Before you spend on AI, check whether you are ready

The instruction usually arrives without much detail. Someone at board level, or a minister, or a managing director has seen what AI can do and wants the organisation to "do something with it". Vendors are already booking demonstrations. A colleague in another agency or bank says they have launched a pilot. You are the person who has to decide where to start and what to spend.

Asking whether your organisation is ready can feel like a delay. In practice, it is what makes the spending useful. You can answer most of the readiness questions in a week, with the people you already have, before any money leaves the budget.

Why pilots stall

The published evidence points less at the AI model than at what surrounds it. In July 2024, Gartner predicted that at least 30% of generative AI projects would be abandoned after proof of concept by the end of 2025. It named four causes: poor data quality, inadequate risk controls, escalating costs and unclear business value. In February 2025, Gartner reported that 63% of organisations in a survey of 1,203 data management leaders either did not have, or were unsure whether they had, the right data management practices for AI. It predicted that through 2026, organisations would abandon 60% of AI projects that are not supported by AI-ready data.

These are global surveys and forecasts, not Nigerian measurements. The useful point is what they describe: most of the reasons a pilot stalls exist before the first tool is bought. That means you can find them in advance.

A self-check you can run this week

Start with one candidate use case, not "AI" in general. Readiness depends on the job. An organisation may be ready to draft internal meeting summaries and not ready to answer customer enquiries about account balances.

Gather four to six people, which we come back to below. For each question, record one of three answers: clear yes, partly or no / don't know. Treat "don't know" as a real answer. It usually points to the gap.

The questions are grouped into four areas: digital transformation, cybersecurity, operational excellence and innovation readiness.

1. Digital transformation: is the foundation in place?

• Is the information this use case needs already digital, and can someone say exactly where it lives?
• Who owns that information, and could you get a representative sample within a week?
• Do the relevant systems share information, or does it move by email, spreadsheet and paper?

A weak answer sounds like "it's all in the system", followed by nobody being able to name the system. Imagine a hypothetical state agency that wants AI to summarise permit applications. Half the applications arrive as scanned paper forms, and each office uses a different version. The first useful project there is a consistent intake form, not an AI tool.

2. Cybersecurity: can you protect what AI will touch?

• Do you know which of this information is personal or sensitive, and is it labelled as such?
• Is there a written rule on what staff may paste into public AI tools?
• If a vendor hosts the system, do you know where your data goes, who can see it, what they keep and how you would get it back?
• Can you see who accessed what, and when?

A weak answer is discovering that staff already use free chatbots with customer or patient information, and nobody has said yes or no. This matters beyond security. In October 2026, the Nigeria Data Protection Commission urged ministries, departments and agencies to prioritise compliance with the Nigeria Data Protection Act 2023. Any AI use case that involves personal data falls within that conversation.

3. Operational excellence: is the process clear enough to improve?

• Can you draw the current process on one page, with a named owner at each step?
• Do you measure it today, for example turnaround time, backlog or error rate?
• When the AI output is wrong, who notices, who corrects it and who is accountable?

Without a baseline, you cannot tell whether a pilot helped. Consider a hypothetical bank complaints team that wants AI to classify incoming complaints. Each branch logs complaints differently, and nobody has a current turnaround figure. A pilot there could look promising in a demonstration and still prove nothing.

4. Innovation readiness: can the organisation absorb the change?

• Is there a named sponsor with budget authority and time to make decisions?
• Will the people who use the system help shape it, and will they have time to learn it?
• Have you agreed in advance what result would justify expanding the pilot, and what would stop it?
• Can procurement and legal support a small, time-limited pilot, or does every purchase follow a full tender?

A weak answer is a pilot whose success measure is "let's see", or one that frontline staff first hear about on launch day.

Why IT should not answer alone

It is tempting to hand this checklist to the IT department. IT can speak with authority on systems, integration and security controls. It usually cannot say whether the process works the same way in every branch, whether the records are reliable, what the regulator expects, or whether the clinicians, officers or relationship managers who will use the output will trust it.

A better group includes the business owner of the process, someone from IT or data, someone from risk, compliance or legal, a frontline supervisor and someone from finance or procurement.

The disagreements between them are often the most useful result. If IT rates the data "good" and operations says half the records are re-keyed by hand, that gap is exactly where a pilot would have stalled. It is better to find it in a meeting room than three months into a vendor contract.

Reading your results

• Mostly clear yes: choose a narrow pilot, record the baseline and agree the stop rule before you start.
• A mix: close the gaps first. A data clean-up, a staff usage policy or a one-page process map often costs less than the tool and makes the eventual pilot easier to judge.
• Mostly no or don't know: hold the tool budget. The most valuable spend right now is on foundations, and that work will serve any technology you adopt later.

None of these outcomes means you are behind. They mean you know where you stand, and you can explain your decision to the people who asked you to "do something with AI".

If you want a structured version

A self-check like this is a good start, but it depends on who is in the room and how candid they are. If you want a more structured picture, VIS offers Vordia, an organisational readiness assessment that covers the same four areas: digital transformation, cybersecurity, operational excellence and innovation readiness.

Vordia collects input from several stakeholders, not just one department. It produces a gap analysis, readiness dashboards and an executive report you can take to leadership. It also includes progress tracking, so you can follow the gaps as you work on them.

The guided conversational assessment costs NGN 29,000 for the first five organisations that sign up by 28 November 2026. Places at this launch price are limited to five organisations. After that, the price is NGN 49,000. If you would prefer a one-on-one or in-person assessment, VIS prices it according to your requirements.

Request a Vordia assessment and the team will reply within two business days. They will walk you through how it works and agree with you whether it is the right first step.

Whichever route you take, run the check before you sign the next AI contract. You will spend with more confidence, and you will have a clear answer when someone asks why you started where you did.

• AI readiness
• Digital transformation
• Cybersecurity
• Operational excellence
• Leadership
• Nigeria