Next Roundtable · Book your seat
Technology

Three-quarters of UK technology employers are struggling to fill roles.

The skills they need most (AI, data, cybersecurity) are exactly the skills that take years to develop and cannot be hired fast enough to meet current demand.

Get in touch See case studies →
75%

Can't fill roles

5yr

IT hardest to source

7.5 million

UK adults lack essential digital skills for work

The challenge

IT and data skills have been the hardest to source in the UK for five consecutive years. In Q1 2025, 75% of technology employers reported difficulty filling roles while simultaneously planning to hire. The gap is structural. Educational pipelines produce graduates faster than the market can absorb in some areas, while the specific competencies employers need most (AI and automation, cloud architecture, cybersecurity engineering) remain chronically undersupplied. A 2024 government review found that 7.5 million UK adults, 18% of the working-age population, lack the essential digital skills needed for the workplace. A further 23.4 million people (approximately 60% of the workforce) cannot complete all 20 digital tasks that government and industry have defined as essential for employment. These are not niche gaps. They are a structural deficit that no employer can hire their way out of alone.

For early talent specifically, the challenge is twofold. AI-assisted applications have made it genuinely difficult to distinguish candidates who have strong AI literacy from those who used AI tools to produce a convincing submission. The skills that matter most (AI fluency, learning agility, the ability to adapt quickly to new tools) are the hardest to assess from a written application and the most commercially critical to get right. Meanwhile 71% of UK organisations report a persistent cybersecurity skills shortage, up from 57% the previous year. AI-related job postings are growing at more than three times the rate of the broader jobs market. The talent required is in acute short supply and being competed for by every sector simultaneously.

The government’s June 2025 announcement of a partnership with industry to train 7.5 million workers in essential AI skills by 2030 signals the scale of the structural investment required. The Level 4 AI and Automation Practitioner Apprenticeship, launched in March 2026, creates a funded route through the Growth and Skills Levy for employers who want to build that capability from entry level. Employers who build the early talent pipeline for these roles now will have a structural advantage within three years.

75% cannot fill roles

the shortage is persistent enough to become a structural brake on delivery.

5 straight years

IT and data have remained hardest to source, so employers cannot assume the market will correct itself.

AI use is not AI fluency

the real problem is separating polished applications from demonstrable technical judgement.

Our approach

Talent Assess measures AI Skills as one of its four core competency dimensions. Candidates demonstrate how they actually use AI tools through real work tasks, not a knowledge test. The individual calibration model in the Video Interview Platform is particularly relevant here: it establishes each candidate’s natural communication baseline through a non-assessed warm-up question, then scores structured interview responses against that individual baseline rather than a population norm. The result is a reliable signal of authentic technical communication that is distinct from AI-polished presentation.

The Talent Portal supports the end-to-end recruitment process for technology apprenticeship cohorts, including the new Level 4 AI and Automation Practitioner standard. Interest-led matching identifies candidates who are genuinely drawn to technical roles rather than applying speculatively. Talent Connect manages engagement across what can be a longer recruitment cycle for specialist technical programmes, with automated nudges, in-flight analytics and multi-channel communication.

Test applied AI skills

candidates complete structured tasks that reveal real AI fluency, not borrowed language about tools.

Calibrate the person, not the polish

baseline scoring reduces the advantage held by candidates with more coaching or interview practice.

Keep scarce candidates engaged

specialist cohorts move through longer technical hiring cycles without dropping out of the funnel.

Relevant products

Platform

Talent Portal

An early talent ATS built on a headhunting model. Attraction, matching, engagement, eligibility, compliance, communications and reporting — in one platform built for this market.

  • Interest-led matching and proactive candidate search — find candidates on their interests and location before they apply
  • Multi-channel communications across email, WhatsApp, SMS and phone, all from one candidate record
  • Integrated eligibility checking, compliance management and reporting built for early talent routes
Explore Talent Portal →
Platform

Talent Assess

Most assessment tools measure how well a candidate performs under pressure for 45 minutes. Talent Assess measures what they actually did over weeks of real engagement. The data is richer, more reliable and harder to game.

  • Intent Score ranks candidates by motivation using four behavioural-science constructs — attitude, perceived fit, self-efficacy and engagement actions. Not clicks or page views
  • Skills assessment across Communication, Critical Thinking, Learning Agility and AI Skills through real work tasks
Explore Talent Assess →
Platform

Video Interview Platform

Scoring each candidate against their own natural communication baseline removes the structural advantage that accents, coaching access and social capital would otherwise confer. That is a design decision, not a compliance patch.

  • Individual calibration model — every candidate scored against their own baseline, not a population norm
  • Live bias monitoring dashboard tracks scoring patterns across the review team in real time, flagging statistical anomalies before they become decisions
  • Every scoring decision, rubric and bias alert is documented — if a decision is challenged, the evidence is there
Explore Video Interview Platform →
Solution · Attract

Candidate Attraction & Engagement

Most early talent processes are designed to filter the people who find you. Very few are designed to find the people who should. In-flight campaign data. Multi-channel engagement. Real-time pipeline analytics.

  • In-flight conversion data by channel, message and candidate group — visible while the campaign is live
  • Pipeline diversity visible in real time, against targets set at the start of the campaign
  • Candidate drop-off recovered automatically through multi-channel follow-up
Explore Candidate Attraction & Engagement →

Common questions

How do I assess AI skills in early talent candidates?

The challenge with assessing AI skills in early talent is that most entry-level candidates have used AI tools but few have used them in a structured, reflective way. What employers actually need is not proof of tool use. It is evidence of AI fluency: the ability to use AI purposefully, evaluate its outputs critically, and apply the results to real work. The most reliable way to assess this is through a structured task rather than a question-and-answer format. Candidates who have genuinely developed AI fluency will demonstrate it through how they approach a problem; those who have not will rely on AI to answer questions about AI, which produces a circular and unreliable result. Talent Assess measures AI fluency as a distinct competency dimension, assessed through task completion rather than self-report.

How do I tell if a candidate used AI to write their application?

The short answer is that you cannot tell reliably from the application itself. AI detection tools have high false positive rates, and more experienced candidates have learned to edit AI outputs to evade detection. The more useful question is why it matters: if the application is the primary evidence base, you are dependent on a document that may tell you nothing accurate about the candidate. The structural answer is to build evidence upstream of the application (through Immersive Work Experience, through structured work tasks via Talent Assess) so the application is corroborating a picture you have already built, not the only picture you have.

What are the hardest tech skills to recruit in the UK right now?

According to the Manpower Experis 2025 Talent Shortage Survey, IT and data skills have been the hardest to source in the UK for five consecutive years, a position unchanged since 2020. Within that, AI and machine learning engineers, data analysts and scientists, and cybersecurity engineers are the most sought-after roles with the lowest available supply. 71% of UK organisations reported a persistent cybersecurity skills shortage in 2024, up from 57% the previous year. AI-related job postings are growing at more than three times the average rate across all job categories.

What is the Level 4 AI and Automation Practitioner Apprenticeship?

The Level 4 AI and Automation Practitioner Apprenticeship was launched in March 2026. It is designed for employers who want to develop applied AI skills across their workforce at entry and junior level, rather than relying on graduate recruitment alone. The standard is fundable through the Growth and Skills Levy. It covers AI tool application, automation, data literacy and responsible AI use in the workplace. For technology employers with existing levy balances, it is a route to building a pipeline of AI-competent early talent that does not depend on the graduate hiring market.

Talk to us about building a technology talent pipeline that identifies genuine AI literacy.

Get in touch