At AI Compass Lagos 2026, the National Information Technology Development Agency did not present artificial intelligence as a distant possibility. Its message to banks, insurers, payment companies and fintechs was more demanding: AI is already changing financial services, so leaders must decide how to scale it without losing accountability, resilience or public trust.

Barr. Abdullahi Abubakar Aliyu delivered the prepared keynote on behalf of NITDA Director-General and CEO Kashifu Inuwa Abdullahi. Early in the address, he reframed the industry’s challenge:

“The question before the financial services industry is no longer whether artificial intelligence will transform the sector. It already is.”

The useful questions, he said, are how quickly institutions can move from experimentation to scale, how AI can create measurable economic value, and how the sector can build “the trust, security, accountability and inclusion necessary for AI-powered financial services to succeed.”

The Question Has Changed

Nigeria already operates at a scale where a weak technology decision can become a public-interest problem. Central Bank of Nigeria data shows that the country recorded 22.42 billion e-payment transactions worth about ₦1.559 quadrillion in the first half of 2024.

The keynote described that scale as “a policy fact.” What institutions get right or wrong about AI governance in payments can affect millions of people quickly.

CBN’s Payments System Vision 2028 reflects the same tension. Its six principles are interoperability, security, inclusion, innovation, trust and collaboration. The keynote argued that these principles also belong at the centre of financial-services AI.

“The future of digital finance will not simply be about making financial services more digital. It will be about making them more intelligent.”

That intelligence can support onboarding, fraud detection, credit assessment, forecasting, claims, customer support and operations. But adoption is not the goal by itself.

“The opportunity is not simply to deploy more AI. The opportunity is to deploy the right AI in the right places with the right safeguards.”

The Problem Is What Happens After the Pilot

The keynote’s sharpest business point was not about model capability. It was about institutional follow-through.

“The problem is not that we are failing to experiment. The problem is what happens after the pilot.”

PwC’s 2026 Africa analysis supports that concern. It found that 82% of surveyed African organisations had participated in AI pilots, compared with 88% among its AI-leader group. Yet median AI spending was 2% of revenue in Africa, against 5% among AI leaders, and only 32% of surveyed African organisations believed their current investment was sufficient to achieve their AI goals.

The gap is not interest. It is the work required to connect a useful model to live systems, controlled data, customer journeys, named decision owners and reliable measurement.

As the keynote put it:

“The harder question is whether that solution can be deployed safely across millions of customers, integrated into legacy systems, governed appropriately, monitored continuously and trusted by both the institution and its customers.”

That changes how success should be measured.

“The next phase of AI in financial services must therefore be defined not by the number of pilots we launch, but by the value we successfully scale.”

Governance Should Create Certainty

Financial institutions sometimes treat governance as the stage that begins after a technical team proves a use case. The keynote argued for the opposite: controls must grow with the effect an AI system can have on a person’s life.

“We should regulate the risk associated with AI, not regulate innovation out of existence.”

At the same time, it warned, “innovation cannot be allowed to outrun accountability.”

That risk-based idea is already visible in Nigeria’s existing policy and legal architecture. The Nigeria Data Protection Act 2023 requires fair, lawful and transparent processing of personal data, accountability, security and impact assessment for likely high-risk processing. It also provides safeguards around decisions based solely on automated processing where those decisions have legal or similarly significant effects.

Nigeria’s National Artificial Intelligence Strategy 2025 sets a wider policy direction through five pillars, including responsible and ethical AI and an AI governance framework. It is a strategy, not a standalone AI Act. In the capital market, SEC Nigeria’s robo-adviser rules offer a concrete example of algorithmic accountability through testing, monitoring, documentation, disclosure and human intervention.

The keynote summarised the value of clear rules in four words:

“Good governance creates certainty.”

It added: “What financial institutions actually want is not less regulation. It is predictable regulation.”

Because one financial product can involve a bank, fintech, cloud provider, payment processor, model provider and several data systems, no single institution can cover every risk alone.

“The future therefore requires regulatory coordination rather than regulatory competition.”

Data Can Scale Exclusion Too

The strongest warning in the address concerned the raw material used to build and operate AI.

“Poor-quality data produces poor-quality AI, and biased data produces biased outcomes.”

In financial services, that can affect who receives credit, whose claim is flagged, which transaction is blocked, or which customer is offered a product. Explainability is not enough if the outcome is still unfair. Human oversight is not meaningful if the reviewer lacks the authority, information or time to challenge the model.

The keynote made the consequence plain:

“We risk automating yesterday’s inequalities into tomorrow’s financial systems.”

For leaders, this turns data quality, customer recourse, fairness testing, model monitoring and documented override paths into production requirements. They are not side notes for an ethics statement.

Risk Changes at Scale

“AI changes the nature, speed and scale of the risk we are managing,” the keynote said.

A pilot can succeed while hiding the dependencies that will matter in production. A model may drift. Fraud tactics may change. An upstream data source may fail. A cloud region or model provider may become unavailable. A vendor may change its model, price or terms.

The Financial Stability Board identifies third-party concentration, cyber risk, market correlations, and model risk and data governance among the AI-related vulnerabilities relevant to financial stability. This is why financial institutions need more than vendor assurances. They need model and dependency inventories, independent testing for high-impact use cases, continuous monitoring, fallback procedures, incident plans and practical exit options.

Trust becomes visible when the system fails. Can a claim, payment, onboarding process or customer-support journey continue safely? Can a human intervene? Can the institution explain what happened and reconstruct the decision?

Trust Is the Infrastructure for Scale

The keynote did not frame trust as a brake on progress.

“Treat trust not as an obstacle to innovation, but as an enabler of it.”

That is the central lesson for financial-services leaders. Trusted AI is not a promise added to a launch announcement. It is an operating system made of lawful data, clear accountability, controlled access, human decision rights, continuous monitoring and resilience across every critical dependency.

“The future of digital financial services will not be determined by who adopts AI first. It will be determined by who can scale AI responsibly, securely and inclusively.”

The Takeaway

The next phase of financial-services AI will not belong to the institution with the most pilots. It will belong to the institution that can choose a valuable workflow, define the human decision point, prove the outcome, and keep the service safe under real volume and real scrutiny.

“The future of financial services will be intelligent. Our task is to ensure that it is also trusted.”

Curacel builds AI infrastructure that helps financial-services businesses digitise operations and scale service delivery across emerging markets. Talk to Curacel about one workflow your team is ready to move from pilot to production.