Most organisations have now adopted AI in some form. The tools are live, people are using them and adoption numbers look good. What has not changed is how most of those organisations actually work.
Grant Thornton’s 2026 AI Impact Survey of 950 senior leaders found that only 12 per cent believe their workforce is ready to use AI effectively. In the same survey, 39 per cent of CIOs and CTOs said the workforce was prepared, compared with just 7 per cent of COOs. That means the people buying the technology believe the organisation is ready, while the people running the operations do not.
The same survey found that organisations with fully integrated AI were nearly four times more likely to report revenue growth than those still running pilots and that the difference came down to whether leadership, governance and working practices had changed around the technology.
In other words, adoption alone is not enough.
The board hears the deployment story
When the CIO briefs the board on AI progress, the update usually focuses on deployment. But there are questions underneath that most organisations are not asking. Can the people on the ground actually use AI in their work? Does anyone know what to do when it goes wrong? The AI strategy ends up built on whichever version of reality reaches the boardroom first.
More than three-quarters of senior leaders in the Grant Thornton survey said they were not fully confident they could pass an AI governance audit. If the people accountable for AI outcomes cannot say the governance works, the organisation has a leadership problem dressed up as a technology programme.
Shadow AI is a leadership signal
While leaders debate readiness, employees have already made their own decisions. A BlackFog survey of 2,000 workers in the UK and US, published in January 2026, found that 49 per cent are using AI tools their employer has not approved. Among C-suite members and presidents, 69 per cent said they were comfortable with this, prioritising speed over security.
Sensitive data flowing into unapproved tools is a real cybersecurity risk, but shadow AI also tells you that the official way of doing things is failing. Employees go around the system because it does not work well enough for their needs. The tools they pick, the tasks they use them for and the workarounds they build show leadership exactly where AI is useful and where it falls short. When nearly half the workforce is finding its own way, the system is the problem.
That information is already moving through the organisation, but leadership cannot see it. Until the organisation builds the means to capture it, adoption will keep growing and capability will stay flat.
Training is where most organisations start and stop
Reaching for training as a solution is understandable, but it only addresses AI awareness. The harder question is whether anyone has changed how the work gets done.
Docebo surveyed 2,000 enterprise respondents across six countries for its 2026 AI Readiness Gap report. It found that 85 per cent of employees say the training they receive does not help them use AI in their role and only 9 per cent of organisations have used AI to redefine their workflows. So, awareness is growing, but the way people actually work has barely moved. That is the difference between knowing what AI can do and knowing how to make it work inside the job you already have.
Capability develops when someone uses AI on a real task, reviews the output with their manager and they work out together where the tool helped and where it did not. It develops when someone can say, “I tried this with AI and it did not work”. The team then learns from it rather than moving on. That process needs governance that sits at the point of work, not in a policy document three levels above it. Training can prepare people for that, but it cannot replace it.
The gap keeps widening
AI’s development cycle moves in weeks, while organisational change moves in quarters and years. Every model release widens the distance between what is possible and what most organisations can actually handle, and that distance keeps compounding.
The organisations closing this gap are the ones treating AI as a management discipline. They are changing how people are managed, supported and held accountable while also ensuring that people using AI are guided, governed and able to learn from what they find. They are helping their people figure out how to use AI on real tasks, providing governance that catches problems where the work happens and creating systems that turn what individuals discover into something the whole team can use.
Other companies will continue to track licences, run workshops and report numbers, while the distance between what they have adopted and what they actually build keeps growing. They will adopt AI, but they will not change.
About the author: Fahed Bizzari is an organisational AI strategist who argues that while most organisations have adopted AI, few have become empowered by it. He has over twenty years’ experience of helping organisations restructure how they work around technology. He is the Managing Partner of AI empowerment consultancy, Bellamy Alden, whose clients include L’Oréal, Fugro, Atlas Copco, Dubai Police and Shiseido. His work has been cited in MIT Sloan Management Review and The National, and he is the author of the upcoming book AI Empowerment for Business Leaders (Rethink Press).