McKinsey, BCG, and Deloitte all published the same finding in early 2026. Organisations that layer advanced AI technology onto existing structures without redesigning workflows get almost nothing back. Sometimes less than nothing.
They call it the "silicon ceiling." The tools are purchased. The licences are active. The frontline adoption stalls completely. Expensive technology sits unused because nobody was trained to work with it, nobody redesigned the processes around it, and nobody thought about what the humans in the system actually needed to change.
74% of major corporations say their training systems are failing to keep pace with the demand for new skills. The global corporate learning market is worth $400 billion and most of it is broken.
Schools should pay attention. Because schools are about to hit the same ceiling.
The school version of the silicon ceiling
A school buys a suite of AI tools. Maybe it is a chatbot platform. Maybe it is an AI-assisted marking tool. Maybe it is a content generation system for lesson planning. The purchase is made. The INSET is delivered. Two terms later, three teachers use it regularly and everyone else has forgotten their login.
This pattern repeats across the country. It is not a technology problem. It is a literacy problem. The teachers were not trained to understand what the tool does, why it works the way it does, or how to evaluate whether its outputs are reliable. They were shown how to click the buttons and left to figure out the rest.
The corporate research is explicit about this. You cannot capture productivity gains from AI without intentionally redesigning how people work. That means rethinking workflows, decision points, and the relationship between human judgement and machine output. A one-hour demo does not do that. A single INSET does not do that.
What works instead
Cognizant launched an AI-native learning platform in early 2026 called Skillspring. It does not pull workers out of their jobs for a training day. It delivers learning conversationally, in context, matched to the task the person is actually doing. It maps skills acquisition to measurable business outcomes in real time.
That is the model that works. Not "here is a course about AI." Instead: "here is the task in front of you, and here is how AI connects to it."
Schools need exactly the same shift. A Geography teacher does not need a course on machine learning. She needs a tool that takes her existing lesson on population data and shows her, in that moment, how it connects to AI literacy. With activities she can use that afternoon. With coaching language she can say out loud to her class. With support, challenge and differentiation for the students who need a different entry point.
That is contextual learning. It is what the best corporate L&D teams have spent billions working out. And it is what schools can access right now for a fraction of the cost.
The leadership gap
The research also identifies a leadership problem. Organisations hit the silicon ceiling when leadership treats AI adoption as a technology project rather than a people project. The technology is easy. Getting humans to change how they work is hard.
In schools, that means AI literacy cannot live with the IT technician or the Computing department. It has to be owned by curriculum leadership. The person who decides what gets taught in every subject needs to decide where AI literacy fits. And they need a system that makes that deliverable, not just a vision statement in the school development plan.
The World Economic Forum is clear that the bottleneck to capturing AI's economic value is not computational power or algorithmic sophistication. It is foundational human literacy. That applies to businesses. It applies to governments. And it applies to schools.
AILitKit is built for that bottleneck. It gives curriculum leaders a way to deliver AI literacy across every subject without depending on every teacher becoming an AI expert first. The tool does the subject-specific mapping. The teacher does the teaching.
No silicon ceiling. No unused licences. No training day that everyone forgets. Just a tool that works the way teachers already work, starting with the lesson they are already going to deliver.
Matthew Wemyss is the founder of AILitKit and IN&ED, and author of AI in Education: An Educator's Handbook.