Banks, telcos driving Nigeria’s data governance — Apampa

In this interview with JUSTICE OKAMGBA, Metaheuristic’s Chief Executive Officer and Founder, Toye Apampa, explains why Nigerian banks are more ready to adopt AI, while other sectors trail behind due to weak data governance and inadequate foundational systems

You monitor Nigeria’s digital governance progress. What key gaps are currently holding the country back?

I’ve spent the last fourteen years inside some of Britain’s most data-heavy organisations: Lloyds, British Airways, Marks & Spencer, WorldRemit, and Rank Group. In environments like those, data governance isn’t a nice-to-have. It is how you keep regulators at bay and the business running, whether you are deciding if someone qualifies for credit or ensuring a gambling platform isn’t serving a vulnerable customer.

Most of my conversations with senior executives across those organisations came back to the same question: can we trust the numbers we are seeing? Underneath that was almost always the same problem. User data lived in one system, transactional data in another, and marketing data in a third. When the time came to stitch it all together for a single view of the customer, the identifiers didn’t match. Sometimes the data wasn’t even there at the point of stitching.

Processing had to be rerun to unpick legacy joins that had been feeding incorrect insights into the boardroom for months. By the time you fix it, decisions have already been made. That is the reality of data governance at scale. It is not glamorous, but it is the foundation every downstream decision rests on.

When I look at Nigeria through that lens, the frameworks are more impressive than people give them credit for. The Nigeria Data Protection Act and the NITDA AI Code of Practice are serious, well-drafted documents. The UNDP Digital Development Compass scores Nigeria’s Data and Privacy framework at 4.17 out of 5. Metaheuristic, my data and AI consultancy (mheuristic.com), built the Nigeria Digital Governance Tracker on top of that Compass framework specifically because the underlying data were too important to leave in a global dashboard. Decision-makers here needed it surfaced in a form they could actually use.

The gap lies in the plumbing underneath the policy. Where the policy scores high, implementation capacity sits at 2.5. The laws exist, but the operational muscle to comply with them, the ability to actually execute these mandates across ministries and private enterprises, is the next frontier for the Nigerian digital economy.

How ready are African organisations for AI adoption beyond the hype?

Across the board, readiness is often lower than the noise suggests, and I say that as a technologist, not a critic. I have helped UK institutions prepare for AI. It is slow, unglamorous work: cataloguing data, documenting its origin, and defining what a customer means across departments that have used different names for twenty years. None of that makes the news, yet it is the entire difference between an AI model you can trust and one you cannot defend.

The risk for many emerging markets is that the conversation about AI has arrived before the conversation about data foundations has had a chance to finish. Boards are being asked to approve AI roadmaps before anyone in the business can confidently describe the data those roadmaps depend on. The organisations that will win this decade aren’t the ones deploying the flashiest models; they are the ones quietly doing the foundational work first.

What does AI readiness actually entail for a typical business?

Strip away the jargon. AI readiness comes down to three questions any business owner should be able to answer before spending a single naira on AI.

Do you know what data you have? You need an actual inventory of where it lives and who owns it.

Can you trust it? Is it clean, consistent, and documented, or is it a pile of spreadsheets that nobody has audited in years?

Can you legally use it? Under the Nigeria Data Protection Act, “it was in our system already” is not a legal basis for AI processing.

I have sat in London boardrooms and watched senior executives realise, mid-meeting, that the insight on the screen was built on data joined incorrectly for months. Different systems, different user IDs, and gaps that nobody had flagged because the organisation had grown faster than its governance. You cannot reliably train a model on data you cannot reliably stitch together.

That global struggle is exactly why I built EventParity (eventparity.com). The platform is designed to automate the boring but essential governance work of mapping data flows and tracking compliance—so that businesses can actually use AI without tripping over regulatory or ethical wires.

Nigeria is advancing digital transformation. Are governance structures keeping up?

No, and the gap is widening. Digital transformation is something you can buy; governance is something you must build through people, process, and practice.

When digital outpaces governance, you end up with shiny services sitting on shaky foundations. Nigeria has a chance to avoid the catch-up trap I watched the UK go through. By integrating governance tools from day one, the country can leapfrog the legacy mistakes of the West. If the gap widens, Nigeria builds digital infrastructure it cannot ultimately defend.

Based on your tracker, which sectors in Nigeria are set to lead in digital governance, and which are likely to fall behind?

The pattern usually follows the strength of the regulator. Banking, under the CBN, and telecommunications, under the NCC, are positioned to lead because they have been forced to build governance muscle for decades.

The lagging sectors are typically health and education. In these fields, data is frequently scattered across incompatible systems. When data doesn’t speak to other data, critical decisions in hospitals and schools are made on fragmented information.

The NHS in the UK is a cautionary tale. Different trusts run different software with limited integration, and the cost of fixing it has become so high that most departments put it off indefinitely. Patients suffer the consequences. Every new digital initiative inherits the legacy problem. Nigeria has the opportunity to avoid that trajectory, but only if data integration is built in from the start rather than bolted on a decade later.

How can regulators strike a balance between innovation and risk as AI advances?

The biggest lesson from working in highly regulated UK sectors is that prescriptive regulation ages badly. Regulators should tell organisations what outcome they must achieve, protected data and explainable decisions, and let them determine the how.

Three things are vital. First, supervised environments where organisations can pilot new AI systems under a regulator’s watch, with a clear path from successful tests to live operation. Second, capacity-building helps institutions learn the tools of compliance rather than just punishing them for falling short. Third, inter-agency coordination, ensuring that NITDA, the NDPC, and the NCC aren’t pulling businesses in different directions.

What are the most frequent data governance mistakes organisations make?

I’ve seen four mistakes repeated globally. The first is treating data governance as an IT problem. It is an accountability problem. If only your junior engineer can explain your data, your structure is broken.

The second is buying tools too early. Procuring a platform before you have defined your business terms just automates the confusion. I’ve watched organisations spend millions while two departments still couldn’t agree on what a customer was.

The third is defaulting to centralisation. Trying to pull all data into one giant lake is often a recipe for technical and political failure. The lake usually fills with duplicates rather than insights.

The fourth is governance as an afterthought. Trying to fix governance after an AI model is live is like trying to install a foundation after the house is already built.

If you had to choose a single policy reform to speed up Nigeria’s digital future, what would it be?

Make data maturity a mandatory condition of public-sector AI procurement.

Before any agency spends public funds on an AI system, it should be required to produce a standardised assessment of the data that system will rely on. Is it clean? Is it legal under the NDPA?

This creates an immediate economic incentive to invest in governance. It forces institutions to confront the state of their plumbing before they buy the taps.

If Nigeria gets this right, the picture in ten years is one where a state ministry can deploy an AI model with confidence, knowing the data underneath it is defensible and where a citizen can trust that automated decisions are traceable. If it gets it wrong, we end up with a decade of expensive pilots and a slow erosion of public trust. The difference is whether we do the unglamorous work of governance now or pay for it publicly later.