Co-Pilot Approach: AI In Public Sector Development

Last month, we delivered a comprehensive email platform for a Nigerian state government in 28 days. The same project would typically take us four to five months using traditional development approaches. The reason for this difference? Artificial intelligence.

For most of 2023, AI sat at the margins of our work at Nugi Technologies. We used it for basic coding assistance and documentation support, the way most development teams do. But by late last year, it became clear that treating AI as a peripheral tool was limiting our potential to serve the public sector clients who depend on us.

When you’re building systems that citizens interact with daily, such as email platforms, e-governance tools, and service delivery applications, efficiency gains translate directly into better public services. A month saved in development means citizens get access to improved government services a month sooner.

Rethinking The Development Cycle

Our biggest AI-driven transformation has been in our approach to the development cycle itself, starting with our requirement gathering process. What used to take our product teams up to two weeks (research, documentation, market validation) now happens in a few days. We can feed raw ideas into AI tools and get structured requirements instantly, solving 50 per cent to 60 per cent of the initial problem upfront.

Instead of spending hours on manual research or crafting basic documentation, our team now focuses on understanding complex government workflows and designing solutions that truly serve citizen needs.

In the design phase, AI helps us generate detailed user stories and convert design mockups into functional code. We have tools that can structure database architectures and generate APIs, making handoffs between teams smoother. But here’s the lesson: AI is only as effective as the person guiding it. Developers who frame problems in clear sequences (step one ⇾ step two ⇾ step three, etc.) consistently get better results than those who don’t. The better the prompt, the better the outcome.

The Human Factor Remains Critical

Some engineers initially worried that leaning on AI would make them less of a developer. I understand that concern, but I see it differently. When AI takes on routine code generation, developers can spend their energy on what matters most: understanding workflows, making strategic architectural decisions, and designing for the realities of government systems.

This shift required training and, more importantly, a cultural reset in how we think about efficiency. Our developers now dedicate more time to evaluating internal systems, studying citizen needs, and making creative decisions about user experience. For example, when AI produces multiple solutions to a single problem, the real value comes from the developer’s judgment in selecting the one with the best performance or fit within government constraints.

In practice, this has made our teams faster, and in many cases better – not because AI does the work for us, but because it lets our developers become strategists and domain experts instead of just code writers. They’re solving the puzzle of how to make government services more intuitive for citizens who don’t want to fill out lengthy forms or navigate complex processes.

Automated testing now catches vulnerabilities and performance issues that human reviewers might miss due to fatigue or time constraints. But we still maintain strong human oversight. I chair an internal committee with senior engineers that reviews all AI-generated code, benchmarks it against both our internal standards and international coding practices, and ensures consistency across projects.

Working with public sector clients presents unique considerations that consumer-facing applications don’t require. We handle sensitive citizen data, and our systems affect how people access essential services. A failure isn’t just an inconvenience, but can impact daily life. Standardised boilerplates and templates guide our AI prompts, ensuring outputs align with both company expectations and global security requirements.

Our AI usage must comply with strict privacy protection policies such as the Nigerian Data Protection Regulation (NDPR) and the General Data Protection Regulation (GDPR). For non-technical end users, however, the backend complexity doesn’t matter. What matters is the experience: how intuitive is the interface? How quickly can they complete a task? Nigerians, for instance, dislike lengthy forms, so we use AI to streamline workflows, often cutting four steps down to two. The quality assurance teams then ensure the system performs reliably before deployment.

While these lessons shape how we build, I believe the same principles apply to government. Responsible adoption, strong human oversight, and citizen-focused design are just as critical when public institutions deploy AI.

It represents progress toward creating frameworks for ethical adoption while still accelerating innovation. But more is needed: expanded training programmes that bring together government, academia, and industry stakeholders. When universities offer machine learning and AI programmes, graduates are better prepared to tackle real-world challenges with these tools from the start of their careers.

For start-ups building AI-powered solutions for public service, my advice is: embrace AI capabilities, but never forget the human factor. AI-powered solutions are increasingly in demand because they reduce costs and shorten delivery timelines while often improving quality outcomes. The companies that will succeed are those that view AI as an amplifier of human capabilities rather than a replacement for human judgment.

AI has come to stay, that much is certain. But the key to harnessing its potential, whether you’re building internal tools or citizen-facing government services, is maintaining the right balance. AI handles complexity, speed, and automation. Humans provide insight, creativity, and ethics. When we strike that balance correctly, we don’t just build better products. We build more accountable institutions.

That’s the journey we’re on, and it’s one I believe will reshape how public services work across Africa.

 This article was written by Godswill Adie, CTO, Nugi Technologies.