The five readiness gaps that stall Nigerian AI pilots — and how to close them

Most AI pilots in Nigerian enterprises do not fail loudly. They stall quietly — a proof of concept that impressed a steering committee, then never reached production; a vendor demo that went nowhere after procurement questions; a data science hire who left after a year of fighting for access to data. When we run Muller AI Readiness Index assessments, the same five gaps appear with remarkable consistency, and almost every stalled initiative traces back to one of them.

Gap 1: Ambition without a ranked opportunity portfolio

The most common state we find is enthusiasm without prioritisation: a leadership team that wants "AI in the business" but has never ranked use cases by value, feasibility, and risk. The result is either paralysis or, worse, a first project chosen by vendor availability rather than business value. The fix is disciplined and fast: a structured opportunity scan — the core of our two-week Opportunity Sprint — that produces a ranked portfolio with an honest business case for the top three candidates. Organisations with a ranked portfolio make funding decisions in weeks; organisations without one debate indefinitely.

Gap 2: Data that exists but cannot be reached

Nigerian enterprises rarely lack data. They lack reachable data. Customer records split across a core banking system, a CRM, and departmental spreadsheets; scanned documents that were digitised as images and never OCR'd; call logs sitting with an outsourced contact centre under a contract that never anticipated analytics. Readiness scores on the data dimension are routinely the lowest we measure — and the most overestimated by executives beforehand. Closing this gap does not require a multi-year data warehouse programme. It requires scoping the first use case around data you can actually access in ninety days, and using that pilot to justify the deeper plumbing.

Gap 3: Capability concentrated in one person

A single champion — often a head of digital or an ambitious CTO — carries the entire AI agenda. When they are promoted, poached, or simply overloaded, the initiative dies. Institutions that sustain momentum spread capability deliberately: executives who can interrogate an AI business case, middle managers who can specify a use case, and a delivery bench that does not depend on one heroic individual. This is the design logic behind our Executive Masterclass and corporate cohorts — not awareness training, but building enough shared fluency that the organisation's AI judgment survives personnel changes.

Gap 4: Governance treated as a launch-day formality

Pilots stall at the exact moment they threaten to succeed: the compliance review before production. Legal asks where the data goes; nobody knows. Risk asks who is accountable for model errors; nobody has decided. The pilot enters a review loop it never exits. The fix is sequencing: governance artefacts — data protection impact assessment, model risk classification, human-oversight design — are produced during the pilot, not after it. A pilot that arrives at the production gate with its governance file complete gets approved. One that arrives without it becomes shelfware.

Gap 5: No defined success gates

Ask a stalled pilot's sponsor what success looked like, and you will usually hear a retrospective rationalisation. The pilot began without quantified gates — target accuracy, cost-per-transaction threshold, adoption rate — so it could neither pass nor fail; it could only continue or fade. Boards should refuse pilot budgets that arrive without go/no-go criteria. Our Pilot-in-a-Box structure enforces this: success gates are agreed in writing before a line of code is deployed, and the final gate review produces one of two outcomes — a production business case, or a documented decision to stop with lessons captured. Both are wins; the drift between them is the loss.

Readiness is a score, not a feeling

Every one of these gaps is measurable, and measured gaps become closeable plans. That is the premise of the Readiness Index: score the organisation honestly across strategy, data, people, and governance; identify the binding constraint; fix that constraint first. If you want to know where your organisation stands, the five-minute self-assessment on this site will give you a directional score — and a conversation with us will give you the full diagnostic.