Breaking News
Your Staff are Already Using AI, Is Your Company Ready?
Kayode Jaiyeola, a technology lead who governs AI rollouts for regulated organisations in the UK, says the real risk in workplace AI is not the tool itself. It is what the tool is allowed to see.
Before Kayode Jaiyeola arrived at one UK organisation, it had already switched on Microsoft Copilot for its staff. The AI assistant worked as promised. It also, given the right prompt, could reach the company’s payroll.
“It’s one thing to have the AI itself,” Jaiyeola says. “It’s another thing to put controls on it.”
Fixing that gap has become a large part of his work. With over 15 years in technology, from Etisalat and MedPlus in Nigeria to financial services, healthcare and the public sector in Britain, he has moved with the industry from on-site data centres to the cloud, and now to AI. Today, he leads AI adoption and governance for organisations where a data leak is not an embarrassment but a regulatory event.
His message to Nigerian boardrooms is direct: your employees are already using AI. The question is whether your company is ready for it.
The numbers back him up. In Microsoft and LinkedIn’s 2024 Work Trend Index, a survey of 31,000 knowledge workers in 31 countries, three in four said they use AI at work. Of those, 78 per cent brought their own AI tools rather than wait for their employers to provide them. At small and medium-sized companies, the figure rose to 80 per cent.
So the question is no longer whether staff will use AI, but whether their company is ready when they do.
In the payroll case, the fix was not dramatic. Using Microsoft Purview, a data protection tool, the organisation identified its sensitive information and applied labels that tell AI assistants what they can and cannot touch.
“It’s simple, to be fair,” he says. “Companies need to identify that data, such as payroll or other sensitive information, apply the right sensitivity labels, and let the tool do its job. Everyone wants to jump on the AI wagon. But are you doing it in a way that also protects the company’s interest?“
Readiness, in his view, starts long before any software is switched on. First, identify where AI adds measurable value. A retailer whose customer service team spends its day on refunds and missing deliveries is a good candidate for an AI assistant. Second, check the data. “AI deals with what you put in,” he says. “If the data isn’t accurate, there’s really no point.” Third, settle the basics: who owns the system, what it costs, what security controls apply and what outcome you expect.
When he built customer service assistants for a retail business, handling refunds, delivery complaints and website orders, the groundwork came first. He mapped how the business actually handled each request, spoke to the staff who dealt with customers every day, and set firm limits on what the assistant could and could not access.
The bigger risk, he says, sits with employees pasting company or customer information into free AI tools. His answer is not a blanket ban but what he calls a company bubble: enterprise versions of tools such as ChatGPT and Copilot, configured so that anything staff type stays within the organisation’s own environment.
In financial services, his team went further, adding software that scans what employees enter into the AI tool. If someone pastes card details or personal data, it is flagged, and the user is told to remove it before going any further.
Should companies ban public AI tools altogether?
“It depends on the organisation,” he says.
Highly regulated sectors such as banking and healthcare need tight controls. Everyone else needs guidance and a safe alternative. “If you don’t provide approved environments within the company, you force people to go to AI tools that aren’t regulated. People will find a way to use them one way or another.”
Having written AI policies for staff, Jaiyeola keeps them to a handful of rules.
● Approved tools only, especially when handling company or customer personal data.
● Label the data, so the AI knows what it may and may not look at.
● Company devices and accounts: no pasting company data into a personal AI account. If you need a new AI tool, raise it with IT and work it out together.
● A human stays accountable. “AI can help you draft something and help with your task,” he says. “But as a human being, you need to check the output. You see people copying things from AI blindly and pasting them. I’m like, really, guys? Really?”
Why Nigeria should pay attention
The risks are no longer limited to leaks. At one company he worked with, criminals used AI to clone the chief executive’s voice and tried to use it to manipulate clients and customers. For Nigerian banks and fintechs already fighting fraud, that should ring alarm bells.
Those same institutions, along with insurers and hospitals, hold vast amounts of personal data now governed by the Nigeria Data Protection Act 2023. Their staff have the same free AI tools on their phones as anyone in London. An assistant that can surface payroll in a UK firm can just as easily surface customer account details in Lagos.
For a Nigerian SME, his approach comes down to three starting steps: know what data you hold and label what is sensitive; give staff an approved AI tool so they are not tempted by unapproved ones; and write simple rules that keep a human responsible for whatever the AI produces.
None of this makes him wary of the technology. Asked what excites him most about AI, his answer is short. “The unknown,” he says. “Where is this leading us? We don’t know yet.”
What he does know is where the responsibility sits. “AI is fantastic,” he says. “At the end of the day, it’s about how you use it, how you govern it, and how you protect sensitive data.”
