What the data and commentary are telling us
Recent survey releases and economic commentary point to a resilient but cooling labour market for graduates. The HESA Graduate Outcomes 2022/23 release shows that 88% of graduates were in work or further study 15 months after graduation and 59% were in full time employment. Unemployment among that cohort stood at 6% and 5% were in full time further study. Subjective measures remain strong with 85% describing their activity as meaningful and 77% saying it aligned with their future plans. Source HESA Graduate Outcomes 2022/23 published 17 July 2025.
At the macro level, think tanks and commentators are flagging structural pressures. Weak productivity growth, rising costs and a wave of business closures are reshaping demand and reallocating jobs across sectors. The Resolution Foundation has described a period of “creative destruction” where less efficient firms are being replaced, and employment is shifting between sectors. Source Jo Faragher, Personnel Today, January 2026.
On the ground many employers report sharp increases in applicant numbers. At Cripps we have seen graduate applications rise by 22% and apprenticeship applications by 94% this year. 2025 saw applications more than double; we are expecting this trend to continue. Large organisations routinely receive hundreds of thousands of applications for flagship programmes, which makes manual shortlisting impractical without technological support.
Why AI is being used and why it worries people
There is a simple operational logic behind the adoption of AI in recruitment. Automated systems scale. They can apply consistent rules to thousands of applications, reduce reviewer hours, and speed up shortlisting. For organisations facing a surge in volume, automation is not a luxury, it is a necessity. But automation brings real risks. Off the shelf models and poorly designed filters can be brittle. They can reproduce historical bias embedded in training data. They can prioritise easily measurable signals over the human qualities that matter for long term performance. And when automated steps are opaque, applicants feel dehumanised and left in the dark.
Headlines that recruitment is “inhumane” when driven by automation capture a genuine anxiety. The problem is not AI per se. The problem is how it is designed, deployed, and governed.
A balanced view
AI in volume recruitment is a useful tool when used correctly. It can help employers identify talent they would otherwise miss and free human assessors to focus on higher value judgement. But it must be paired with deliberate design choices that protect fairness and the applicant experience. If organisations treat AI as a shortcut to remove human involvement entirely, the result will be poorer hiring decisions and reputational damage. If they treat AI as an augmentation that reduces administrative burden while preserving human oversight, the result can be better outcomes for both employers and candidates.
Practical recommendations for organisations and applicants
For organisations
- Run public insight sessions to explain roles, the application process and the organisation’s culture and to answer questions in real time.
- Publish clear guidance and short videos that explain each stage of the process and what is being assessed.
- Design assessments deliberately and test the end-to-end process for bias and fairness before launch.
- Provide human contact points in addition to automated agents so applicants can ask questions.
- Use human judgement for later stages and ensure all assessors are trained and use consistent criteria.
- Make reasonable adjustments straightforward and visible so applicants with accessibility needs can participate fully.
For applicants
- Target your applications to roles that match your skills and eligibility rather than applying everywhere. Quality beats quantity.
- Respect eligibility criteria but recognise that eligibility is different from doing every task in a job description. If you meet the core requirements, apply.
- Use AI as a drafting tool only. Use generative tools for research and to structure answers, but always personalise, fact check and add your own insight.
- Request adjustments early for timed tasks or accessibility needs and follow the organisation’s guidance on how to do so.
Conclusion
The UK talent market is under pressure from economic headwinds and structural change. AI is already embedded in many high‑volume recruitment processes and will remain a feature of modern hiring. That is not inherently bad. The test for employers is whether they use AI to scale a fair, transparent and humane process or whether they use it to hide behind opaque filters that amplify frustration. Used well, AI can be a powerful ally in matching people to roles. Used badly, it will compound the very problems it was meant to solve. Employers must therefore invest in design, testing and human oversight and commit to a positive applicant experience. Applicants must adapt their approach and use new tools wisely while asserting their individuality. Getting that balance right is the only way to turn this perfect storm into an opportunity.
