From the Floor: Reflections from the Skills Symposium

The graduate skills gap is not a new problem. The urgency surrounding it, however, undeniably is. As recruitment, technology, and the labour market accelerate, the distance between what universities develop and what employers need continues to widen, with direct consequences for the students navigating that gap. 

To address it, Cappfinity convened senior leaders from over 35 universities and 15 employers for a day structured around four sessions with the objective of connecting people, and surfacing shared thinking, practices and learnings. This report summarises the key findings from that day and their implications for careers and employability professionals working to close the gap. 

Key findings: 

• The curriculum universities can realistically deliver is falling further behind the skills employers will need next. 

• Skills only transfer between education and employment when both sides share a common vocabulary for describing them and that vocabulary is largely missing. 

• Employers are divided on AI use in recruitment, yet unified in expecting AI proficiency from day one, leaving students to navigate the contradiction alone. 

• Judgement, proactivity, and curiosity are the capabilities that determine whether AI becomes an asset or a crutch for graduates. 

When Curricula Can’t Keep Up 

The clearest finding from universities was that the half-life of a relevant curriculum is shrinking. By the time a course is designed, validated, and delivered, the skills it was built around are already beginning to give way to the next wave of demand.  

Research presented by the University of East London illustrated this directly. The gaps graduates face today centre on communication, time management, and commercial awareness. Looking ahead five years, the priorities shift to AI literacy, emotional intelligence, and cultural agility. This shift is not an anomaly to correct for but a pattern to design for. 

Graduate role specifications of the future do not yet exist in a defined form, which makes them difficult to teach directly or hire for. This points to a different kind of ambition: rather than curricula that must constantly adapt to keep up, universities need to build human capabilities durable enough to outlast any single job specification, whilst remaining flexible enough to scaffold whatever comes next.  

The Case for a Shared Language 

The challenge with skills is rarely whether students have the right ones, but whether they can articulate them in a way that resonates with the people hiring them. As Cappfinity Co-Founder Nicky Garcea noted, that gap is structural; skills need a common framework before they can be meaningfully transferred between education and employment, and that common framework is precisely what is missing at a wider level.  

Both Kingston University and IPSOS have worked to close that gap from their respective sides. Kingston has embedded nine core attributes into the student experience, giving learners and educational leaders a consistent framework to build from. IPSOS developed a "leading self" model focused on helping people understand how they think, so they can interrogate their own assumptions and communicate their strengths more clearly. The approaches differ, but the underlying belief is the same: without a shared vocabulary, the transition from education into employment remains needlessly opaque. 

Both organisations point to the same requirement: structures that work across both worlds, giving students the vocabulary to demonstrate what they bring, giving employers confidence in the talent coming through the door, and making the transition into work a legible process rather than an uncertain one.  

Embedding Skills Structurally, Not Just as an Add-On 

A shared language only works if it's built into the student experience from the start, rather than bolted on at the edges. That means graduate attributes need to be shaped by ongoing consultation and woven through the full arc of a student's journey, supporting genuine reflection over time rather than a one-off exercise. Done well, students build a cumulative picture of their strengths and development areas, that they can draw on when it matters most.  

The University of London’s experience of reaching tens of thousands of students globally sharpens this point. At that scale, clarity is non-negotiable. A deliberately limited set of core employability skills, mapped onto existing academic curriculum, makes development that is already happening visible: naming the skills students are building, contextualising why they’re building them, and explaining how. Consistent shorthand gives students a reference point they can return to across their studies and gives employers something legible at the other end.  

A common vocabulary, consistently applied, gives every graduate a credible way to demonstrate their capabilities regardless of where they started. For a student from a well-connected background, that framework is useful. For a first-generation graduate who has never been shown how to translate their experience into language employers recognise, it is something far more significant. It is the difference between being legible to the market and being invisible to it. It’s not a marginal benefit – it’s a structural one, and one of the most compelling reasons to take the shared language agenda seriously. 

The Tension Around AI 

The AI conversation surfaced a contradiction that no single institution has resolved. Employers currently sit across a wide spectrum: some prohibit AI use in the recruitment process entirely, with proctored assessments and locked browser environments; others discourage it without enforcing boundaries; still others say nothing at all. What cuts across all of them, is the expectation that the successful candidate will be AI-proficient from day one. The restriction on using AI to apply, and the requirements to use it once hired, sit in direct tension and students are navigating that contradiction largely on their own. 

Sam Littel, an engineering student at Loughborough University, actively navigating the internship market, described his own approach: using AI to sharpen his applications, grounding them in each organisation’s context while preserving his own voice and reasoning. His experience raises the more useful question – not whether students should use AI, but whether they are being equipped to use it well. This is not an individual workaround; it is an adaptation to a system that has yet to decide what it wants.  

This points to a further structural gap. For universities to respond meaningfully to employer expectations around AI, those expectations need to be clearly articulated. What does “AI-ready” actually mean?  

The Capabilities That Keep the Graduate in the Driving Seat 

The capabilities that determine whether AI becomes a genuine asset or a crutch are judgement, proactivity, curiosity. As Nicky Garcea observed, these are the same qualities that make a graduate effective in any context, and they are the ones most at risk of atrophying if AI is used as a shortcut rather than a scaffold. The students most likely to navigate this tension successfully are not those who avoid the technology, but those who have developed the awareness and critical thinking to use it well. Sam's approach illustrates this, he used AI to interrogate and sharpen his own thinking, not to replace it. The tool was in service of his judgement, not the other way around. 

The task for universities is not to teach avoidance or passive reliance, but to develop the human capabilities that keep the graduate, rather than the tool, in the driving seat. This is the same argument as the one for a shared skills language; the goal is graduates who know what they bring, can articulate it clearly, and can apply it with judgement, whatever tools or contexts they find themselves working within. 

Recommendations 

The gap between education and employment is real and consequential, but it is addressable, provided the response is shared and connected rather than siloed. 

Universities need a framework that supports consistent skills of development across the curriculum, rather than treating employability as an add-on. Progressive institutions - Kingston University, the University of London and the University of East London among them - are already building this into practice and offer a starting model for others. 

Employers need to articulate what “AI-ready” means in practice and resolve the contradiction between restricting AI in recruitment while requiring it on the job, so students are not left to interpret unstated expectations on their own. 

Students need consistent language to demonstrate what they can bring, developed through structured reflection over time rather than assembled at the point of application. 

This infrastructure will not emerge by accident. It must be designed, embedded, and held consistently across the graduate journey into employment. The tools, roles, and demands of the market will keep shifting; what should not shift is the standard for the graduate on the other side of it: someone who knows what they can bring, can articulate it clearly, and can apply it with judgement, whatever tools or contexts they find themselves working within. 

If you enjoyed the insights from this session, you can check out the full clips from the day here. 

Follow our LinkedIn to keep in touch and find out more about the next symposium and any other upcoming events: click here.