi, etc.), and the system responds meaningfully. This “natural language revolution” lowers barriers, democratizes access, and makes advanced technology usable by non-technical staff.

Samisa Abeysinghe Presenting at AI Worksop at Temple Trees

The Art of Prompting and Articulation

If natural language is the bridge, prompting is the skill that determines how effectively you cross it. In my workshops, we emphasize the art of articulation: how literacy, structured thinking, and curiosity shape the quality of AI interaction.

Government officers who can write clear, logical prompts unlock better AI outputs. For example, using chain-of-thought prompting helps AI break down complex procurement processes step by step. Similarly, tailored prompts can train junior officers in budgeting, project development, or audit standards.

This is not just technical—it’s about cultivating a new kind of literacy in public service.

Applications in Finance, Audit, and Policy

One of the most exciting parts of our training has been showing executives domain-specific prompts for critical government functions:

Budget Planning & New Project Development: AI can help analyze allocations, test scenarios, and support strategic planning.
Monitoring & Evaluation: AI prompts guide officers in tracking project outcomes and measuring impact.
Audit Procedures: AI can explain methodologies, compliance requirements, and reporting standards—helping officers master government audits faster.

By embedding Gen AI into financial governance, we are not just automating tasks—we are accelerating institutional learning.

Risks, Bias, and Responsible Use

Every workshop also spends significant time on risks. Gen AI can produce misinformation, mishandle sensitive data, or reinforce societal biases. I often use the penguin analogy from our training: when a system sees only penguins as birds, it might falsely assume all birds can’t fly.

This illustrates how biased or incomplete training data can lead to flawed outputs. In government, such errors are not academic—they affect people’s lives. That is why principles of transparency, accountability, and fairness must be embedded into every deployment.

Citizen Centricity and Inclusion Though AI Enabled Public Sector Transofrmation

High-Value Use Cases for Citizen Services

To anchor vision in practice, we identify the top five high-value Gen AI use cases for government:

Automated Customer Service – 24/7 multilingual citizen support.
Fraud Detection – AI-powered analysis to safeguard public funds.
Resource Allocation – Optimizing how scarce resources are distributed.
Policy Analysis – Turning large datasets into insights for better decision-making.
Digital Document Processing – Automating paperwork, reducing delays.

Each of these use cases, when implemented responsibly, can cut waiting times, improve trust, and expand access to government services.

Moving Forward

The transformation of public service through AI is not just about efficiency—it is about equity, accessibility, and trust. Governments should start with low-risk pilots, invest in workforce upskilling, adopt responsible AI frameworks, and deploy secure platforms designed for sensitive public data.

The ultimate goal is not merely digitization. It is to reimagine public service as human-centered, empathetic, and anticipatory—a system where AI helps officers and citizens alike navigate complexity with clarity.

Gen AI offers governments a mirror: it reflects back the values they choose to embed. Fairness, transparency, and inclusion are not optional—they are the very foundation of trust.