Across Nigeria, there is growing recognition that artificial intelligence will influence how governments deliver services, how universities support learning, and how public institutions make decisions.
Most leaders are no longer asking, “What is AI?”
They’re asking something far more practical: “Where do we even begin?”
For many ministries, agencies, public universities, and state institutions, that question is difficult to answer.
There is no shortage of webinars, conferences, or consultants explaining why AI matters. Yet there is far less practical guidance on how a public institution with limited budgets, ageing systems, competing priorities, and overworked staff should take its first meaningful step.
In Part 3 of this series, we explored the divide between organisations that are moving forward with AI and those that are still waiting to start. In Part 4, we looked at what that divide means specifically for Africa.
This article is for the Nigerian public institution that has heard enough of the argument and is ready to do something about it.
That gap between awareness and action is where many institutions find themselves today. The good news is that getting started does not require a multimillion-naira project or an organisation-wide transformation. It only requires making the right first decisions.
Start with the problem, not the technology
One of the most common mistakes organisations make is beginning with a tool. Someone attends an AI conference where a vendor gives an impressive demonstration.
Leadership becomes excited and wants to implement immediately. The first question becomes: “Which AI platform should we buy?”
That is the wrong place to start. Instead, ask: What is the single biggest operational problem that repeatedly wastes time, creates unnecessary errors, or delays service delivery?
It might be processing student records. Responding to citizen enquiries. Managing procurement documents. Tracking expenditure. Preparing reports that staff spend days compiling manually.
Choose one problem. Not five. Not an entire digital transformation programme. One problem that matters enough to solve. Technology should follow the problem, not define it.
Get your data in order first
AI depends on information. Unfortunately, information inside many public institutions exists in several different places at once.
Some records live in spreadsheets. Others are stored in filing cabinets. Important conversations happen in email chains or WhatsApp groups. Sometimes the most valuable knowledge exists only in the minds of experienced staff.
That knowledge needs to be documented before it disappears. Simple steps like creating standard operating procedures, recording how decisions are made, and conducting structured handover sessions when staff transition out of roles can preserve institutional knowledge that would otherwise be lost.
An AI system cannot learn from what was never written down. Before introducing AI, institutions should ask a simple question:
Could someone new understand our information without relying on institutional memory?
If the answer is no, AI is unlikely to perform well. Improving data quality does not sound exciting, but it is one of the highest-value investments an institution can make.
Organise records. Standardise formats. Remove duplicates. Identify missing information. Document processes.
This work lays the foundation not only for AI but for better decision-making across the organisation.
Start small, and prove that it works
Public institutions often feel pressure to launch ambitious digital transformation programmes.
In reality, large-scale projects frequently struggle because they attempt to change too much at once.
A better approach is to begin with a pilot.
Choose one department. One process. One measurable objective. Run the project for a defined period. Measure the results carefully.
- Did processing times improve?
- Were fewer errors made?
- Did staff spend less time on repetitive work?
- Was the service better for citizens or students?
Evidence matters. A successful pilot creates credibility. Consider a state ministry that automates the processing of just one category of application, pension requests, for example.
If processing time drops from three weeks to three days, that result alone justifies the next conversation about expanding the pilot.
It demonstrates that AI is more than a buzzword and provides leadership with real data to support future investment. Small successes often achieve far more than ambitious announcements.
Practical first use cases
Not every AI project needs to be groundbreaking. Some of the most valuable applications solve ordinary administrative problems.
Document processing
Many public institutions process large numbers of forms, applications, letters, and reports every day. AI can help categorise documents, extract key information, and reduce manual administrative work.
Responding to common enquiries
Universities receive repeated questions about admissions, registration, examinations, and fees.
Government agencies answer similar questions about services, requirements, and procedures.
An AI assistant can handle straightforward enquiries while allowing staff to focus on more complex cases.
Critically, these systems must always offer a clear and easy route to a human being, not as a last resort, but as a standard feature. In public service, citizen trust depends on knowing that a real person is accessible when it matters.
An AI that feels like a wall rather than a door will do more harm than good.
Meeting summaries and report drafting
Public institutions produce countless meeting minutes and reports. AI can generate first drafts and summaries, helping staff spend less time formatting documents and more time reviewing decisions and actions.
Human oversight remains essential, but drafting becomes significantly faster.
Service delivery analysis
Every institution generates data, waiting times, application volumes, complaint patterns, response times.
AI can help identify trends that might otherwise go unnoticed, giving leaders better evidence for operational improvements.
Budget monitoring
Finance teams spend considerable time reviewing expenditure and identifying unusual spending patterns.
AI can support this work by highlighting anomalies or trends for further investigation. It does not replace financial oversight, it strengthens it.
Bring your people along
Technology without people is simply software that nobody uses. One of the biggest barriers to AI adoption is fear.
Some employees worry that automation means redundancy. Others fear making mistakes with unfamiliar technology.
Some simply assume AI is another short-lived initiative that will disappear after a few months. These concerns deserve honest conversations rather than reassurance without evidence.
Leadership should explain what AI is intended to achieve. Staff should understand that the goal is to reduce repetitive administrative work, not eliminate the expertise and judgement that public service depends on.
Training also matters. Employees do not need to become AI engineers. They need enough confidence to understand what the technology can do, recognise its limitations, and use it responsibly in their daily work.
When people understand the purpose behind change, they are far more likely to support it.
Working within procurement realities
Public procurement is often seen as an obstacle to innovation. While procurement rules are essential for accountability and transparency, they can make it difficult to adopt rapidly evolving technologies.
That does not mean institutions must wait indefinitely.
Many practical options already exist. Tools like Hugging Face, which hosts thousands of openly available AI models, or Llama, Meta’s open-source large language model, can be deployed without expensive licensing fees.
For document processing and automation, tools like Apache OpenNLP or spaCy offer practical starting points.
None of these require significant upfront investment, they require people with the skills to configure and maintain them, which is a more solvable problem than budget.
Universities can also partner with researchers and academic departments to test new ideas. Pilot projects can demonstrate value before larger investments are considered.
Collaborating with Nigerian AI startups can also be beneficial. Local companies often understand the country’s regulatory environment, infrastructure challenges, and user needs better than organisations designing products primarily for international markets.
Innovation does not always require the largest budget. Sometimes it requires the right partnerships.
What success looks like after one year
Success is not an institution that suddenly becomes “AI-powered.”
It is not a headline announcing a digital revolution.
It is something much more practical.
One process that is noticeably better than it was twelve months earlier. One department with evidence that change improved efficiency or service delivery.
One team that understands how AI fits into its work, and where it doesn’t.
Those achievements may seem modest.
In reality, they create something invaluable: confidence.
Confidence to expand thoughtfully rather than rush blindly. Confidence to make future decisions based on evidence instead of hype. And confidence that innovation can be both responsible and practical.
Closing thought
The distance between where many Nigerian public institutions are today and where they could be is often smaller than it appears. The gap is not measured by the sophistication of the technology available.
It is measured by decisions. The decision to identify one genuine problem worth solving. The decision to organise data before chasing new tools. The decision to test a small idea, measure its impact, and learn from the results.
Meaningful transformation rarely begins with a grand announcement. It begins with one practical step that makes tomorrow’s work better than today’s.
That is how institutions build trust in new technology. And that is how the AI divide begins to close, not all at once, but one informed decision at a time.
