The Institute of Student Employers published its annual Student Development Survey in early 2026. It is a substantial piece of research — 144 employers across eight sectors, covering everything from programme design to retention rates to budget allocation. As a benchmarking tool, it is useful. But read alongside the wider conversation about AI and capability formation, it surfaces something less expected: a quiet signal that what employers most want may be harder to form than what they currently measure.
The survey becomes particularly interesting when the skills employers prioritise are compared with the areas where new entrants most often fall below expectations.
What employers say they need
The survey’s skills data is, on the surface, familiar. Problem solving and decision making, working with others, motivation and self-awareness: these sit at the top of what employers recruit for, consistently, across both graduate and school and college leaver pipelines. They are the transferable capabilities that have appeared in employer surveys for decades.
But the readiness data tells a different story. When employers assess what new hires actually bring on entry, three areas stand out as consistently below expectations: adaptability, motivation and self-awareness, and awareness of the wider context. These are not peripheral skills. They are the same capabilities employers most frequently say they need.
Around a third of employers rated graduates as below expectations in each of these three areas. For school and college leavers, adaptability reaches the same threshold. The gap is persistent, and it is growing in visibility.
The distinction the data keeps raising
There is a distinction worth drawing carefully here, because the ISE data invites it without quite making it.
Employers are largely satisfied with what might be called foundational performance: reading, writing, numeracy, listening, basic communication. These skills are meeting expectations in most cases. What is not meeting expectations — consistently, across sectors, at both entry routes — are the qualities that sit underneath performance: the capacity to adapt when context changes, to sustain motivation and direction without constant external structure, to understand how one’s role sits within a wider system. Increasingly, this also includes organisational understanding: the ability to recognise how organisations make decisions, manage uncertainty, shift priorities, and coordinate action in practice.
These are not skills in the conventional sense. They are not produced by instruction, measurable through assessment, or developed through short interventions. They form more slowly, through accumulated engagement with uncertainty, real consequence, and reflective adjustment over time.
The survey also records that employers are almost twice as likely to develop digital literacy as they are to recruit for it — suggesting that technical and AI-related skills are increasingly being treated as developable after hiring. What employers are still trying to recruit for, and finding increasingly hard to find, are the underlying qualities that technical development cannot easily substitute for: judgement, contextual reasoning, self-direction.
What AI is doing to the formation question
The survey’s AI section adds a further layer. Employers are not, on the whole, anticipating large-scale role replacement. What most expect — 87% across all sectors — is that entry-level roles will be reshaped rather than eliminated. The skills that will matter more are critical thinking and judgement, adaptability, and the capacity to evaluate AI outputs responsibly. The skills that will matter less are the routine cognitive tasks — research, drafting, data processing — that once structured the early developmental years of many professional careers.
This creates a quiet but consequential problem. The routine tasks that AI is automating were not merely administrative. They were the environments in which early-career professionals gradually developed contextual understanding, pattern recognition, and professional judgement. If those tasks disappear, so do the conditions under which capability has historically formed. The question is therefore not only whether AI changes entry-level work, but whether the developmental environments that once supported judgement, pattern recognition and professional identity remain available in the same way.
Employers appear aware of this tension, even if the survey does not articulate it directly. Over half have increased the time devoted to professional conduct during induction — the transition from education to workplace norms, the development of self-awareness and organisational understanding. These are not skills being bolted on. They are capabilities that development teams are recognising they can no longer assume will arrive pre-formed.
A formation question, not only a recruitment one
The ISE data is drawn from organisations reflecting on what arrives at their door. It is, necessarily, a post-entry view. But the readiness gaps it describes — the adaptability deficit, the self-awareness shortfall, the absence of wider contextual understanding — point to a formation question that begins earlier than recruitment and cannot be resolved by it.
These qualities are not formed through instruction. They are formed through structured engagement with real uncertainty — through encountering problems that do not have predetermined answers, reasoning under conditions of incomplete information, and building the capacity to take and defend a position. That kind of formation is not easily compressed into induction, however well designed.
What the survey may be signalling — without quite naming it — is that the conditions for capability formation need to exist upstream of organisational entry. And that the distinction between performance and capability, which has long been implicit in employer expectations, is becoming harder to ignore as AI makes high-quality performance outputs increasingly easy to generate while the underlying capability remains unverified.
A convergence worth noticing
The ISE report is one data point. But it sits alongside other signals pointing in a consistent direction: the question of how capability forms — not how performance is produced — is becoming more consequential, not less, as AI becomes more embedded in both learning and work.
Employers are investing in AI literacy, in digital skills, in structured induction for professional conduct. These are rational responses to visible gaps. But the deeper gaps — adaptability, self-direction, contextual judgement — are not the kind that respond well to programme additions. They require environments designed for capability formation from earlier, and sustained across a longer developmental arc.
That is not a criticism of what employers are doing. It is a reflection of where the challenge is located. If the formation environments that once produced adaptability, self-awareness, and contextual reasoning are changing structurally, the responsibility for maintaining them cannot sit at the point of entry alone.
The signals are converging. The question of where capability forms — and whether it is being formed — is becoming the central one. The ISE data suggests employers are beginning to feel the consequence of that question, even where they have not yet fully named it.
Disclaimer
This article reflects the working perspectives of CognateUK and is intended to support informed discussion. It does not constitute advice or represent the official positions of any affiliated organisations or partners.
Image caption: The ISE's 2026 Student Development Survey raises questions about where adaptability, judgement and contextual understanding are formed.
