Speak to almost any UK business today and you will find employees using AI. While leadership is welcoming this, they also face the challenge that their people may be adopting the technology faster than organisations are preparing them to use it safely. An employee drafts a customer response with a public AI chatbot to save time. A team analyses data outside approved systems because it is faster than waiting for access. A manager experiments with public AI tools to prioritise workloads or respond to suppliers. These AI tools don’t appear on a roadmap, but they are having a big impact on how work gets done across organisations.
This is what we call “shadow AI”. Our research shows that 68% of UK organisations say employees are using unapproved AI tools at least occasionally. Almost half (44%) of businesses have already seen data or IP exposure, and 43% report security vulnerabilities as a result of shadow AI. These figures should concern boards and risk teams.
It’s clear that people are excited by and see value in AI, and we shouldn’t be trying to stop this behaviour – rather, we should be equipping employees with the skills, tools, and culture to innovate and thrive alongside this transformative technology. What we do need to address, however, is why employees are often moving faster than the organisations that are meant to be leading them.
Investment outpacing enablement
UK businesses are investing heavily in AI. UK organisations, on average, will have spent £15.94 million on AI last year, with plans to increase that investment by an average of 40% over the next two years. Returns are already visible, with companies reporting an average ROI of 17% in 2025, expected to rise to 32% by 2027.
But 60% of businesses say their employees have not completed comprehensive AI training. There is a risk that organisations are asking people to work alongside a powerful new technology without giving them the tools, skills, confidence, or clarity to use it properly. So shadow AI is not rebellion, rather improvisation. Most employees are not trying to bypass governance. They are trying to do their jobs better, faster, and with the tools they see around them every day.
Our research shows that only 7% of UK organisations have an enterprise-wide AI strategy, so AI is being implemented in fragments. People need to understand what is approved, what data is safe to use, and where they could be using AI in their role. AI adoption cannot be treated purely as a technology rollout.
Expanding beyond the UK’s technical skills focus
The Government, and Skills England, released a new report in October 2025 highlighting the UK’s AI skills gap, which if addressed could help unlock up to £400 billion of growth potential by 2030.
The next phase of AI will only amplify this gap, as UK leaders are already looking ahead with 55% saying Agentic AI is influencing their strategic planning over the next two years.
While the AI skills gap conversation is often focussed on technical expertise, it’s also important to recognise that a lot of the challenges are cultural. Companies need to consider investment in employee development and tools as part of their AI strategy, and the shift in company culture and championing AI tools as part of day-to-day work. Those tools need to be clearly communicated. Employees need to understand how outputs are generated, what data is being used, and how human judgement still matters.
Training here is not just about teaching people how to prompt better. It is about helping them feel confident, responsible, and supported.
The skills opportunity starts earlier than you think
If we wait until people are already in senior positions to address the AI skills gap, we are too late. Driving stronger links and collaboration between UK business, education, the third sector, and early careers is key – alongside a focus on lifelong learning throughout an individual’s career journey. This includes, for example, collaboration with social enterprise to support STEM skills and careers, apprenticeships that combine digital skills with real business context, and employers working together through user groups and ecosystems to raise capability across industries, not just within their own walls.
As one of the UK’s largest technology employers, SAP sees first-hand how early exposure, practical training and partnerships shape confidence and career paths. Solely producing more technologists won’t completely fill the AI skills gap, we need to build a workforce that can work intelligently alongside and in collaboration with technology.
What leaders need to do differently in 2026
The organisations making the most progress are not necessarily the ones experimenting the most loudly. They are the ones who are also providing role-based training linked to real workflows, they set guardrails that can be easily understood and followed – not policies that live in documents no one reads, and they are the ones who ensure their people are a core pillar of their company’s AI strategy.
They talk about AI openly and foster a culture of innovation and experimentation. They acknowledge uncertainty. They invite feedback from the front line. And they treat partners as accelerators, not afterthoughts, drawing on external expertise to move faster and more responsibly.
The UK does not lack AI ambition. The scale of investment makes that clear, but we’ve got to back it up. Shadow AI is a reminder that people are ready, even when organisations are not. Success comes from people investment, and businesses that stop treating AI as a side project and start building the skills, trust, and confidence needed to scale the technology across the organisation properly.
What steps are you taking to upskill and empower your people to innovate and thrive alongside AI, and unleash its potential?
